From 7b7acdf369aabb8069e45d4e624b9f78644dc065 Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Fri, 13 May 2022 10:51:44 +0800 Subject: [PATCH 01/16] Support --iter in export.py (#360) --- .../pruned_transducer_stateless2/export.py | 42 ++++++++++++++++--- .../pruned_transducer_stateless3/export.py | 42 ++++++++++++++++--- 2 files changed, 74 insertions(+), 10 deletions(-) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/export.py b/egs/librispeech/ASR/pruned_transducer_stateless2/export.py index b5757ee8c..6b3a7a9ff 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/export.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/export.py @@ -51,7 +51,11 @@ import sentencepiece as spm import torch from train import get_params, get_transducer_model -from icefall.checkpoint import average_checkpoints, load_checkpoint +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) from icefall.utils import str2bool @@ -64,8 +68,19 @@ def get_parser(): "--epoch", type=int, default=28, - help="It specifies the checkpoint to use for decoding." - "Note: Epoch counts from 0.", + help="""It specifies the checkpoint to use for averaging. + Note: Epoch counts from 0. + You can specify --avg to use more checkpoints for model averaging.""", + ) + + parser.add_argument( + "--iter", + type=int, + default=0, + help="""If positive, --epoch is ignored and it + will use the checkpoint exp_dir/checkpoint-iter.pt. + You can specify --avg to use more checkpoints for model averaging. + """, ) parser.add_argument( @@ -74,7 +89,7 @@ def get_parser(): default=15, help="Number of checkpoints to average. Automatically select " "consecutive checkpoints before the checkpoint specified by " - "'--epoch'. ", + "'--epoch' and '--iter'", ) parser.add_argument( @@ -141,7 +156,24 @@ def main(): model.to(device) - if params.avg == 1: + if params.iter > 0: + filenames = find_checkpoints(params.exp_dir, iteration=-params.iter)[ + : params.avg + ] + if len(filenames) == 0: + raise ValueError( + f"No checkpoints found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + elif len(filenames) < params.avg: + raise ValueError( + f"Not enough checkpoints ({len(filenames)}) found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) else: start = params.epoch - params.avg + 1 diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/export.py b/egs/librispeech/ASR/pruned_transducer_stateless3/export.py index 29acc7181..0cdb0b957 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/export.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/export.py @@ -52,7 +52,11 @@ import sentencepiece as spm import torch from train import get_params, get_transducer_model -from icefall.checkpoint import average_checkpoints, load_checkpoint +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) from icefall.utils import str2bool @@ -65,8 +69,19 @@ def get_parser(): "--epoch", type=int, default=28, - help="It specifies the checkpoint to use for decoding." - "Note: Epoch counts from 0.", + help="""It specifies the checkpoint to use for averaging. + Note: Epoch counts from 0. + You can specify --avg to use more checkpoints for model averaging.""", + ) + + parser.add_argument( + "--iter", + type=int, + default=0, + help="""If positive, --epoch is ignored and it + will use the checkpoint exp_dir/checkpoint-iter.pt. + You can specify --avg to use more checkpoints for model averaging. + """, ) parser.add_argument( @@ -75,7 +90,7 @@ def get_parser(): default=15, help="Number of checkpoints to average. Automatically select " "consecutive checkpoints before the checkpoint specified by " - "'--epoch'. ", + "'--epoch' and '--iter'", ) parser.add_argument( @@ -142,7 +157,24 @@ def main(): model.to(device) - if params.avg == 1: + if params.iter > 0: + filenames = find_checkpoints(params.exp_dir, iteration=-params.iter)[ + : params.avg + ] + if len(filenames) == 0: + raise ValueError( + f"No checkpoints found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + elif len(filenames) < params.avg: + raise ValueError( + f"Not enough checkpoints ({len(filenames)}) found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) else: start = params.epoch - params.avg + 1 From 48a6a9a54918874736b9f5b0235c76b63e266807 Mon Sep 17 00:00:00 2001 From: Guanbo Wang Date: Thu, 12 May 2022 23:03:26 -0400 Subject: [PATCH 02/16] GigaSpeech RNN-T experiments (#318) * Copy RNN-T recipe from librispeech * flake8 * flake8 * Update params * gigaspeech decode * black * Update results * syntax highlight * Update RESULTS.md * typo --- .flake8 | 1 + egs/gigaspeech/ASR/README.md | 7 +- egs/gigaspeech/ASR/RESULTS.md | 80 +- .../pruned_transducer_stateless2/__init__.py | 0 .../asr_datamodule.py | 416 +++++++ .../beam_search.py | 766 ++++++++++++ .../pruned_transducer_stateless2/conformer.py | 1038 +++++++++++++++++ .../pruned_transducer_stateless2/decode.py | 557 +++++++++ .../pruned_transducer_stateless2/decoder.py | 103 ++ .../encoder_interface.py | 43 + .../pruned_transducer_stateless2/export.py | 182 +++ .../gigaspeech_scoring.py | 1 + .../pruned_transducer_stateless2/joiner.py | 67 ++ .../ASR/pruned_transducer_stateless2/model.py | 193 +++ .../ASR/pruned_transducer_stateless2/optim.py | 331 ++++++ .../pruned_transducer_stateless2/scaling.py | 702 +++++++++++ .../ASR/pruned_transducer_stateless2/train.py | 977 ++++++++++++++++ 17 files changed, 5458 insertions(+), 6 deletions(-) create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/__init__.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py create mode 100755 egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py create mode 100755 egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py create mode 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/gigaspeech_scoring.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py create mode 100644 egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py create mode 100755 egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py diff --git a/.flake8 b/.flake8 index 190387886..8c497fac3 100644 --- a/.flake8 +++ b/.flake8 @@ -9,6 +9,7 @@ per-file-ignores = egs/tedlium3/ASR/*/conformer.py: E501, egs/gigaspeech/ASR/*/conformer.py: E501, egs/librispeech/ASR/pruned_transducer_stateless2/*.py: E501, + egs/gigaspeech/ASR/pruned_transducer_stateless2/*.py: E501, egs/librispeech/ASR/pruned_transducer_stateless4/*.py: E501, egs/librispeech/ASR/*/optim.py: E501, egs/librispeech/ASR/*/scaling.py: E501, diff --git a/egs/gigaspeech/ASR/README.md b/egs/gigaspeech/ASR/README.md index 7796ef2a0..1fca69e8b 100644 --- a/egs/gigaspeech/ASR/README.md +++ b/egs/gigaspeech/ASR/README.md @@ -13,8 +13,9 @@ ln -sfv /path/to/GigaSpeech download/GigaSpeech ``` ## Performance Record -| | Dev | Test | -|-----|-------|-------| -| WER | 10.47 | 10.58 | +| | Dev | Test | +|--------------------------------|-------|-------| +| `conformer_ctc` | 10.47 | 10.58 | +| `pruned_transducer_stateless2` | 10.52 | 10.62 | See [RESULTS](/egs/gigaspeech/ASR/RESULTS.md) for details. diff --git a/egs/gigaspeech/ASR/RESULTS.md b/egs/gigaspeech/ASR/RESULTS.md index b29e893da..eaed9b9a0 100644 --- a/egs/gigaspeech/ASR/RESULTS.md +++ b/egs/gigaspeech/ASR/RESULTS.md @@ -1,4 +1,78 @@ ## Results +### GigaSpeech BPE training results (Pruned Transducer 2) + +#### 2022-05-12 + +#### Conformer encoder + embedding decoder + +Conformer encoder + non-recurrent decoder. The encoder is a +reworked version of the conformer encoder, with many changes. The +decoder contains only an embedding layer, a Conv1d (with kernel +size 2) and a linear layer (to transform tensor dim). k2 pruned +RNN-T loss is used. + +Results are: + +| | Dev | Test | +|----------------------|-------|-------| +| greedy search | 10.59 | 10.87 | +| fast beam search | 10.56 | 10.80 | +| modified beam search | 10.52 | 10.62 | + +To reproduce the above result, use the following commands for training: + +```bash +cd egs/gigaspeech/ASR +./prepare.sh +export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" +./pruned_transducer_stateless2/train.py \ + --max-duration 120 \ + --num-workers 1 \ + --world-size 8 \ + --exp-dir pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --use-fp16 True +``` + +and the following commands for decoding: + +```bash +# greedy search +./pruned_transducer_stateless2/decode.py \ + --epoch 29 \ + --avg 11 \ + --decoding-method greedy_search \ + --exp-dir pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --max-duration 20 \ + --num-workers 1 + +# fast beam search +./pruned_transducer_stateless2/decode.py \ + --epoch 29 \ + --avg 9 \ + --decoding-method fast_beam_search \ + --exp-dir pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --max-duration 20 \ + --num-workers 1 + +# modified beam search +./pruned_transducer_stateless2/decode.py \ + --epoch 29 \ + --avg 8 \ + --decoding-method modified_beam_search \ + --exp-dir pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --max-duration 20 \ + --num-workers 1 +``` + +Pretrained model is available at + + +The tensorboard log for training is available at + ### GigaSpeech BPE training results (Conformer-CTC) @@ -20,7 +94,7 @@ Scale values used in n-gram LM rescoring and attention rescoring for the best WE To reproduce the above result, use the following commands for training: -``` +```bash cd egs/gigaspeech/ASR ./prepare.sh export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" @@ -34,7 +108,7 @@ export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" and the following command for decoding: -``` +```bash ./conformer_ctc/decode.py \ --epoch 18 \ --avg 6 \ @@ -59,7 +133,7 @@ Scale values used in n-gram LM rescoring and attention rescoring for the best WE To reproduce the above result, use the training commands above, and the following command for decoding: -``` +```bash ./conformer_ctc/decode.py \ --epoch 18 \ --avg 6 \ diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/__init__.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/__init__.py new file mode 100644 index 000000000..e69de29bb diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py new file mode 100644 index 000000000..ff3d3b07a --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py @@ -0,0 +1,416 @@ +# Copyright 2021 Piotr Żelasko +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +import argparse +import inspect +import logging +from functools import lru_cache +from pathlib import Path +from typing import Any, Dict, Optional + +import torch +from lhotse import CutSet, Fbank, FbankConfig, load_manifest +from lhotse.dataset import ( + BucketingSampler, + CutConcatenate, + CutMix, + DynamicBucketingSampler, + K2SpeechRecognitionDataset, + PrecomputedFeatures, + SingleCutSampler, + SpecAugment, +) +from lhotse.dataset.input_strategies import OnTheFlyFeatures +from lhotse.utils import fix_random_seed +from torch.utils.data import DataLoader + +from icefall.utils import str2bool + + +class _SeedWorkers: + def __init__(self, seed: int): + self.seed = seed + + def __call__(self, worker_id: int): + fix_random_seed(self.seed + worker_id) + + +class GigaSpeechAsrDataModule: + """ + DataModule for k2 ASR experiments. + It assumes there is always one train and valid dataloader, + but there can be multiple test dataloaders (e.g. LibriSpeech test-clean + and test-other). + + It contains all the common data pipeline modules used in ASR + experiments, e.g.: + - dynamic batch size, + - bucketing samplers, + - cut concatenation, + - augmentation, + - on-the-fly feature extraction + + This class should be derived for specific corpora used in ASR tasks. + """ + + def __init__(self, args: argparse.Namespace): + self.args = args + + @classmethod + def add_arguments(cls, parser: argparse.ArgumentParser): + group = parser.add_argument_group( + title="ASR data related options", + description="These options are used for the preparation of " + "PyTorch DataLoaders from Lhotse CutSet's -- they control the " + "effective batch sizes, sampling strategies, applied data " + "augmentations, etc.", + ) + group.add_argument( + "--manifest-dir", + type=Path, + default=Path("data/fbank"), + help="Path to directory with train/valid/test cuts.", + ) + group.add_argument( + "--max-duration", + type=int, + default=200.0, + help="Maximum pooled recordings duration (seconds) in a " + "single batch. You can reduce it if it causes CUDA OOM.", + ) + group.add_argument( + "--bucketing-sampler", + type=str2bool, + default=True, + help="When enabled, the batches will come from buckets of " + "similar duration (saves padding frames).", + ) + group.add_argument( + "--num-buckets", + type=int, + default=30, + help="The number of buckets for the DynamicBucketingSampler" + "(you might want to increase it for larger datasets).", + ) + group.add_argument( + "--concatenate-cuts", + type=str2bool, + default=False, + help="When enabled, utterances (cuts) will be concatenated " + "to minimize the amount of padding.", + ) + group.add_argument( + "--duration-factor", + type=float, + default=1.0, + help="Determines the maximum duration of a concatenated cut " + "relative to the duration of the longest cut in a batch.", + ) + group.add_argument( + "--gap", + type=float, + default=1.0, + help="The amount of padding (in seconds) inserted between " + "concatenated cuts. This padding is filled with noise when " + "noise augmentation is used.", + ) + group.add_argument( + "--on-the-fly-feats", + type=str2bool, + default=False, + help="When enabled, use on-the-fly cut mixing and feature " + "extraction. Will drop existing precomputed feature manifests " + "if available.", + ) + group.add_argument( + "--shuffle", + type=str2bool, + default=True, + help="When enabled (=default), the examples will be " + "shuffled for each epoch.", + ) + group.add_argument( + "--return-cuts", + type=str2bool, + default=True, + help="When enabled, each batch will have the " + "field: batch['supervisions']['cut'] with the cuts that " + "were used to construct it.", + ) + + group.add_argument( + "--num-workers", + type=int, + default=2, + help="The number of training dataloader workers that " + "collect the batches.", + ) + + group.add_argument( + "--enable-spec-aug", + type=str2bool, + default=True, + help="When enabled, use SpecAugment for training dataset.", + ) + + group.add_argument( + "--spec-aug-time-warp-factor", + type=int, + default=80, + help="Used only when --enable-spec-aug is True. " + "It specifies the factor for time warping in SpecAugment. " + "Larger values mean more warping. " + "A value less than 1 means to disable time warp.", + ) + + group.add_argument( + "--enable-musan", + type=str2bool, + default=True, + help="When enabled, select noise from MUSAN and mix it " + "with training dataset. ", + ) + + # GigaSpeech specific arguments + group.add_argument( + "--subset", + type=str, + default="XL", + help="Select the GigaSpeech subset (XS|S|M|L|XL)", + ) + group.add_argument( + "--small-dev", + type=str2bool, + default=False, + help="Should we use only 1000 utterances for dev " + "(speeds up training)", + ) + + def train_dataloaders( + self, + cuts_train: CutSet, + sampler_state_dict: Optional[Dict[str, Any]] = None, + ) -> DataLoader: + """ + Args: + cuts_train: + CutSet for training. + sampler_state_dict: + The state dict for the training sampler. + """ + + transforms = [] + if self.args.enable_musan: + logging.info("Enable MUSAN") + logging.info("About to get Musan cuts") + cuts_musan = load_manifest( + self.args.manifest_dir / "cuts_musan.json.gz" + ) + transforms.append( + CutMix( + cuts=cuts_musan, prob=0.5, snr=(10, 20), preserve_id=True + ) + ) + else: + logging.info("Disable MUSAN") + + if self.args.concatenate_cuts: + logging.info( + f"Using cut concatenation with duration factor " + f"{self.args.duration_factor} and gap {self.args.gap}." + ) + # Cut concatenation should be the first transform in the list, + # so that if we e.g. mix noise in, it will fill the gaps between + # different utterances. + transforms = [ + CutConcatenate( + duration_factor=self.args.duration_factor, gap=self.args.gap + ) + ] + transforms + + input_transforms = [] + if self.args.enable_spec_aug: + logging.info("Enable SpecAugment") + logging.info( + f"Time warp factor: {self.args.spec_aug_time_warp_factor}" + ) + # Set the value of num_frame_masks according to Lhotse's version. + # In different Lhotse's versions, the default of num_frame_masks is + # different. + num_frame_masks = 10 + num_frame_masks_parameter = inspect.signature( + SpecAugment.__init__ + ).parameters["num_frame_masks"] + if num_frame_masks_parameter.default == 1: + num_frame_masks = 2 + logging.info(f"Num frame mask: {num_frame_masks}") + input_transforms.append( + SpecAugment( + time_warp_factor=self.args.spec_aug_time_warp_factor, + num_frame_masks=num_frame_masks, + features_mask_size=27, + num_feature_masks=2, + frames_mask_size=100, + ) + ) + else: + logging.info("Disable SpecAugment") + + logging.info("About to create train dataset") + train = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_transforms=input_transforms, + return_cuts=self.args.return_cuts, + ) + + if self.args.on_the_fly_feats: + # NOTE: the PerturbSpeed transform should be added only if we + # remove it from data prep stage. + # Add on-the-fly speed perturbation; since originally it would + # have increased epoch size by 3, we will apply prob 2/3 and use + # 3x more epochs. + # Speed perturbation probably should come first before + # concatenation, but in principle the transforms order doesn't have + # to be strict (e.g. could be randomized) + # transforms = [PerturbSpeed(factors=[0.9, 1.1], p=2/3)] + transforms # noqa + # Drop feats to be on the safe side. + train = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_strategy=OnTheFlyFeatures( + Fbank(FbankConfig(num_mel_bins=80)) + ), + input_transforms=input_transforms, + return_cuts=self.args.return_cuts, + ) + + if self.args.bucketing_sampler: + logging.info("Using DynamicBucketingSampler.") + train_sampler = DynamicBucketingSampler( + cuts_train, + max_duration=self.args.max_duration, + shuffle=self.args.shuffle, + num_buckets=self.args.num_buckets, + drop_last=True, + ) + else: + logging.info("Using SingleCutSampler.") + train_sampler = SingleCutSampler( + cuts_train, + max_duration=self.args.max_duration, + shuffle=self.args.shuffle, + ) + logging.info("About to create train dataloader") + + if sampler_state_dict is not None: + logging.info("Loading sampler state dict") + train_sampler.load_state_dict(sampler_state_dict) + + # 'seed' is derived from the current random state, which will have + # previously been set in the main process. + seed = torch.randint(0, 100000, ()).item() + worker_init_fn = _SeedWorkers(seed) + + train_dl = DataLoader( + train, + sampler=train_sampler, + batch_size=None, + num_workers=self.args.num_workers, + persistent_workers=False, + worker_init_fn=worker_init_fn, + ) + + return train_dl + + def valid_dataloaders(self, cuts_valid: CutSet) -> DataLoader: + transforms = [] + if self.args.concatenate_cuts: + transforms = [ + CutConcatenate( + duration_factor=self.args.duration_factor, gap=self.args.gap + ) + ] + transforms + + logging.info("About to create dev dataset") + if self.args.on_the_fly_feats: + validate = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_strategy=OnTheFlyFeatures( + Fbank(FbankConfig(num_mel_bins=80)) + ), + return_cuts=self.args.return_cuts, + ) + else: + validate = K2SpeechRecognitionDataset( + cut_transforms=transforms, + return_cuts=self.args.return_cuts, + ) + valid_sampler = BucketingSampler( + cuts_valid, + max_duration=self.args.max_duration, + shuffle=False, + ) + logging.info("About to create dev dataloader") + valid_dl = DataLoader( + validate, + sampler=valid_sampler, + batch_size=None, + num_workers=2, + persistent_workers=False, + ) + + return valid_dl + + def test_dataloaders(self, cuts: CutSet) -> DataLoader: + logging.debug("About to create test dataset") + test = K2SpeechRecognitionDataset( + input_strategy=OnTheFlyFeatures(Fbank(FbankConfig(num_mel_bins=80))) + if self.args.on_the_fly_feats + else PrecomputedFeatures(), + return_cuts=self.args.return_cuts, + ) + sampler = BucketingSampler( + cuts, max_duration=self.args.max_duration, shuffle=False + ) + logging.debug("About to create test dataloader") + test_dl = DataLoader( + test, + batch_size=None, + sampler=sampler, + num_workers=self.args.num_workers, + ) + return test_dl + + @lru_cache() + def train_cuts(self) -> CutSet: + logging.info(f"About to get train_{self.args.subset} cuts") + path = self.args.manifest_dir / f"cuts_{self.args.subset}.jsonl.gz" + cuts_train = CutSet.from_jsonl_lazy(path) + return cuts_train + + @lru_cache() + def dev_cuts(self) -> CutSet: + logging.info("About to get dev cuts") + cuts_valid = load_manifest(self.args.manifest_dir / "cuts_DEV.jsonl.gz") + if self.args.small_dev: + return cuts_valid.subset(first=1000) + else: + return cuts_valid + + @lru_cache() + def test_cuts(self) -> CutSet: + logging.info("About to get test cuts") + return load_manifest(self.args.manifest_dir / "cuts_TEST.jsonl.gz") diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py new file mode 100644 index 000000000..2e9bf3e0b --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py @@ -0,0 +1,766 @@ +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from dataclasses import dataclass +from typing import Dict, List, Optional + +import k2 +import torch +from model import Transducer + +from icefall.decode import one_best_decoding +from icefall.utils import get_texts + + +def fast_beam_search( + model: Transducer, + decoding_graph: k2.Fsa, + encoder_out: torch.Tensor, + encoder_out_lens: torch.Tensor, + beam: float, + max_states: int, + max_contexts: int, +) -> List[List[int]]: + """It limits the maximum number of symbols per frame to 1. + + Args: + model: + An instance of `Transducer`. + decoding_graph: + Decoding graph used for decoding, may be a TrivialGraph or a HLG. + encoder_out: + A tensor of shape (N, T, C) from the encoder. + encoder_out_lens: + A tensor of shape (N,) containing the number of frames in `encoder_out` + before padding. + beam: + Beam value, similar to the beam used in Kaldi.. + max_states: + Max states per stream per frame. + max_contexts: + Max contexts pre stream per frame. + Returns: + Return the decoded result. + """ + assert encoder_out.ndim == 3 + + context_size = model.decoder.context_size + vocab_size = model.decoder.vocab_size + + B, T, C = encoder_out.shape + + config = k2.RnntDecodingConfig( + vocab_size=vocab_size, + decoder_history_len=context_size, + beam=beam, + max_contexts=max_contexts, + max_states=max_states, + ) + individual_streams = [] + for i in range(B): + individual_streams.append(k2.RnntDecodingStream(decoding_graph)) + decoding_streams = k2.RnntDecodingStreams(individual_streams, config) + + encoder_out = model.joiner.encoder_proj(encoder_out) + + for t in range(T): + # shape is a RaggedShape of shape (B, context) + # contexts is a Tensor of shape (shape.NumElements(), context_size) + shape, contexts = decoding_streams.get_contexts() + # `nn.Embedding()` in torch below v1.7.1 supports only torch.int64 + contexts = contexts.to(torch.int64) + # decoder_out is of shape (shape.NumElements(), 1, decoder_out_dim) + decoder_out = model.decoder(contexts, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + # current_encoder_out is of shape + # (shape.NumElements(), 1, joiner_dim) + # fmt: off + current_encoder_out = torch.index_select( + encoder_out[:, t:t + 1, :], 0, shape.row_ids(1).to(torch.int64) + ) + # fmt: on + logits = model.joiner( + current_encoder_out.unsqueeze(2), + decoder_out.unsqueeze(1), + project_input=False, + ) + logits = logits.squeeze(1).squeeze(1) + log_probs = logits.log_softmax(dim=-1) + decoding_streams.advance(log_probs) + decoding_streams.terminate_and_flush_to_streams() + lattice = decoding_streams.format_output(encoder_out_lens.tolist()) + + best_path = one_best_decoding(lattice) + hyps = get_texts(best_path) + return hyps + + +def greedy_search( + model: Transducer, encoder_out: torch.Tensor, max_sym_per_frame: int +) -> List[int]: + """Greedy search for a single utterance. + Args: + model: + An instance of `Transducer`. + encoder_out: + A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. + max_sym_per_frame: + Maximum number of symbols per frame. If it is set to 0, the WER + would be 100%. + Returns: + Return the decoded result. + """ + assert encoder_out.ndim == 3 + + # support only batch_size == 1 for now + assert encoder_out.size(0) == 1, encoder_out.size(0) + + blank_id = model.decoder.blank_id + context_size = model.decoder.context_size + + device = model.device + + decoder_input = torch.tensor( + [blank_id] * context_size, device=device, dtype=torch.int64 + ).reshape(1, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + + encoder_out = model.joiner.encoder_proj(encoder_out) + + T = encoder_out.size(1) + t = 0 + hyp = [blank_id] * context_size + + # Maximum symbols per utterance. + max_sym_per_utt = 1000 + + # symbols per frame + sym_per_frame = 0 + + # symbols per utterance decoded so far + sym_per_utt = 0 + + while t < T and sym_per_utt < max_sym_per_utt: + if sym_per_frame >= max_sym_per_frame: + sym_per_frame = 0 + t += 1 + continue + + # fmt: off + current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) + # fmt: on + logits = model.joiner( + current_encoder_out, decoder_out.unsqueeze(1), project_input=False + ) + # logits is (1, 1, 1, vocab_size) + + y = logits.argmax().item() + if y != blank_id: + hyp.append(y) + decoder_input = torch.tensor( + [hyp[-context_size:]], device=device + ).reshape(1, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + + sym_per_utt += 1 + sym_per_frame += 1 + else: + sym_per_frame = 0 + t += 1 + hyp = hyp[context_size:] # remove blanks + + return hyp + + +def greedy_search_batch( + model: Transducer, encoder_out: torch.Tensor +) -> List[List[int]]: + """Greedy search in batch mode. It hardcodes --max-sym-per-frame=1. + Args: + model: + The transducer model. + encoder_out: + Output from the encoder. Its shape is (N, T, C), where N >= 1. + Returns: + Return a list-of-list of token IDs containing the decoded results. + len(ans) equals to encoder_out.size(0). + """ + assert encoder_out.ndim == 3 + assert encoder_out.size(0) >= 1, encoder_out.size(0) + + device = model.device + + batch_size = encoder_out.size(0) + T = encoder_out.size(1) + + blank_id = model.decoder.blank_id + context_size = model.decoder.context_size + + hyps = [[blank_id] * context_size for _ in range(batch_size)] + + decoder_input = torch.tensor( + hyps, + device=device, + dtype=torch.int64, + ) # (batch_size, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + encoder_out = model.joiner.encoder_proj(encoder_out) + + # decoder_out: (batch_size, 1, decoder_out_dim) + for t in range(T): + current_encoder_out = encoder_out[:, t : t + 1, :].unsqueeze(2) # noqa + # current_encoder_out's shape: (batch_size, 1, 1, encoder_out_dim) + logits = model.joiner( + current_encoder_out, decoder_out.unsqueeze(1), project_input=False + ) + # logits'shape (batch_size, 1, 1, vocab_size) + + logits = logits.squeeze(1).squeeze(1) # (batch_size, vocab_size) + assert logits.ndim == 2, logits.shape + y = logits.argmax(dim=1).tolist() + emitted = False + for i, v in enumerate(y): + if v != blank_id: + hyps[i].append(v) + emitted = True + if emitted: + # update decoder output + decoder_input = [h[-context_size:] for h in hyps] + decoder_input = torch.tensor( + decoder_input, + device=device, + dtype=torch.int64, + ) + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + + ans = [h[context_size:] for h in hyps] + return ans + + +@dataclass +class Hypothesis: + # The predicted tokens so far. + # Newly predicted tokens are appended to `ys`. + ys: List[int] + + # The log prob of ys. + # It contains only one entry. + log_prob: torch.Tensor + + @property + def key(self) -> str: + """Return a string representation of self.ys""" + return "_".join(map(str, self.ys)) + + +class HypothesisList(object): + def __init__(self, data: Optional[Dict[str, Hypothesis]] = None) -> None: + """ + Args: + data: + A dict of Hypotheses. Its key is its `value.key`. + """ + if data is None: + self._data = {} + else: + self._data = data + + @property + def data(self) -> Dict[str, Hypothesis]: + return self._data + + def add(self, hyp: Hypothesis) -> None: + """Add a Hypothesis to `self`. + + If `hyp` already exists in `self`, its probability is updated using + `log-sum-exp` with the existed one. + + Args: + hyp: + The hypothesis to be added. + """ + key = hyp.key + if key in self: + old_hyp = self._data[key] # shallow copy + torch.logaddexp( + old_hyp.log_prob, hyp.log_prob, out=old_hyp.log_prob + ) + else: + self._data[key] = hyp + + def get_most_probable(self, length_norm: bool = False) -> Hypothesis: + """Get the most probable hypothesis, i.e., the one with + the largest `log_prob`. + + Args: + length_norm: + If True, the `log_prob` of a hypothesis is normalized by the + number of tokens in it. + Returns: + Return the hypothesis that has the largest `log_prob`. + """ + if length_norm: + return max( + self._data.values(), key=lambda hyp: hyp.log_prob / len(hyp.ys) + ) + else: + return max(self._data.values(), key=lambda hyp: hyp.log_prob) + + def remove(self, hyp: Hypothesis) -> None: + """Remove a given hypothesis. + + Caution: + `self` is modified **in-place**. + + Args: + hyp: + The hypothesis to be removed from `self`. + Note: It must be contained in `self`. Otherwise, + an exception is raised. + """ + key = hyp.key + assert key in self, f"{key} does not exist" + del self._data[key] + + def filter(self, threshold: torch.Tensor) -> "HypothesisList": + """Remove all Hypotheses whose log_prob is less than threshold. + + Caution: + `self` is not modified. Instead, a new HypothesisList is returned. + + Returns: + Return a new HypothesisList containing all hypotheses from `self` + with `log_prob` being greater than the given `threshold`. + """ + ans = HypothesisList() + for _, hyp in self._data.items(): + if hyp.log_prob > threshold: + ans.add(hyp) # shallow copy + return ans + + def topk(self, k: int) -> "HypothesisList": + """Return the top-k hypothesis.""" + hyps = list(self._data.items()) + + hyps = sorted(hyps, key=lambda h: h[1].log_prob, reverse=True)[:k] + + ans = HypothesisList(dict(hyps)) + return ans + + def __contains__(self, key: str): + return key in self._data + + def __iter__(self): + return iter(self._data.values()) + + def __len__(self) -> int: + return len(self._data) + + def __str__(self) -> str: + s = [] + for key in self: + s.append(key) + return ", ".join(s) + + +def _get_hyps_shape(hyps: List[HypothesisList]) -> k2.RaggedShape: + """Return a ragged shape with axes [utt][num_hyps]. + + Args: + hyps: + len(hyps) == batch_size. It contains the current hypothesis for + each utterance in the batch. + Returns: + Return a ragged shape with 2 axes [utt][num_hyps]. Note that + the shape is on CPU. + """ + num_hyps = [len(h) for h in hyps] + + # torch.cumsum() is inclusive sum, so we put a 0 at the beginning + # to get exclusive sum later. + num_hyps.insert(0, 0) + + num_hyps = torch.tensor(num_hyps) + row_splits = torch.cumsum(num_hyps, dim=0, dtype=torch.int32) + ans = k2.ragged.create_ragged_shape2( + row_splits=row_splits, cached_tot_size=row_splits[-1].item() + ) + return ans + + +def modified_beam_search( + model: Transducer, + encoder_out: torch.Tensor, + beam: int = 4, +) -> List[List[int]]: + """Beam search in batch mode with --max-sym-per-frame=1 being hardcoded. + + Args: + model: + The transducer model. + encoder_out: + Output from the encoder. Its shape is (N, T, C). + beam: + Number of active paths during the beam search. + Returns: + Return a list-of-list of token IDs. ans[i] is the decoding results + for the i-th utterance. + """ + assert encoder_out.ndim == 3, encoder_out.shape + + batch_size = encoder_out.size(0) + T = encoder_out.size(1) + + blank_id = model.decoder.blank_id + context_size = model.decoder.context_size + device = model.device + B = [HypothesisList() for _ in range(batch_size)] + for i in range(batch_size): + B[i].add( + Hypothesis( + ys=[blank_id] * context_size, + log_prob=torch.zeros(1, dtype=torch.float32, device=device), + ) + ) + + encoder_out = model.joiner.encoder_proj(encoder_out) + + for t in range(T): + current_encoder_out = encoder_out[:, t : t + 1, :].unsqueeze(2) # noqa + # current_encoder_out's shape is (batch_size, 1, 1, encoder_out_dim) + + hyps_shape = _get_hyps_shape(B).to(device) + + A = [list(b) for b in B] + B = [HypothesisList() for _ in range(batch_size)] + + ys_log_probs = torch.cat( + [hyp.log_prob.reshape(1, 1) for hyps in A for hyp in hyps] + ) # (num_hyps, 1) + + decoder_input = torch.tensor( + [hyp.ys[-context_size:] for hyps in A for hyp in hyps], + device=device, + dtype=torch.int64, + ) # (num_hyps, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False).unsqueeze(1) + decoder_out = model.joiner.decoder_proj(decoder_out) + # decoder_out is of shape (num_hyps, 1, 1, joiner_dim) + + # Note: For torch 1.7.1 and below, it requires a torch.int64 tensor + # as index, so we use `to(torch.int64)` below. + current_encoder_out = torch.index_select( + current_encoder_out, + dim=0, + index=hyps_shape.row_ids(1).to(torch.int64), + ) # (num_hyps, 1, 1, encoder_out_dim) + + logits = model.joiner( + current_encoder_out, + decoder_out, + project_input=False, + ) # (num_hyps, 1, 1, vocab_size) + + logits = logits.squeeze(1).squeeze(1) # (num_hyps, vocab_size) + + log_probs = logits.log_softmax(dim=-1) # (num_hyps, vocab_size) + + log_probs.add_(ys_log_probs) + + vocab_size = log_probs.size(-1) + + log_probs = log_probs.reshape(-1) + + row_splits = hyps_shape.row_splits(1) * vocab_size + log_probs_shape = k2.ragged.create_ragged_shape2( + row_splits=row_splits, cached_tot_size=log_probs.numel() + ) + ragged_log_probs = k2.RaggedTensor( + shape=log_probs_shape, value=log_probs + ) + + for i in range(batch_size): + topk_log_probs, topk_indexes = ragged_log_probs[i].topk(beam) + + topk_hyp_indexes = (topk_indexes // vocab_size).tolist() + topk_token_indexes = (topk_indexes % vocab_size).tolist() + + for k in range(len(topk_hyp_indexes)): + hyp_idx = topk_hyp_indexes[k] + hyp = A[i][hyp_idx] + + new_ys = hyp.ys[:] + new_token = topk_token_indexes[k] + if new_token != blank_id: + new_ys.append(new_token) + + new_log_prob = topk_log_probs[k] + new_hyp = Hypothesis(ys=new_ys, log_prob=new_log_prob) + B[i].add(new_hyp) + + best_hyps = [b.get_most_probable(length_norm=True) for b in B] + ans = [h.ys[context_size:] for h in best_hyps] + + return ans + + +def _deprecated_modified_beam_search( + model: Transducer, + encoder_out: torch.Tensor, + beam: int = 4, +) -> List[int]: + """It limits the maximum number of symbols per frame to 1. + + It decodes only one utterance at a time. We keep it only for reference. + The function :func:`modified_beam_search` should be preferred as it + supports batch decoding. + + + Args: + model: + An instance of `Transducer`. + encoder_out: + A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. + beam: + Beam size. + Returns: + Return the decoded result. + """ + + assert encoder_out.ndim == 3 + + # support only batch_size == 1 for now + assert encoder_out.size(0) == 1, encoder_out.size(0) + blank_id = model.decoder.blank_id + context_size = model.decoder.context_size + + device = model.device + + T = encoder_out.size(1) + + B = HypothesisList() + B.add( + Hypothesis( + ys=[blank_id] * context_size, + log_prob=torch.zeros(1, dtype=torch.float32, device=device), + ) + ) + encoder_out = model.joiner.encoder_proj(encoder_out) + + for t in range(T): + # fmt: off + current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) + # current_encoder_out is of shape (1, 1, 1, encoder_out_dim) + # fmt: on + A = list(B) + B = HypothesisList() + + ys_log_probs = torch.cat([hyp.log_prob.reshape(1, 1) for hyp in A]) + # ys_log_probs is of shape (num_hyps, 1) + + decoder_input = torch.tensor( + [hyp.ys[-context_size:] for hyp in A], + device=device, + dtype=torch.int64, + ) + # decoder_input is of shape (num_hyps, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False).unsqueeze(1) + decoder_out = model.joiner.decoder_proj(decoder_out) + # decoder_output is of shape (num_hyps, 1, 1, joiner_dim) + + current_encoder_out = current_encoder_out.expand( + decoder_out.size(0), 1, 1, -1 + ) # (num_hyps, 1, 1, encoder_out_dim) + + logits = model.joiner( + current_encoder_out, + decoder_out, + project_input=False, + ) + # logits is of shape (num_hyps, 1, 1, vocab_size) + logits = logits.squeeze(1).squeeze(1) + + # now logits is of shape (num_hyps, vocab_size) + log_probs = logits.log_softmax(dim=-1) + + log_probs.add_(ys_log_probs) + + log_probs = log_probs.reshape(-1) + topk_log_probs, topk_indexes = log_probs.topk(beam) + + # topk_hyp_indexes are indexes into `A` + topk_hyp_indexes = topk_indexes // logits.size(-1) + topk_token_indexes = topk_indexes % logits.size(-1) + + topk_hyp_indexes = topk_hyp_indexes.tolist() + topk_token_indexes = topk_token_indexes.tolist() + + for i in range(len(topk_hyp_indexes)): + hyp = A[topk_hyp_indexes[i]] + new_ys = hyp.ys[:] + new_token = topk_token_indexes[i] + if new_token != blank_id: + new_ys.append(new_token) + new_log_prob = topk_log_probs[i] + new_hyp = Hypothesis(ys=new_ys, log_prob=new_log_prob) + B.add(new_hyp) + + best_hyp = B.get_most_probable(length_norm=True) + ys = best_hyp.ys[context_size:] # [context_size:] to remove blanks + + return ys + + +def beam_search( + model: Transducer, + encoder_out: torch.Tensor, + beam: int = 4, +) -> List[int]: + """ + It implements Algorithm 1 in https://arxiv.org/pdf/1211.3711.pdf + + espnet/nets/beam_search_transducer.py#L247 is used as a reference. + + Args: + model: + An instance of `Transducer`. + encoder_out: + A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. + beam: + Beam size. + Returns: + Return the decoded result. + """ + assert encoder_out.ndim == 3 + + # support only batch_size == 1 for now + assert encoder_out.size(0) == 1, encoder_out.size(0) + blank_id = model.decoder.blank_id + context_size = model.decoder.context_size + + device = model.device + + decoder_input = torch.tensor( + [blank_id] * context_size, + device=device, + dtype=torch.int64, + ).reshape(1, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + + encoder_out = model.joiner.encoder_proj(encoder_out) + + T = encoder_out.size(1) + t = 0 + + B = HypothesisList() + B.add(Hypothesis(ys=[blank_id] * context_size, log_prob=0.0)) + + max_sym_per_utt = 20000 + + sym_per_utt = 0 + + decoder_cache: Dict[str, torch.Tensor] = {} + + while t < T and sym_per_utt < max_sym_per_utt: + # fmt: off + current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) + # fmt: on + A = B + B = HypothesisList() + + joint_cache: Dict[str, torch.Tensor] = {} + + # TODO(fangjun): Implement prefix search to update the `log_prob` + # of hypotheses in A + + while True: + y_star = A.get_most_probable() + A.remove(y_star) + + cached_key = y_star.key + + if cached_key not in decoder_cache: + decoder_input = torch.tensor( + [y_star.ys[-context_size:]], + device=device, + dtype=torch.int64, + ).reshape(1, context_size) + + decoder_out = model.decoder(decoder_input, need_pad=False) + decoder_out = model.joiner.decoder_proj(decoder_out) + decoder_cache[cached_key] = decoder_out + else: + decoder_out = decoder_cache[cached_key] + + cached_key += f"-t-{t}" + if cached_key not in joint_cache: + logits = model.joiner( + current_encoder_out, + decoder_out.unsqueeze(1), + project_input=False, + ) + + # TODO(fangjun): Scale the blank posterior + log_prob = logits.log_softmax(dim=-1) + # log_prob is (1, 1, 1, vocab_size) + log_prob = log_prob.squeeze() + # Now log_prob is (vocab_size,) + joint_cache[cached_key] = log_prob + else: + log_prob = joint_cache[cached_key] + + # First, process the blank symbol + skip_log_prob = log_prob[blank_id] + new_y_star_log_prob = y_star.log_prob + skip_log_prob + + # ys[:] returns a copy of ys + B.add(Hypothesis(ys=y_star.ys[:], log_prob=new_y_star_log_prob)) + + # Second, process other non-blank labels + values, indices = log_prob.topk(beam + 1) + for i, v in zip(indices.tolist(), values.tolist()): + if i == blank_id: + continue + new_ys = y_star.ys + [i] + new_log_prob = y_star.log_prob + v + A.add(Hypothesis(ys=new_ys, log_prob=new_log_prob)) + + # Check whether B contains more than "beam" elements more probable + # than the most probable in A + A_most_probable = A.get_most_probable() + + kept_B = B.filter(A_most_probable.log_prob) + + if len(kept_B) >= beam: + B = kept_B.topk(beam) + break + + t += 1 + + best_hyp = B.get_most_probable(length_norm=True) + ys = best_hyp.ys[context_size:] # [context_size:] to remove blanks + return ys diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py new file mode 100644 index 000000000..257936b59 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py @@ -0,0 +1,1038 @@ +#!/usr/bin/env python3 +# Copyright (c) 2021 University of Chinese Academy of Sciences (author: Han Zhu) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import copy +import math +import warnings +from typing import Optional, Tuple + +import torch +from encoder_interface import EncoderInterface +from scaling import ( + ActivationBalancer, + BasicNorm, + DoubleSwish, + ScaledConv1d, + ScaledConv2d, + ScaledLinear, +) +from torch import Tensor, nn + +from icefall.utils import make_pad_mask + + +class Conformer(EncoderInterface): + """ + Args: + num_features (int): Number of input features + subsampling_factor (int): subsampling factor of encoder (the convolution layers before transformers) + d_model (int): attention dimension, also the output dimension + nhead (int): number of head + dim_feedforward (int): feedforward dimention + num_encoder_layers (int): number of encoder layers + dropout (float): dropout rate + layer_dropout (float): layer-dropout rate. + cnn_module_kernel (int): Kernel size of convolution module + vgg_frontend (bool): whether to use vgg frontend. + """ + + def __init__( + self, + num_features: int, + subsampling_factor: int = 4, + d_model: int = 256, + nhead: int = 4, + dim_feedforward: int = 2048, + num_encoder_layers: int = 12, + dropout: float = 0.1, + layer_dropout: float = 0.075, + cnn_module_kernel: int = 31, + ) -> None: + super(Conformer, self).__init__() + + self.num_features = num_features + self.subsampling_factor = subsampling_factor + if subsampling_factor != 4: + raise NotImplementedError("Support only 'subsampling_factor=4'.") + + # self.encoder_embed converts the input of shape (N, T, num_features) + # to the shape (N, T//subsampling_factor, d_model). + # That is, it does two things simultaneously: + # (1) subsampling: T -> T//subsampling_factor + # (2) embedding: num_features -> d_model + self.encoder_embed = Conv2dSubsampling(num_features, d_model) + + self.encoder_pos = RelPositionalEncoding(d_model, dropout) + + encoder_layer = ConformerEncoderLayer( + d_model, + nhead, + dim_feedforward, + dropout, + layer_dropout, + cnn_module_kernel, + ) + self.encoder = ConformerEncoder(encoder_layer, num_encoder_layers) + + def forward( + self, x: torch.Tensor, x_lens: torch.Tensor, warmup: float = 1.0 + ) -> Tuple[torch.Tensor, torch.Tensor]: + """ + Args: + x: + The input tensor. Its shape is (batch_size, seq_len, feature_dim). + x_lens: + A tensor of shape (batch_size,) containing the number of frames in + `x` before padding. + warmup: + A floating point value that gradually increases from 0 throughout + training; when it is >= 1.0 we are "fully warmed up". It is used + to turn modules on sequentially. + Returns: + Return a tuple containing 2 tensors: + - embeddings: its shape is (batch_size, output_seq_len, d_model) + - lengths, a tensor of shape (batch_size,) containing the number + of frames in `embeddings` before padding. + """ + x = self.encoder_embed(x) + x, pos_emb = self.encoder_pos(x) + x = x.permute(1, 0, 2) # (N, T, C) -> (T, N, C) + + with warnings.catch_warnings(): + warnings.simplefilter("ignore") + # Caution: We assume the subsampling factor is 4! + lengths = ((x_lens - 1) // 2 - 1) // 2 + assert x.size(0) == lengths.max().item() + mask = make_pad_mask(lengths) + + x = self.encoder( + x, pos_emb, src_key_padding_mask=mask, warmup=warmup + ) # (T, N, C) + + x = x.permute(1, 0, 2) # (T, N, C) ->(N, T, C) + + return x, lengths + + +class ConformerEncoderLayer(nn.Module): + """ + ConformerEncoderLayer is made up of self-attn, feedforward and convolution networks. + See: "Conformer: Convolution-augmented Transformer for Speech Recognition" + + Args: + d_model: the number of expected features in the input (required). + nhead: the number of heads in the multiheadattention models (required). + dim_feedforward: the dimension of the feedforward network model (default=2048). + dropout: the dropout value (default=0.1). + cnn_module_kernel (int): Kernel size of convolution module. + + Examples:: + >>> encoder_layer = ConformerEncoderLayer(d_model=512, nhead=8) + >>> src = torch.rand(10, 32, 512) + >>> pos_emb = torch.rand(32, 19, 512) + >>> out = encoder_layer(src, pos_emb) + """ + + def __init__( + self, + d_model: int, + nhead: int, + dim_feedforward: int = 2048, + dropout: float = 0.1, + layer_dropout: float = 0.075, + cnn_module_kernel: int = 31, + ) -> None: + super(ConformerEncoderLayer, self).__init__() + + self.layer_dropout = layer_dropout + + self.d_model = d_model + + self.self_attn = RelPositionMultiheadAttention( + d_model, nhead, dropout=0.0 + ) + + self.feed_forward = nn.Sequential( + ScaledLinear(d_model, dim_feedforward), + ActivationBalancer(channel_dim=-1), + DoubleSwish(), + nn.Dropout(dropout), + ScaledLinear(dim_feedforward, d_model, initial_scale=0.25), + ) + + self.feed_forward_macaron = nn.Sequential( + ScaledLinear(d_model, dim_feedforward), + ActivationBalancer(channel_dim=-1), + DoubleSwish(), + nn.Dropout(dropout), + ScaledLinear(dim_feedforward, d_model, initial_scale=0.25), + ) + + self.conv_module = ConvolutionModule(d_model, cnn_module_kernel) + + self.norm_final = BasicNorm(d_model) + + # try to ensure the output is close to zero-mean (or at least, zero-median). + self.balancer = ActivationBalancer( + channel_dim=-1, min_positive=0.45, max_positive=0.55, max_abs=6.0 + ) + + self.dropout = nn.Dropout(dropout) + + def forward( + self, + src: Tensor, + pos_emb: Tensor, + src_mask: Optional[Tensor] = None, + src_key_padding_mask: Optional[Tensor] = None, + warmup: float = 1.0, + ) -> Tensor: + """ + Pass the input through the encoder layer. + + Args: + src: the sequence to the encoder layer (required). + pos_emb: Positional embedding tensor (required). + src_mask: the mask for the src sequence (optional). + src_key_padding_mask: the mask for the src keys per batch (optional). + warmup: controls selective bypass of of layers; if < 1.0, we will + bypass layers more frequently. + + Shape: + src: (S, N, E). + pos_emb: (N, 2*S-1, E) + src_mask: (S, S). + src_key_padding_mask: (N, S). + S is the source sequence length, N is the batch size, E is the feature number + """ + src_orig = src + + warmup_scale = min(0.1 + warmup, 1.0) + # alpha = 1.0 means fully use this encoder layer, 0.0 would mean + # completely bypass it. + if self.training: + alpha = ( + warmup_scale + if torch.rand(()).item() <= (1.0 - self.layer_dropout) + else 0.1 + ) + else: + alpha = 1.0 + + # macaron style feed forward module + src = src + self.dropout(self.feed_forward_macaron(src)) + + # multi-headed self-attention module + src_att = self.self_attn( + src, + src, + src, + pos_emb=pos_emb, + attn_mask=src_mask, + key_padding_mask=src_key_padding_mask, + )[0] + src = src + self.dropout(src_att) + + # convolution module + src = src + self.dropout(self.conv_module(src)) + + # feed forward module + src = src + self.dropout(self.feed_forward(src)) + + src = self.norm_final(self.balancer(src)) + + if alpha != 1.0: + src = alpha * src + (1 - alpha) * src_orig + + return src + + +class ConformerEncoder(nn.Module): + r"""ConformerEncoder is a stack of N encoder layers + + Args: + encoder_layer: an instance of the ConformerEncoderLayer() class (required). + num_layers: the number of sub-encoder-layers in the encoder (required). + + Examples:: + >>> encoder_layer = ConformerEncoderLayer(d_model=512, nhead=8) + >>> conformer_encoder = ConformerEncoder(encoder_layer, num_layers=6) + >>> src = torch.rand(10, 32, 512) + >>> pos_emb = torch.rand(32, 19, 512) + >>> out = conformer_encoder(src, pos_emb) + """ + + def __init__(self, encoder_layer: nn.Module, num_layers: int) -> None: + super().__init__() + self.layers = nn.ModuleList( + [copy.deepcopy(encoder_layer) for i in range(num_layers)] + ) + self.num_layers = num_layers + + def forward( + self, + src: Tensor, + pos_emb: Tensor, + mask: Optional[Tensor] = None, + src_key_padding_mask: Optional[Tensor] = None, + warmup: float = 1.0, + ) -> Tensor: + r"""Pass the input through the encoder layers in turn. + + Args: + src: the sequence to the encoder (required). + pos_emb: Positional embedding tensor (required). + mask: the mask for the src sequence (optional). + src_key_padding_mask: the mask for the src keys per batch (optional). + + Shape: + src: (S, N, E). + pos_emb: (N, 2*S-1, E) + mask: (S, S). + src_key_padding_mask: (N, S). + S is the source sequence length, T is the target sequence length, N is the batch size, E is the feature number + + """ + output = src + + for i, mod in enumerate(self.layers): + output = mod( + output, + pos_emb, + src_mask=mask, + src_key_padding_mask=src_key_padding_mask, + warmup=warmup, + ) + + return output + + +class RelPositionalEncoding(torch.nn.Module): + """Relative positional encoding module. + + See : Appendix B in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" + Modified from https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/transformer/embedding.py + + Args: + d_model: Embedding dimension. + dropout_rate: Dropout rate. + max_len: Maximum input length. + + """ + + def __init__( + self, d_model: int, dropout_rate: float, max_len: int = 5000 + ) -> None: + """Construct an PositionalEncoding object.""" + super(RelPositionalEncoding, self).__init__() + self.d_model = d_model + self.dropout = torch.nn.Dropout(p=dropout_rate) + self.pe = None + self.extend_pe(torch.tensor(0.0).expand(1, max_len)) + + def extend_pe(self, x: Tensor) -> None: + """Reset the positional encodings.""" + if self.pe is not None: + # self.pe contains both positive and negative parts + # the length of self.pe is 2 * input_len - 1 + if self.pe.size(1) >= x.size(1) * 2 - 1: + # Note: TorchScript doesn't implement operator== for torch.Device + if self.pe.dtype != x.dtype or str(self.pe.device) != str( + x.device + ): + self.pe = self.pe.to(dtype=x.dtype, device=x.device) + return + # Suppose `i` means to the position of query vecotr and `j` means the + # position of key vector. We use position relative positions when keys + # are to the left (i>j) and negative relative positions otherwise (i Tuple[Tensor, Tensor]: + """Add positional encoding. + + Args: + x (torch.Tensor): Input tensor (batch, time, `*`). + + Returns: + torch.Tensor: Encoded tensor (batch, time, `*`). + torch.Tensor: Encoded tensor (batch, 2*time-1, `*`). + + """ + self.extend_pe(x) + pos_emb = self.pe[ + :, + self.pe.size(1) // 2 + - x.size(1) + + 1 : self.pe.size(1) // 2 # noqa E203 + + x.size(1), + ] + return self.dropout(x), self.dropout(pos_emb) + + +class RelPositionMultiheadAttention(nn.Module): + r"""Multi-Head Attention layer with relative position encoding + + See reference: "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" + + Args: + embed_dim: total dimension of the model. + num_heads: parallel attention heads. + dropout: a Dropout layer on attn_output_weights. Default: 0.0. + + Examples:: + + >>> rel_pos_multihead_attn = RelPositionMultiheadAttention(embed_dim, num_heads) + >>> attn_output, attn_output_weights = multihead_attn(query, key, value, pos_emb) + """ + + def __init__( + self, + embed_dim: int, + num_heads: int, + dropout: float = 0.0, + ) -> None: + super(RelPositionMultiheadAttention, self).__init__() + self.embed_dim = embed_dim + self.num_heads = num_heads + self.dropout = dropout + self.head_dim = embed_dim // num_heads + assert ( + self.head_dim * num_heads == self.embed_dim + ), "embed_dim must be divisible by num_heads" + + self.in_proj = ScaledLinear(embed_dim, 3 * embed_dim, bias=True) + self.out_proj = ScaledLinear( + embed_dim, embed_dim, bias=True, initial_scale=0.25 + ) + + # linear transformation for positional encoding. + self.linear_pos = ScaledLinear(embed_dim, embed_dim, bias=False) + # these two learnable bias are used in matrix c and matrix d + # as described in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" Section 3.3 + self.pos_bias_u = nn.Parameter(torch.Tensor(num_heads, self.head_dim)) + self.pos_bias_v = nn.Parameter(torch.Tensor(num_heads, self.head_dim)) + self.pos_bias_u_scale = nn.Parameter(torch.zeros(()).detach()) + self.pos_bias_v_scale = nn.Parameter(torch.zeros(()).detach()) + self._reset_parameters() + + def _pos_bias_u(self): + return self.pos_bias_u * self.pos_bias_u_scale.exp() + + def _pos_bias_v(self): + return self.pos_bias_v * self.pos_bias_v_scale.exp() + + def _reset_parameters(self) -> None: + nn.init.normal_(self.pos_bias_u, std=0.01) + nn.init.normal_(self.pos_bias_v, std=0.01) + + def forward( + self, + query: Tensor, + key: Tensor, + value: Tensor, + pos_emb: Tensor, + key_padding_mask: Optional[Tensor] = None, + need_weights: bool = True, + attn_mask: Optional[Tensor] = None, + ) -> Tuple[Tensor, Optional[Tensor]]: + r""" + Args: + query, key, value: map a query and a set of key-value pairs to an output. + pos_emb: Positional embedding tensor + key_padding_mask: if provided, specified padding elements in the key will + be ignored by the attention. When given a binary mask and a value is True, + the corresponding value on the attention layer will be ignored. When given + a byte mask and a value is non-zero, the corresponding value on the attention + layer will be ignored + need_weights: output attn_output_weights. + attn_mask: 2D or 3D mask that prevents attention to certain positions. A 2D mask will be broadcasted for all + the batches while a 3D mask allows to specify a different mask for the entries of each batch. + + Shape: + - Inputs: + - query: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, E is + the embedding dimension. + - key: :math:`(S, N, E)`, where S is the source sequence length, N is the batch size, E is + the embedding dimension. + - value: :math:`(S, N, E)` where S is the source sequence length, N is the batch size, E is + the embedding dimension. + - pos_emb: :math:`(N, 2*L-1, E)` where L is the target sequence length, N is the batch size, E is + the embedding dimension. + - key_padding_mask: :math:`(N, S)` where N is the batch size, S is the source sequence length. + If a ByteTensor is provided, the non-zero positions will be ignored while the position + with the zero positions will be unchanged. If a BoolTensor is provided, the positions with the + value of ``True`` will be ignored while the position with the value of ``False`` will be unchanged. + - attn_mask: 2D mask :math:`(L, S)` where L is the target sequence length, S is the source sequence length. + 3D mask :math:`(N*num_heads, L, S)` where N is the batch size, L is the target sequence length, + S is the source sequence length. attn_mask ensure that position i is allowed to attend the unmasked + positions. If a ByteTensor is provided, the non-zero positions are not allowed to attend + while the zero positions will be unchanged. If a BoolTensor is provided, positions with ``True`` + is not allowed to attend while ``False`` values will be unchanged. If a FloatTensor + is provided, it will be added to the attention weight. + + - Outputs: + - attn_output: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, + E is the embedding dimension. + - attn_output_weights: :math:`(N, L, S)` where N is the batch size, + L is the target sequence length, S is the source sequence length. + """ + return self.multi_head_attention_forward( + query, + key, + value, + pos_emb, + self.embed_dim, + self.num_heads, + self.in_proj.get_weight(), + self.in_proj.get_bias(), + self.dropout, + self.out_proj.get_weight(), + self.out_proj.get_bias(), + training=self.training, + key_padding_mask=key_padding_mask, + need_weights=need_weights, + attn_mask=attn_mask, + ) + + def rel_shift(self, x: Tensor) -> Tensor: + """Compute relative positional encoding. + + Args: + x: Input tensor (batch, head, time1, 2*time1-1). + time1 means the length of query vector. + + Returns: + Tensor: tensor of shape (batch, head, time1, time2) + (note: time2 has the same value as time1, but it is for + the key, while time1 is for the query). + """ + (batch_size, num_heads, time1, n) = x.shape + assert n == 2 * time1 - 1 + # Note: TorchScript requires explicit arg for stride() + batch_stride = x.stride(0) + head_stride = x.stride(1) + time1_stride = x.stride(2) + n_stride = x.stride(3) + return x.as_strided( + (batch_size, num_heads, time1, time1), + (batch_stride, head_stride, time1_stride - n_stride, n_stride), + storage_offset=n_stride * (time1 - 1), + ) + + def multi_head_attention_forward( + self, + query: Tensor, + key: Tensor, + value: Tensor, + pos_emb: Tensor, + embed_dim_to_check: int, + num_heads: int, + in_proj_weight: Tensor, + in_proj_bias: Tensor, + dropout_p: float, + out_proj_weight: Tensor, + out_proj_bias: Tensor, + training: bool = True, + key_padding_mask: Optional[Tensor] = None, + need_weights: bool = True, + attn_mask: Optional[Tensor] = None, + ) -> Tuple[Tensor, Optional[Tensor]]: + r""" + Args: + query, key, value: map a query and a set of key-value pairs to an output. + pos_emb: Positional embedding tensor + embed_dim_to_check: total dimension of the model. + num_heads: parallel attention heads. + in_proj_weight, in_proj_bias: input projection weight and bias. + dropout_p: probability of an element to be zeroed. + out_proj_weight, out_proj_bias: the output projection weight and bias. + training: apply dropout if is ``True``. + key_padding_mask: if provided, specified padding elements in the key will + be ignored by the attention. This is an binary mask. When the value is True, + the corresponding value on the attention layer will be filled with -inf. + need_weights: output attn_output_weights. + attn_mask: 2D or 3D mask that prevents attention to certain positions. A 2D mask will be broadcasted for all + the batches while a 3D mask allows to specify a different mask for the entries of each batch. + + Shape: + Inputs: + - query: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, E is + the embedding dimension. + - key: :math:`(S, N, E)`, where S is the source sequence length, N is the batch size, E is + the embedding dimension. + - value: :math:`(S, N, E)` where S is the source sequence length, N is the batch size, E is + the embedding dimension. + - pos_emb: :math:`(N, 2*L-1, E)` or :math:`(1, 2*L-1, E)` where L is the target sequence + length, N is the batch size, E is the embedding dimension. + - key_padding_mask: :math:`(N, S)` where N is the batch size, S is the source sequence length. + If a ByteTensor is provided, the non-zero positions will be ignored while the zero positions + will be unchanged. If a BoolTensor is provided, the positions with the + value of ``True`` will be ignored while the position with the value of ``False`` will be unchanged. + - attn_mask: 2D mask :math:`(L, S)` where L is the target sequence length, S is the source sequence length. + 3D mask :math:`(N*num_heads, L, S)` where N is the batch size, L is the target sequence length, + S is the source sequence length. attn_mask ensures that position i is allowed to attend the unmasked + positions. If a ByteTensor is provided, the non-zero positions are not allowed to attend + while the zero positions will be unchanged. If a BoolTensor is provided, positions with ``True`` + are not allowed to attend while ``False`` values will be unchanged. If a FloatTensor + is provided, it will be added to the attention weight. + + Outputs: + - attn_output: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, + E is the embedding dimension. + - attn_output_weights: :math:`(N, L, S)` where N is the batch size, + L is the target sequence length, S is the source sequence length. + """ + + tgt_len, bsz, embed_dim = query.size() + assert embed_dim == embed_dim_to_check + assert key.size(0) == value.size(0) and key.size(1) == value.size(1) + + head_dim = embed_dim // num_heads + assert ( + head_dim * num_heads == embed_dim + ), "embed_dim must be divisible by num_heads" + + scaling = float(head_dim) ** -0.5 + + if torch.equal(query, key) and torch.equal(key, value): + # self-attention + q, k, v = nn.functional.linear( + query, in_proj_weight, in_proj_bias + ).chunk(3, dim=-1) + + elif torch.equal(key, value): + # encoder-decoder attention + # This is inline in_proj function with in_proj_weight and in_proj_bias + _b = in_proj_bias + _start = 0 + _end = embed_dim + _w = in_proj_weight[_start:_end, :] + if _b is not None: + _b = _b[_start:_end] + q = nn.functional.linear(query, _w, _b) + + # This is inline in_proj function with in_proj_weight and in_proj_bias + _b = in_proj_bias + _start = embed_dim + _end = None + _w = in_proj_weight[_start:, :] + if _b is not None: + _b = _b[_start:] + k, v = nn.functional.linear(key, _w, _b).chunk(2, dim=-1) + + else: + # This is inline in_proj function with in_proj_weight and in_proj_bias + _b = in_proj_bias + _start = 0 + _end = embed_dim + _w = in_proj_weight[_start:_end, :] + if _b is not None: + _b = _b[_start:_end] + q = nn.functional.linear(query, _w, _b) + + # This is inline in_proj function with in_proj_weight and in_proj_bias + _b = in_proj_bias + _start = embed_dim + _end = embed_dim * 2 + _w = in_proj_weight[_start:_end, :] + if _b is not None: + _b = _b[_start:_end] + k = nn.functional.linear(key, _w, _b) + + # This is inline in_proj function with in_proj_weight and in_proj_bias + _b = in_proj_bias + _start = embed_dim * 2 + _end = None + _w = in_proj_weight[_start:, :] + if _b is not None: + _b = _b[_start:] + v = nn.functional.linear(value, _w, _b) + + if attn_mask is not None: + assert ( + attn_mask.dtype == torch.float32 + or attn_mask.dtype == torch.float64 + or attn_mask.dtype == torch.float16 + or attn_mask.dtype == torch.uint8 + or attn_mask.dtype == torch.bool + ), "Only float, byte, and bool types are supported for attn_mask, not {}".format( + attn_mask.dtype + ) + if attn_mask.dtype == torch.uint8: + warnings.warn( + "Byte tensor for attn_mask is deprecated. Use bool tensor instead." + ) + attn_mask = attn_mask.to(torch.bool) + + if attn_mask.dim() == 2: + attn_mask = attn_mask.unsqueeze(0) + if list(attn_mask.size()) != [1, query.size(0), key.size(0)]: + raise RuntimeError( + "The size of the 2D attn_mask is not correct." + ) + elif attn_mask.dim() == 3: + if list(attn_mask.size()) != [ + bsz * num_heads, + query.size(0), + key.size(0), + ]: + raise RuntimeError( + "The size of the 3D attn_mask is not correct." + ) + else: + raise RuntimeError( + "attn_mask's dimension {} is not supported".format( + attn_mask.dim() + ) + ) + # attn_mask's dim is 3 now. + + # convert ByteTensor key_padding_mask to bool + if ( + key_padding_mask is not None + and key_padding_mask.dtype == torch.uint8 + ): + warnings.warn( + "Byte tensor for key_padding_mask is deprecated. Use bool tensor instead." + ) + key_padding_mask = key_padding_mask.to(torch.bool) + + q = (q * scaling).contiguous().view(tgt_len, bsz, num_heads, head_dim) + k = k.contiguous().view(-1, bsz, num_heads, head_dim) + v = v.contiguous().view(-1, bsz * num_heads, head_dim).transpose(0, 1) + + src_len = k.size(0) + + if key_padding_mask is not None: + assert key_padding_mask.size(0) == bsz, "{} == {}".format( + key_padding_mask.size(0), bsz + ) + assert key_padding_mask.size(1) == src_len, "{} == {}".format( + key_padding_mask.size(1), src_len + ) + + q = q.transpose(0, 1) # (batch, time1, head, d_k) + + pos_emb_bsz = pos_emb.size(0) + assert pos_emb_bsz in (1, bsz) # actually it is 1 + p = self.linear_pos(pos_emb).view(pos_emb_bsz, -1, num_heads, head_dim) + p = p.transpose(1, 2) # (batch, head, 2*time1-1, d_k) + + q_with_bias_u = (q + self._pos_bias_u()).transpose( + 1, 2 + ) # (batch, head, time1, d_k) + + q_with_bias_v = (q + self._pos_bias_v()).transpose( + 1, 2 + ) # (batch, head, time1, d_k) + + # compute attention score + # first compute matrix a and matrix c + # as described in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" Section 3.3 + k = k.permute(1, 2, 3, 0) # (batch, head, d_k, time2) + matrix_ac = torch.matmul( + q_with_bias_u, k + ) # (batch, head, time1, time2) + + # compute matrix b and matrix d + matrix_bd = torch.matmul( + q_with_bias_v, p.transpose(-2, -1) + ) # (batch, head, time1, 2*time1-1) + matrix_bd = self.rel_shift(matrix_bd) + + attn_output_weights = ( + matrix_ac + matrix_bd + ) # (batch, head, time1, time2) + + attn_output_weights = attn_output_weights.view( + bsz * num_heads, tgt_len, -1 + ) + + assert list(attn_output_weights.size()) == [ + bsz * num_heads, + tgt_len, + src_len, + ] + + if attn_mask is not None: + if attn_mask.dtype == torch.bool: + attn_output_weights.masked_fill_(attn_mask, float("-inf")) + else: + attn_output_weights += attn_mask + + if key_padding_mask is not None: + attn_output_weights = attn_output_weights.view( + bsz, num_heads, tgt_len, src_len + ) + attn_output_weights = attn_output_weights.masked_fill( + key_padding_mask.unsqueeze(1).unsqueeze(2), + float("-inf"), + ) + attn_output_weights = attn_output_weights.view( + bsz * num_heads, tgt_len, src_len + ) + + attn_output_weights = nn.functional.softmax(attn_output_weights, dim=-1) + attn_output_weights = nn.functional.dropout( + attn_output_weights, p=dropout_p, training=training + ) + + attn_output = torch.bmm(attn_output_weights, v) + assert list(attn_output.size()) == [bsz * num_heads, tgt_len, head_dim] + attn_output = ( + attn_output.transpose(0, 1) + .contiguous() + .view(tgt_len, bsz, embed_dim) + ) + attn_output = nn.functional.linear( + attn_output, out_proj_weight, out_proj_bias + ) + + if need_weights: + # average attention weights over heads + attn_output_weights = attn_output_weights.view( + bsz, num_heads, tgt_len, src_len + ) + return attn_output, attn_output_weights.sum(dim=1) / num_heads + else: + return attn_output, None + + +class ConvolutionModule(nn.Module): + """ConvolutionModule in Conformer model. + Modified from https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/conformer/convolution.py + + Args: + channels (int): The number of channels of conv layers. + kernel_size (int): Kernerl size of conv layers. + bias (bool): Whether to use bias in conv layers (default=True). + + """ + + def __init__( + self, channels: int, kernel_size: int, bias: bool = True + ) -> None: + """Construct an ConvolutionModule object.""" + super(ConvolutionModule, self).__init__() + # kernerl_size should be a odd number for 'SAME' padding + assert (kernel_size - 1) % 2 == 0 + + self.pointwise_conv1 = ScaledConv1d( + channels, + 2 * channels, + kernel_size=1, + stride=1, + padding=0, + bias=bias, + ) + + # after pointwise_conv1 we put x through a gated linear unit (nn.functional.glu). + # For most layers the normal rms value of channels of x seems to be in the range 1 to 4, + # but sometimes, for some reason, for layer 0 the rms ends up being very large, + # between 50 and 100 for different channels. This will cause very peaky and + # sparse derivatives for the sigmoid gating function, which will tend to make + # the loss function not learn effectively. (for most layers the average absolute values + # are in the range 0.5..9.0, and the average p(x>0), i.e. positive proportion, + # at the output of pointwise_conv1.output is around 0.35 to 0.45 for different + # layers, which likely breaks down as 0.5 for the "linear" half and + # 0.2 to 0.3 for the part that goes into the sigmoid. The idea is that if we + # constrain the rms values to a reasonable range via a constraint of max_abs=10.0, + # it will be in a better position to start learning something, i.e. to latch onto + # the correct range. + self.deriv_balancer1 = ActivationBalancer( + channel_dim=1, max_abs=10.0, min_positive=0.05, max_positive=1.0 + ) + + self.depthwise_conv = ScaledConv1d( + channels, + channels, + kernel_size, + stride=1, + padding=(kernel_size - 1) // 2, + groups=channels, + bias=bias, + ) + + self.deriv_balancer2 = ActivationBalancer( + channel_dim=1, min_positive=0.05, max_positive=1.0 + ) + + self.activation = DoubleSwish() + + self.pointwise_conv2 = ScaledConv1d( + channels, + channels, + kernel_size=1, + stride=1, + padding=0, + bias=bias, + initial_scale=0.25, + ) + + def forward(self, x: Tensor) -> Tensor: + """Compute convolution module. + + Args: + x: Input tensor (#time, batch, channels). + + Returns: + Tensor: Output tensor (#time, batch, channels). + + """ + # exchange the temporal dimension and the feature dimension + x = x.permute(1, 2, 0) # (#batch, channels, time). + + # GLU mechanism + x = self.pointwise_conv1(x) # (batch, 2*channels, time) + + x = self.deriv_balancer1(x) + x = nn.functional.glu(x, dim=1) # (batch, channels, time) + + # 1D Depthwise Conv + x = self.depthwise_conv(x) + + x = self.deriv_balancer2(x) + x = self.activation(x) + + x = self.pointwise_conv2(x) # (batch, channel, time) + + return x.permute(2, 0, 1) + + +class Conv2dSubsampling(nn.Module): + """Convolutional 2D subsampling (to 1/4 length). + + Convert an input of shape (N, T, idim) to an output + with shape (N, T', odim), where + T' = ((T-1)//2 - 1)//2, which approximates T' == T//4 + + It is based on + https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/transformer/subsampling.py # noqa + """ + + def __init__( + self, + in_channels: int, + out_channels: int, + layer1_channels: int = 8, + layer2_channels: int = 32, + layer3_channels: int = 128, + ) -> None: + """ + Args: + in_channels: + Number of channels in. The input shape is (N, T, in_channels). + Caution: It requires: T >=7, in_channels >=7 + out_channels + Output dim. The output shape is (N, ((T-1)//2 - 1)//2, out_channels) + layer1_channels: + Number of channels in layer1 + layer1_channels: + Number of channels in layer2 + """ + assert in_channels >= 7 + super().__init__() + + self.conv = nn.Sequential( + ScaledConv2d( + in_channels=1, + out_channels=layer1_channels, + kernel_size=3, + padding=1, + ), + ActivationBalancer(channel_dim=1), + DoubleSwish(), + ScaledConv2d( + in_channels=layer1_channels, + out_channels=layer2_channels, + kernel_size=3, + stride=2, + ), + ActivationBalancer(channel_dim=1), + DoubleSwish(), + ScaledConv2d( + in_channels=layer2_channels, + out_channels=layer3_channels, + kernel_size=3, + stride=2, + ), + ActivationBalancer(channel_dim=1), + DoubleSwish(), + ) + self.out = ScaledLinear( + layer3_channels * (((in_channels - 1) // 2 - 1) // 2), out_channels + ) + # set learn_eps=False because out_norm is preceded by `out`, and `out` + # itself has learned scale, so the extra degree of freedom is not + # needed. + self.out_norm = BasicNorm(out_channels, learn_eps=False) + # constrain median of output to be close to zero. + self.out_balancer = ActivationBalancer( + channel_dim=-1, min_positive=0.45, max_positive=0.55 + ) + + def forward(self, x: torch.Tensor) -> torch.Tensor: + """Subsample x. + + Args: + x: + Its shape is (N, T, idim). + + Returns: + Return a tensor of shape (N, ((T-1)//2 - 1)//2, odim) + """ + # On entry, x is (N, T, idim) + x = x.unsqueeze(1) # (N, T, idim) -> (N, 1, T, idim) i.e., (N, C, H, W) + x = self.conv(x) + # Now x is of shape (N, odim, ((T-1)//2 - 1)//2, ((idim-1)//2 - 1)//2) + b, c, t, f = x.size() + x = self.out(x.transpose(1, 2).contiguous().view(b, t, c * f)) + # Now x is of shape (N, ((T-1)//2 - 1))//2, odim) + x = self.out_norm(x) + x = self.out_balancer(x) + return x + + +if __name__ == "__main__": + feature_dim = 50 + c = Conformer(num_features=feature_dim, d_model=128, nhead=4) + batch_size = 5 + seq_len = 20 + # Just make sure the forward pass runs. + f = c( + torch.randn(batch_size, seq_len, feature_dim), + torch.full((batch_size,), seq_len, dtype=torch.int64), + warmup=0.5, + ) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py new file mode 100755 index 000000000..076feb7a9 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py @@ -0,0 +1,557 @@ +#!/usr/bin/env python3 +# +# Copyright 2021 Xiaomi Corporation (Author: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +""" +Usage: +(1) greedy search +./pruned_transducer_stateless2/decode.py \ + --epoch 28 \ + --avg 15 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method greedy_search + +(2) beam search +./pruned_transducer_stateless2/decode.py \ + --epoch 28 \ + --avg 15 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method beam_search \ + --beam-size 4 + +(3) modified beam search +./pruned_transducer_stateless2/decode.py \ + --epoch 28 \ + --avg 15 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method modified_beam_search \ + --beam-size 4 + +(4) fast beam search +./pruned_transducer_stateless2/decode.py \ + --epoch 28 \ + --avg 15 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 1500 \ + --decoding-method fast_beam_search \ + --beam 4 \ + --max-contexts 4 \ + --max-states 8 +""" + + +import argparse +import logging +from collections import defaultdict +from pathlib import Path +from typing import Dict, List, Optional, Tuple + +import k2 +import sentencepiece as spm +import torch +import torch.nn as nn +from asr_datamodule import GigaSpeechAsrDataModule +from beam_search import ( + beam_search, + fast_beam_search, + greedy_search, + greedy_search_batch, + modified_beam_search, +) +from gigaspeech_scoring import asr_text_post_processing +from train import get_params, get_transducer_model + +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) +from icefall.utils import ( + AttributeDict, + setup_logger, + store_transcripts, + write_error_stats, +) + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--epoch", + type=int, + default=29, + help="It specifies the checkpoint to use for decoding." + "Note: Epoch counts from 0.", + ) + parser.add_argument( + "--avg", + type=int, + default=8, + help="Number of checkpoints to average. Automatically select " + "consecutive checkpoints before the checkpoint specified by " + "'--epoch'. ", + ) + + parser.add_argument( + "--avg-last-n", + type=int, + default=0, + help="""If positive, --epoch and --avg are ignored and it + will use the last n checkpoints exp_dir/checkpoint-xxx.pt + where xxx is the number of processed batches while + saving that checkpoint. + """, + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="The experiment dir", + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--decoding-method", + type=str, + default="greedy_search", + help="""Possible values are: + - greedy_search + - beam_search + - modified_beam_search + - fast_beam_search + """, + ) + + parser.add_argument( + "--beam-size", + type=int, + default=4, + help="""An interger indicating how many candidates we will keep for each + frame. Used only when --decoding-method is beam_search or + modified_beam_search.""", + ) + + parser.add_argument( + "--beam", + type=float, + default=4, + help="""A floating point value to calculate the cutoff score during beam + search (i.e., `cutoff = max-score - beam`), which is the same as the + `beam` in Kaldi. + Used only when --decoding-method is fast_beam_search""", + ) + + parser.add_argument( + "--max-contexts", + type=int, + default=4, + help="""Used only when --decoding-method is + fast_beam_search""", + ) + + parser.add_argument( + "--max-states", + type=int, + default=8, + help="""Used only when --decoding-method is + fast_beam_search""", + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + parser.add_argument( + "--max-sym-per-frame", + type=int, + default=1, + help="""Maximum number of symbols per frame. + Used only when --decoding_method is greedy_search""", + ) + + return parser + + +def post_processing( + results: List[Tuple[List[str], List[str]]], +) -> List[Tuple[List[str], List[str]]]: + new_results = [] + for ref, hyp in results: + new_ref = asr_text_post_processing(" ".join(ref)).split() + new_hyp = asr_text_post_processing(" ".join(hyp)).split() + new_results.append((new_ref, new_hyp)) + return new_results + + +def decode_one_batch( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + batch: dict, + decoding_graph: Optional[k2.Fsa] = None, +) -> Dict[str, List[List[str]]]: + """Decode one batch and return the result in a dict. The dict has the + following format: + + - key: It indicates the setting used for decoding. For example, + if greedy_search is used, it would be "greedy_search" + If beam search with a beam size of 7 is used, it would be + "beam_7" + - value: It contains the decoding result. `len(value)` equals to + batch size. `value[i]` is the decoding result for the i-th + utterance in the given batch. + Args: + params: + It's the return value of :func:`get_params`. + model: + The neural model. + sp: + The BPE model. + batch: + It is the return value from iterating + `lhotse.dataset.K2SpeechRecognitionDataset`. See its documentation + for the format of the `batch`. + decoding_graph: + The decoding graph. Can be either a `k2.trivial_graph` or HLG, Used + only when --decoding_method is fast_beam_search. + Returns: + Return the decoding result. See above description for the format of + the returned dict. + """ + device = model.device + feature = batch["inputs"] + assert feature.ndim == 3 + + feature = feature.to(device) + # at entry, feature is (N, T, C) + + supervisions = batch["supervisions"] + feature_lens = supervisions["num_frames"].to(device) + + encoder_out, encoder_out_lens = model.encoder( + x=feature, x_lens=feature_lens + ) + hyps = [] + + if params.decoding_method == "fast_beam_search": + hyp_tokens = fast_beam_search( + model=model, + decoding_graph=decoding_graph, + encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, + beam=params.beam, + max_contexts=params.max_contexts, + max_states=params.max_states, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + elif ( + params.decoding_method == "greedy_search" + and params.max_sym_per_frame == 1 + ): + hyp_tokens = greedy_search_batch( + model=model, + encoder_out=encoder_out, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + elif params.decoding_method == "modified_beam_search": + hyp_tokens = modified_beam_search( + model=model, + encoder_out=encoder_out, + beam=params.beam_size, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + else: + batch_size = encoder_out.size(0) + + for i in range(batch_size): + # fmt: off + encoder_out_i = encoder_out[i:i+1, :encoder_out_lens[i]] + # fmt: on + if params.decoding_method == "greedy_search": + hyp = greedy_search( + model=model, + encoder_out=encoder_out_i, + max_sym_per_frame=params.max_sym_per_frame, + ) + elif params.decoding_method == "beam_search": + hyp = beam_search( + model=model, + encoder_out=encoder_out_i, + beam=params.beam_size, + ) + else: + raise ValueError( + f"Unsupported decoding method: {params.decoding_method}" + ) + hyps.append(sp.decode(hyp).split()) + + if params.decoding_method == "greedy_search": + return {"greedy_search": hyps} + elif params.decoding_method == "fast_beam_search": + return { + ( + f"beam_{params.beam}_" + f"max_contexts_{params.max_contexts}_" + f"max_states_{params.max_states}" + ): hyps + } + else: + return {f"beam_size_{params.beam_size}": hyps} + + +def decode_dataset( + dl: torch.utils.data.DataLoader, + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + decoding_graph: Optional[k2.Fsa] = None, +) -> Dict[str, List[Tuple[List[str], List[str]]]]: + """Decode dataset. + + Args: + dl: + PyTorch's dataloader containing the dataset to decode. + params: + It is returned by :func:`get_params`. + model: + The neural model. + sp: + The BPE model. + decoding_graph: + The decoding graph. Can be either a `k2.trivial_graph` or HLG, Used + only when --decoding_method is fast_beam_search. + Returns: + Return a dict, whose key may be "greedy_search" if greedy search + is used, or it may be "beam_7" if beam size of 7 is used. + Its value is a list of tuples. Each tuple contains two elements: + The first is the reference transcript, and the second is the + predicted result. + """ + num_cuts = 0 + + try: + num_batches = len(dl) + except TypeError: + num_batches = "?" + + log_interval = 100 + + results = defaultdict(list) + for batch_idx, batch in enumerate(dl): + texts = batch["supervisions"]["text"] + + hyps_dict = decode_one_batch( + params=params, + model=model, + sp=sp, + decoding_graph=decoding_graph, + batch=batch, + ) + + for name, hyps in hyps_dict.items(): + this_batch = [] + assert len(hyps) == len(texts) + for hyp_words, ref_text in zip(hyps, texts): + ref_words = ref_text.split() + this_batch.append((ref_words, hyp_words)) + + results[name].extend(this_batch) + + num_cuts += len(texts) + + if batch_idx % log_interval == 0: + batch_str = f"{batch_idx}/{num_batches}" + + logging.info( + f"batch {batch_str}, cuts processed until now is {num_cuts}" + ) + return results + + +def save_results( + params: AttributeDict, + test_set_name: str, + results_dict: Dict[str, List[Tuple[List[int], List[int]]]], +): + test_set_wers = dict() + for key, results in results_dict.items(): + recog_path = ( + params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt" + ) + results = post_processing(results) + store_transcripts(filename=recog_path, texts=results) + logging.info(f"The transcripts are stored in {recog_path}") + + # The following prints out WERs, per-word error statistics and aligned + # ref/hyp pairs. + errs_filename = ( + params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt" + ) + with open(errs_filename, "w") as f: + wer = write_error_stats( + f, f"{test_set_name}-{key}", results, enable_log=True + ) + test_set_wers[key] = wer + + logging.info("Wrote detailed error stats to {}".format(errs_filename)) + + test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1]) + errs_info = ( + params.res_dir + / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt" + ) + with open(errs_info, "w") as f: + print("settings\tWER", file=f) + for key, val in test_set_wers: + print("{}\t{}".format(key, val), file=f) + + s = "\nFor {}, WER of different settings are:\n".format(test_set_name) + note = "\tbest for {}".format(test_set_name) + for key, val in test_set_wers: + s += "{}\t{}{}\n".format(key, val, note) + note = "" + logging.info(s) + + +@torch.no_grad() +def main(): + parser = get_parser() + GigaSpeechAsrDataModule.add_arguments(parser) + args = parser.parse_args() + args.exp_dir = Path(args.exp_dir) + + params = get_params() + params.update(vars(args)) + + assert params.decoding_method in ( + "greedy_search", + "beam_search", + "fast_beam_search", + "modified_beam_search", + ) + params.res_dir = params.exp_dir / params.decoding_method + + params.suffix = f"epoch-{params.epoch}-avg-{params.avg}" + if "fast_beam_search" in params.decoding_method: + params.suffix += f"-beam-{params.beam}" + params.suffix += f"-max-contexts-{params.max_contexts}" + params.suffix += f"-max-states-{params.max_states}" + elif "beam_search" in params.decoding_method: + params.suffix += f"-beam-{params.beam_size}" + else: + params.suffix += f"-context-{params.context_size}" + params.suffix += f"-max-sym-per-frame-{params.max_sym_per_frame}" + + setup_logger(f"{params.res_dir}/log-decode-{params.suffix}") + logging.info("Decoding started") + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", 0) + + logging.info(f"Device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + if params.avg_last_n > 0: + filenames = find_checkpoints(params.exp_dir)[: params.avg_last_n] + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: + load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) + else: + start = params.epoch - params.avg + 1 + filenames = [] + for i in range(start, params.epoch + 1): + if start >= 0: + filenames.append(f"{params.exp_dir}/epoch-{i}.pt") + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + + model.to(device) + model.eval() + model.device = device + + if params.decoding_method == "fast_beam_search": + decoding_graph = k2.trivial_graph(params.vocab_size - 1, device=device) + else: + decoding_graph = None + + num_param = sum([p.numel() for p in model.parameters()]) + logging.info(f"Number of model parameters: {num_param}") + + gigaspeech = GigaSpeechAsrDataModule(args) + + dev_cuts = gigaspeech.dev_cuts() + test_cuts = gigaspeech.test_cuts() + + dev_dl = gigaspeech.test_dataloaders(dev_cuts) + test_dl = gigaspeech.test_dataloaders(test_cuts) + + test_sets = ["dev", "test"] + test_dls = [dev_dl, test_dl] + + for test_set, test_dl in zip(test_sets, test_dls): + results_dict = decode_dataset( + dl=test_dl, + params=params, + model=model, + sp=sp, + decoding_graph=decoding_graph, + ) + + save_results( + params=params, + test_set_name=test_set, + results_dict=results_dict, + ) + + logging.info("Done!") + + +if __name__ == "__main__": + main() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py new file mode 100644 index 000000000..b6d94aaf1 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py @@ -0,0 +1,103 @@ +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import torch +import torch.nn as nn +import torch.nn.functional as F +from scaling import ScaledConv1d, ScaledEmbedding + + +class Decoder(nn.Module): + """This class modifies the stateless decoder from the following paper: + + RNN-transducer with stateless prediction network + https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9054419 + + It removes the recurrent connection from the decoder, i.e., the prediction + network. Different from the above paper, it adds an extra Conv1d + right after the embedding layer. + + TODO: Implement https://arxiv.org/pdf/2109.07513.pdf + """ + + def __init__( + self, + vocab_size: int, + decoder_dim: int, + blank_id: int, + context_size: int, + ): + """ + Args: + vocab_size: + Number of tokens of the modeling unit including blank. + decoder_dim: + Dimension of the input embedding, and of the decoder output. + blank_id: + The ID of the blank symbol. + context_size: + Number of previous words to use to predict the next word. + 1 means bigram; 2 means trigram. n means (n+1)-gram. + """ + super().__init__() + + self.embedding = ScaledEmbedding( + num_embeddings=vocab_size, + embedding_dim=decoder_dim, + padding_idx=blank_id, + ) + self.blank_id = blank_id + + assert context_size >= 1, context_size + self.context_size = context_size + self.vocab_size = vocab_size + if context_size > 1: + self.conv = ScaledConv1d( + in_channels=decoder_dim, + out_channels=decoder_dim, + kernel_size=context_size, + padding=0, + groups=decoder_dim, + bias=False, + ) + + def forward(self, y: torch.Tensor, need_pad: bool = True) -> torch.Tensor: + """ + Args: + y: + A 2-D tensor of shape (N, U). + need_pad: + True to left pad the input. Should be True during training. + False to not pad the input. Should be False during inference. + Returns: + Return a tensor of shape (N, U, decoder_dim). + """ + y = y.to(torch.int64) + embedding_out = self.embedding(y) + if self.context_size > 1: + embedding_out = embedding_out.permute(0, 2, 1) + if need_pad is True: + embedding_out = F.pad( + embedding_out, pad=(self.context_size - 1, 0) + ) + else: + # During inference time, there is no need to do extra padding + # as we only need one output + assert embedding_out.size(-1) == self.context_size + embedding_out = self.conv(embedding_out) + embedding_out = embedding_out.permute(0, 2, 1) + embedding_out = F.relu(embedding_out) + return embedding_out diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py new file mode 100644 index 000000000..257facce4 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py @@ -0,0 +1,43 @@ +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from typing import Tuple + +import torch +import torch.nn as nn + + +class EncoderInterface(nn.Module): + def forward( + self, x: torch.Tensor, x_lens: torch.Tensor + ) -> Tuple[torch.Tensor, torch.Tensor]: + """ + Args: + x: + A tensor of shape (batch_size, input_seq_len, num_features) + containing the input features. + x_lens: + A tensor of shape (batch_size,) containing the number of frames + in `x` before padding. + Returns: + Return a tuple containing two tensors: + - encoder_out, a tensor of (batch_size, out_seq_len, output_dim) + containing unnormalized probabilities, i.e., the output of a + linear layer. + - encoder_out_lens, a tensor of shape (batch_size,) containing + the number of frames in `encoder_out` before padding. + """ + raise NotImplementedError("Please implement it in a subclass") diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py new file mode 100755 index 000000000..b5757ee8c --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py @@ -0,0 +1,182 @@ +#!/usr/bin/env python3 +# +# Copyright 2021 Xiaomi Corporation (Author: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +# This script converts several saved checkpoints +# to a single one using model averaging. +""" +Usage: +./pruned_transducer_stateless2/export.py \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --epoch 20 \ + --avg 10 + +It will generate a file exp_dir/pretrained.pt + +To use the generated file with `pruned_transducer_stateless2/decode.py`, +you can do: + + cd /path/to/exp_dir + ln -s pretrained.pt epoch-9999.pt + + cd /path/to/egs/librispeech/ASR + ./pruned_transducer_stateless2/decode.py \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --epoch 9999 \ + --avg 1 \ + --max-duration 100 \ + --bpe-model data/lang_bpe_500/bpe.model +""" + +import argparse +import logging +from pathlib import Path + +import sentencepiece as spm +import torch +from train import get_params, get_transducer_model + +from icefall.checkpoint import average_checkpoints, load_checkpoint +from icefall.utils import str2bool + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--epoch", + type=int, + default=28, + help="It specifies the checkpoint to use for decoding." + "Note: Epoch counts from 0.", + ) + + parser.add_argument( + "--avg", + type=int, + default=15, + help="Number of checkpoints to average. Automatically select " + "consecutive checkpoints before the checkpoint specified by " + "'--epoch'. ", + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="""It specifies the directory where all training related + files, e.g., checkpoints, log, etc, are saved + """, + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--jit", + type=str2bool, + default=False, + help="""True to save a model after applying torch.jit.script. + """, + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + + return parser + + +def main(): + args = get_parser().parse_args() + args.exp_dir = Path(args.exp_dir) + + assert args.jit is False, "Support torchscript will be added later" + + params = get_params() + params.update(vars(args)) + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", 0) + + logging.info(f"device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + model.to(device) + + if params.avg == 1: + load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) + else: + start = params.epoch - params.avg + 1 + filenames = [] + for i in range(start, params.epoch + 1): + if start >= 0: + filenames.append(f"{params.exp_dir}/epoch-{i}.pt") + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + + model.eval() + + model.to("cpu") + model.eval() + + if params.jit: + logging.info("Using torch.jit.script") + model = torch.jit.script(model) + filename = params.exp_dir / "cpu_jit.pt" + model.save(str(filename)) + logging.info(f"Saved to {filename}") + else: + logging.info("Not using torch.jit.script") + # Save it using a format so that it can be loaded + # by :func:`load_checkpoint` + filename = params.exp_dir / "pretrained.pt" + torch.save({"model": model.state_dict()}, str(filename)) + logging.info(f"Saved to {filename}") + + +if __name__ == "__main__": + formatter = ( + "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + ) + + logging.basicConfig(format=formatter, level=logging.INFO) + main() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/gigaspeech_scoring.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/gigaspeech_scoring.py new file mode 120000 index 000000000..a6a4d12b1 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/gigaspeech_scoring.py @@ -0,0 +1 @@ +../conformer_ctc/gigaspeech_scoring.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py new file mode 100644 index 000000000..35f75ed2a --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py @@ -0,0 +1,67 @@ +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import torch +import torch.nn as nn +from scaling import ScaledLinear + + +class Joiner(nn.Module): + def __init__( + self, + encoder_dim: int, + decoder_dim: int, + joiner_dim: int, + vocab_size: int, + ): + super().__init__() + + self.encoder_proj = ScaledLinear(encoder_dim, joiner_dim) + self.decoder_proj = ScaledLinear(decoder_dim, joiner_dim) + self.output_linear = ScaledLinear(joiner_dim, vocab_size) + + def forward( + self, + encoder_out: torch.Tensor, + decoder_out: torch.Tensor, + project_input: bool = True, + ) -> torch.Tensor: + """ + Args: + encoder_out: + Output from the encoder. Its shape is (N, T, s_range, C). + decoder_out: + Output from the decoder. Its shape is (N, T, s_range, C). + project_input: + If true, apply input projections encoder_proj and decoder_proj. + If this is false, it is the user's responsibility to do this + manually. + Returns: + Return a tensor of shape (N, T, s_range, C). + """ + assert encoder_out.ndim == decoder_out.ndim == 4 + assert encoder_out.shape[:-1] == decoder_out.shape[:-1] + + if project_input: + logit = self.encoder_proj(encoder_out) + self.decoder_proj( + decoder_out + ) + else: + logit = encoder_out + decoder_out + + logit = self.output_linear(torch.tanh(logit)) + + return logit diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py new file mode 100644 index 000000000..599bf2506 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py @@ -0,0 +1,193 @@ +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang, Wei Kang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +import k2 +import torch +import torch.nn as nn +from encoder_interface import EncoderInterface +from scaling import ScaledLinear + +from icefall.utils import add_sos + + +class Transducer(nn.Module): + """It implements https://arxiv.org/pdf/1211.3711.pdf + "Sequence Transduction with Recurrent Neural Networks" + """ + + def __init__( + self, + encoder: EncoderInterface, + decoder: nn.Module, + joiner: nn.Module, + encoder_dim: int, + decoder_dim: int, + joiner_dim: int, + vocab_size: int, + ): + """ + Args: + encoder: + It is the transcription network in the paper. Its accepts + two inputs: `x` of (N, T, encoder_dim) and `x_lens` of shape (N,). + It returns two tensors: `logits` of shape (N, T, encoder_dm) and + `logit_lens` of shape (N,). + decoder: + It is the prediction network in the paper. Its input shape + is (N, U) and its output shape is (N, U, decoder_dim). + It should contain one attribute: `blank_id`. + joiner: + It has two inputs with shapes: (N, T, encoder_dim) and (N, U, decoder_dim). + Its output shape is (N, T, U, vocab_size). Note that its output contains + unnormalized probs, i.e., not processed by log-softmax. + """ + super().__init__() + assert isinstance(encoder, EncoderInterface), type(encoder) + assert hasattr(decoder, "blank_id") + + self.encoder = encoder + self.decoder = decoder + self.joiner = joiner + + self.simple_am_proj = ScaledLinear( + encoder_dim, vocab_size, initial_speed=0.5 + ) + self.simple_lm_proj = ScaledLinear(decoder_dim, vocab_size) + + def forward( + self, + x: torch.Tensor, + x_lens: torch.Tensor, + y: k2.RaggedTensor, + prune_range: int = 5, + am_scale: float = 0.0, + lm_scale: float = 0.0, + warmup: float = 1.0, + ) -> torch.Tensor: + """ + Args: + x: + A 3-D tensor of shape (N, T, C). + x_lens: + A 1-D tensor of shape (N,). It contains the number of frames in `x` + before padding. + y: + A ragged tensor with 2 axes [utt][label]. It contains labels of each + utterance. + prune_range: + The prune range for rnnt loss, it means how many symbols(context) + we are considering for each frame to compute the loss. + am_scale: + The scale to smooth the loss with am (output of encoder network) + part + lm_scale: + The scale to smooth the loss with lm (output of predictor network) + part + warmup: + A value warmup >= 0 that determines which modules are active, values + warmup > 1 "are fully warmed up" and all modules will be active. + Returns: + Return the transducer loss. + + Note: + Regarding am_scale & lm_scale, it will make the loss-function one of + the form: + lm_scale * lm_probs + am_scale * am_probs + + (1-lm_scale-am_scale) * combined_probs + """ + assert x.ndim == 3, x.shape + assert x_lens.ndim == 1, x_lens.shape + assert y.num_axes == 2, y.num_axes + + assert x.size(0) == x_lens.size(0) == y.dim0 + + encoder_out, x_lens = self.encoder(x, x_lens, warmup=warmup) + assert torch.all(x_lens > 0) + + # Now for the decoder, i.e., the prediction network + row_splits = y.shape.row_splits(1) + y_lens = row_splits[1:] - row_splits[:-1] + + blank_id = self.decoder.blank_id + sos_y = add_sos(y, sos_id=blank_id) + + # sos_y_padded: [B, S + 1], start with SOS. + sos_y_padded = sos_y.pad(mode="constant", padding_value=blank_id) + + # decoder_out: [B, S + 1, decoder_dim] + decoder_out = self.decoder(sos_y_padded) + + # Note: y does not start with SOS + # y_padded : [B, S] + y_padded = y.pad(mode="constant", padding_value=0) + + y_padded = y_padded.to(torch.int64) + boundary = torch.zeros( + (x.size(0), 4), dtype=torch.int64, device=x.device + ) + boundary[:, 2] = y_lens + boundary[:, 3] = x_lens + + lm = self.simple_lm_proj(decoder_out) + am = self.simple_am_proj(encoder_out) + + with torch.cuda.amp.autocast(enabled=False): + simple_loss, (px_grad, py_grad) = k2.rnnt_loss_smoothed( + lm=lm.float(), + am=am.float(), + symbols=y_padded, + termination_symbol=blank_id, + lm_only_scale=lm_scale, + am_only_scale=am_scale, + boundary=boundary, + reduction="sum", + return_grad=True, + ) + + # ranges : [B, T, prune_range] + ranges = k2.get_rnnt_prune_ranges( + px_grad=px_grad, + py_grad=py_grad, + boundary=boundary, + s_range=prune_range, + ) + + # am_pruned : [B, T, prune_range, encoder_dim] + # lm_pruned : [B, T, prune_range, decoder_dim] + am_pruned, lm_pruned = k2.do_rnnt_pruning( + am=self.joiner.encoder_proj(encoder_out), + lm=self.joiner.decoder_proj(decoder_out), + ranges=ranges, + ) + + # logits : [B, T, prune_range, vocab_size] + + # project_input=False since we applied the decoder's input projections + # prior to do_rnnt_pruning (this is an optimization for speed). + logits = self.joiner(am_pruned, lm_pruned, project_input=False) + + with torch.cuda.amp.autocast(enabled=False): + pruned_loss = k2.rnnt_loss_pruned( + logits=logits.float(), + symbols=y_padded, + ranges=ranges, + termination_symbol=blank_id, + boundary=boundary, + reduction="sum", + ) + + return (simple_loss, pruned_loss) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py new file mode 100644 index 000000000..432bf8220 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py @@ -0,0 +1,331 @@ +# Copyright 2022 Xiaomi Corp. (authors: Daniel Povey) +# +# See ../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +from typing import List, Optional, Union + +import torch +from torch.optim import Optimizer + + +class Eve(Optimizer): + r""" + Implements Eve algorithm. This is a modified version of AdamW with a special + way of setting the weight-decay / shrinkage-factor, which is designed to make the + rms of the parameters approach a particular target_rms (default: 0.1). This is + for use with networks with 'scaled' versions of modules (see scaling.py), which + will be close to invariant to the absolute scale on the parameter matrix. + + The original Adam algorithm was proposed in `Adam: A Method for Stochastic Optimization`_. + The AdamW variant was proposed in `Decoupled Weight Decay Regularization`_. + Eve is unpublished so far. + + Arguments: + params (iterable): iterable of parameters to optimize or dicts defining + parameter groups + lr (float, optional): learning rate (default: 1e-3) + betas (Tuple[float, float], optional): coefficients used for computing + running averages of gradient and its square (default: (0.9, 0.999)) + eps (float, optional): term added to the denominator to improve + numerical stability (default: 1e-8) + weight_decay (float, optional): weight decay coefficient (default: 3e-4; + this value means that the weight would decay significantly after + about 3k minibatches. Is not multiplied by learning rate, but + is conditional on RMS-value of parameter being > target_rms. + target_rms (float, optional): target root-mean-square value of + parameters, if they fall below this we will stop applying weight decay. + + + .. _Adam\: A Method for Stochastic Optimization: + https://arxiv.org/abs/1412.6980 + .. _Decoupled Weight Decay Regularization: + https://arxiv.org/abs/1711.05101 + .. _On the Convergence of Adam and Beyond: + https://openreview.net/forum?id=ryQu7f-RZ + """ + + def __init__( + self, + params, + lr=1e-3, + betas=(0.9, 0.98), + eps=1e-8, + weight_decay=1e-3, + target_rms=0.1, + ): + + if not 0.0 <= lr: + raise ValueError("Invalid learning rate: {}".format(lr)) + if not 0.0 <= eps: + raise ValueError("Invalid epsilon value: {}".format(eps)) + if not 0.0 <= betas[0] < 1.0: + raise ValueError( + "Invalid beta parameter at index 0: {}".format(betas[0]) + ) + if not 0.0 <= betas[1] < 1.0: + raise ValueError( + "Invalid beta parameter at index 1: {}".format(betas[1]) + ) + if not 0 <= weight_decay <= 0.1: + raise ValueError( + "Invalid weight_decay value: {}".format(weight_decay) + ) + if not 0 < target_rms <= 10.0: + raise ValueError("Invalid target_rms value: {}".format(target_rms)) + defaults = dict( + lr=lr, + betas=betas, + eps=eps, + weight_decay=weight_decay, + target_rms=target_rms, + ) + super(Eve, self).__init__(params, defaults) + + def __setstate__(self, state): + super(Eve, self).__setstate__(state) + + @torch.no_grad() + def step(self, closure=None): + """Performs a single optimization step. + + Arguments: + closure (callable, optional): A closure that reevaluates the model + and returns the loss. + """ + loss = None + if closure is not None: + with torch.enable_grad(): + loss = closure() + + for group in self.param_groups: + for p in group["params"]: + if p.grad is None: + continue + + # Perform optimization step + grad = p.grad + if grad.is_sparse: + raise RuntimeError( + "AdamW does not support sparse gradients" + ) + + state = self.state[p] + + # State initialization + if len(state) == 0: + state["step"] = 0 + # Exponential moving average of gradient values + state["exp_avg"] = torch.zeros_like( + p, memory_format=torch.preserve_format + ) + # Exponential moving average of squared gradient values + state["exp_avg_sq"] = torch.zeros_like( + p, memory_format=torch.preserve_format + ) + + exp_avg, exp_avg_sq = state["exp_avg"], state["exp_avg_sq"] + + beta1, beta2 = group["betas"] + + state["step"] += 1 + bias_correction1 = 1 - beta1 ** state["step"] + bias_correction2 = 1 - beta2 ** state["step"] + + # Decay the first and second moment running average coefficient + exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1) + exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2) + denom = (exp_avg_sq.sqrt() * (bias_correction2 ** -0.5)).add_( + group["eps"] + ) + + step_size = group["lr"] / bias_correction1 + target_rms = group["target_rms"] + weight_decay = group["weight_decay"] + + if p.numel() > 1: + # avoid applying this weight-decay on "scaling factors" + # (which are scalar). + is_above_target_rms = p.norm() > ( + target_rms * (p.numel() ** 0.5) + ) + p.mul_(1 - (weight_decay * is_above_target_rms)) + p.addcdiv_(exp_avg, denom, value=-step_size) + + return loss + + +class LRScheduler(object): + """ + Base-class for learning rate schedulers where the learning-rate depends on both the + batch and the epoch. + """ + + def __init__(self, optimizer: Optimizer, verbose: bool = False): + # Attach optimizer + if not isinstance(optimizer, Optimizer): + raise TypeError( + "{} is not an Optimizer".format(type(optimizer).__name__) + ) + self.optimizer = optimizer + self.verbose = verbose + + for group in optimizer.param_groups: + group.setdefault("initial_lr", group["lr"]) + + self.base_lrs = [ + group["initial_lr"] for group in optimizer.param_groups + ] + + self.epoch = 0 + self.batch = 0 + + def state_dict(self): + """Returns the state of the scheduler as a :class:`dict`. + + It contains an entry for every variable in self.__dict__ which + is not the optimizer. + """ + return { + "base_lrs": self.base_lrs, + "epoch": self.epoch, + "batch": self.batch, + } + + def load_state_dict(self, state_dict): + """Loads the schedulers state. + + Args: + state_dict (dict): scheduler state. Should be an object returned + from a call to :meth:`state_dict`. + """ + self.__dict__.update(state_dict) + + def get_last_lr(self) -> List[float]: + """Return last computed learning rate by current scheduler. Will be a list of float.""" + return self._last_lr + + def get_lr(self): + # Compute list of learning rates from self.epoch and self.batch and + # self.base_lrs; this must be overloaded by the user. + # e.g. return [some_formula(self.batch, self.epoch, base_lr) for base_lr in self.base_lrs ] + raise NotImplementedError + + def step_batch(self, batch: Optional[int] = None) -> None: + # Step the batch index, or just set it. If `batch` is specified, it + # must be the batch index from the start of training, i.e. summed over + # all epochs. + # You can call this in any order; if you don't provide 'batch', it should + # of course be called once per batch. + if batch is not None: + self.batch = batch + else: + self.batch = self.batch + 1 + self._set_lrs() + + def step_epoch(self, epoch: Optional[int] = None): + # Step the epoch index, or just set it. If you provide the 'epoch' arg, + # you should call this at the start of the epoch; if you don't provide the 'epoch' + # arg, you should call it at the end of the epoch. + if epoch is not None: + self.epoch = epoch + else: + self.epoch = self.epoch + 1 + self._set_lrs() + + def _set_lrs(self): + values = self.get_lr() + assert len(values) == len(self.optimizer.param_groups) + + for i, data in enumerate(zip(self.optimizer.param_groups, values)): + param_group, lr = data + param_group["lr"] = lr + self.print_lr(self.verbose, i, lr) + self._last_lr = [group["lr"] for group in self.optimizer.param_groups] + + def print_lr(self, is_verbose, group, lr): + """Display the current learning rate.""" + if is_verbose: + print( + f"Epoch={self.epoch}, batch={self.batch}: adjusting learning rate" + f" of group {group} to {lr:.4e}." + ) + + +class Eden(LRScheduler): + """ + Eden scheduler. + lr = initial_lr * (((batch**2 + lr_batches**2) / lr_batches**2) ** -0.25 * + (((epoch**2 + lr_epochs**2) / lr_epochs**2) ** -0.25)) + + E.g. suggest initial-lr = 0.003 (passed to optimizer). + + Args: + optimizer: the optimizer to change the learning rates on + lr_batches: the number of batches after which we start significantly + decreasing the learning rate, suggest 5000. + lr_epochs: the number of epochs after which we start significantly + decreasing the learning rate, suggest 6 if you plan to do e.g. + 20 to 40 epochs, but may need smaller number if dataset is huge + and you will do few epochs. + """ + + def __init__( + self, + optimizer: Optimizer, + lr_batches: Union[int, float], + lr_epochs: Union[int, float], + verbose: bool = False, + ): + super(Eden, self).__init__(optimizer, verbose) + self.lr_batches = lr_batches + self.lr_epochs = lr_epochs + + def get_lr(self): + factor = ( + (self.batch ** 2 + self.lr_batches ** 2) / self.lr_batches ** 2 + ) ** -0.25 * ( + ((self.epoch ** 2 + self.lr_epochs ** 2) / self.lr_epochs ** 2) + ** -0.25 + ) + return [x * factor for x in self.base_lrs] + + +def _test_eden(): + m = torch.nn.Linear(100, 100) + optim = Eve(m.parameters(), lr=0.003) + + scheduler = Eden(optim, lr_batches=30, lr_epochs=2, verbose=True) + + for epoch in range(10): + scheduler.step_epoch(epoch) # sets epoch to `epoch` + + for step in range(20): + x = torch.randn(200, 100).detach() + x.requires_grad = True + y = m(x) + dy = torch.randn(200, 100).detach() + f = (y * dy).sum() + f.backward() + + optim.step() + scheduler.step_batch() + optim.zero_grad() + print("last lr = ", scheduler.get_last_lr()) + print("state dict = ", scheduler.state_dict()) + + +if __name__ == "__main__": + _test_eden() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py new file mode 100644 index 000000000..f89d2963e --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py @@ -0,0 +1,702 @@ +# Copyright 2022 Xiaomi Corp. (authors: Daniel Povey) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +import collections +from itertools import repeat +from typing import Optional, Tuple + +import torch +import torch.nn as nn +from torch import Tensor + + +def _ntuple(n): + def parse(x): + if isinstance(x, collections.Iterable): + return x + return tuple(repeat(x, n)) + + return parse + + +_single = _ntuple(1) +_pair = _ntuple(2) + + +class ActivationBalancerFunction(torch.autograd.Function): + @staticmethod + def forward( + ctx, + x: Tensor, + channel_dim: int, + min_positive: float, # e.g. 0.05 + max_positive: float, # e.g. 0.95 + max_factor: float, # e.g. 0.01 + min_abs: float, # e.g. 0.2 + max_abs: float, # e.g. 100.0 + ) -> Tensor: + if x.requires_grad: + if channel_dim < 0: + channel_dim += x.ndim + sum_dims = [d for d in range(x.ndim) if d != channel_dim] + xgt0 = x > 0 + proportion_positive = torch.mean( + xgt0.to(x.dtype), dim=sum_dims, keepdim=True + ) + factor1 = ( + (min_positive - proportion_positive).relu() + * (max_factor / min_positive) + if min_positive != 0.0 + else 0.0 + ) + factor2 = ( + (proportion_positive - max_positive).relu() + * (max_factor / (max_positive - 1.0)) + if max_positive != 1.0 + else 0.0 + ) + factor = factor1 + factor2 + if isinstance(factor, float): + factor = torch.zeros_like(proportion_positive) + + mean_abs = torch.mean(x.abs(), dim=sum_dims, keepdim=True) + below_threshold = mean_abs < min_abs + above_threshold = mean_abs > max_abs + + ctx.save_for_backward( + factor, xgt0, below_threshold, above_threshold + ) + ctx.max_factor = max_factor + ctx.sum_dims = sum_dims + return x + + @staticmethod + def backward( + ctx, x_grad: Tensor + ) -> Tuple[Tensor, None, None, None, None, None, None]: + factor, xgt0, below_threshold, above_threshold = ctx.saved_tensors + dtype = x_grad.dtype + scale_factor = ( + (below_threshold.to(dtype) - above_threshold.to(dtype)) + * (xgt0.to(dtype) - 0.5) + * (ctx.max_factor * 2.0) + ) + + neg_delta_grad = x_grad.abs() * (factor + scale_factor) + return x_grad - neg_delta_grad, None, None, None, None, None, None + + +class BasicNorm(torch.nn.Module): + """ + This is intended to be a simpler, and hopefully cheaper, replacement for + LayerNorm. The observation this is based on, is that Transformer-type + networks, especially with pre-norm, sometimes seem to set one of the + feature dimensions to a large constant value (e.g. 50), which "defeats" + the LayerNorm because the output magnitude is then not strongly dependent + on the other (useful) features. Presumably the weight and bias of the + LayerNorm are required to allow it to do this. + + So the idea is to introduce this large constant value as an explicit + parameter, that takes the role of the "eps" in LayerNorm, so the network + doesn't have to do this trick. We make the "eps" learnable. + + Args: + num_channels: the number of channels, e.g. 512. + channel_dim: the axis/dimension corresponding to the channel, + interprted as an offset from the input's ndim if negative. + shis is NOT the num_channels; it should typically be one of + {-2, -1, 0, 1, 2, 3}. + eps: the initial "epsilon" that we add as ballast in: + scale = ((input_vec**2).mean() + epsilon)**-0.5 + Note: our epsilon is actually large, but we keep the name + to indicate the connection with conventional LayerNorm. + learn_eps: if true, we learn epsilon; if false, we keep it + at the initial value. + """ + + def __init__( + self, + num_channels: int, + channel_dim: int = -1, # CAUTION: see documentation. + eps: float = 0.25, + learn_eps: bool = True, + ) -> None: + super(BasicNorm, self).__init__() + self.num_channels = num_channels + self.channel_dim = channel_dim + if learn_eps: + self.eps = nn.Parameter(torch.tensor(eps).log().detach()) + else: + self.register_buffer("eps", torch.tensor(eps).log().detach()) + + def forward(self, x: Tensor) -> Tensor: + assert x.shape[self.channel_dim] == self.num_channels + scales = ( + torch.mean(x ** 2, dim=self.channel_dim, keepdim=True) + + self.eps.exp() + ) ** -0.5 + return x * scales + + +class ScaledLinear(nn.Linear): + """ + A modified version of nn.Linear where the parameters are scaled before + use, via: + weight = self.weight * self.weight_scale.exp() + bias = self.bias * self.bias_scale.exp() + + Args: + Accepts the standard args and kwargs that nn.Linear accepts + e.g. in_features, out_features, bias=False. + + initial_scale: you can override this if you want to increase + or decrease the initial magnitude of the module's output + (affects the initialization of weight_scale and bias_scale). + Another option, if you want to do something like this, is + to re-initialize the parameters. + initial_speed: this affects how fast the parameter will + learn near the start of training; you can set it to a + value less than one if you suspect that a module + is contributing to instability near the start of training. + Nnote: regardless of the use of this option, it's best to + use schedulers like Noam that have a warm-up period. + Alternatively you can set it to more than 1 if you want it to + initially train faster. Must be greater than 0. + """ + + def __init__( + self, + *args, + initial_scale: float = 1.0, + initial_speed: float = 1.0, + **kwargs + ): + super(ScaledLinear, self).__init__(*args, **kwargs) + initial_scale = torch.tensor(initial_scale).log() + self.weight_scale = nn.Parameter(initial_scale.clone().detach()) + if self.bias is not None: + self.bias_scale = nn.Parameter(initial_scale.clone().detach()) + else: + self.register_parameter("bias_scale", None) + + self._reset_parameters( + initial_speed + ) # Overrides the reset_parameters in nn.Linear + + def _reset_parameters(self, initial_speed: float): + std = 0.1 / initial_speed + a = (3 ** 0.5) * std + nn.init.uniform_(self.weight, -a, a) + if self.bias is not None: + nn.init.constant_(self.bias, 0.0) + fan_in = self.weight.shape[1] * self.weight[0][0].numel() + scale = fan_in ** -0.5 # 1/sqrt(fan_in) + with torch.no_grad(): + self.weight_scale += torch.tensor(scale / std).log() + + def get_weight(self): + return self.weight * self.weight_scale.exp() + + def get_bias(self): + return None if self.bias is None else self.bias * self.bias_scale.exp() + + def forward(self, input: Tensor) -> Tensor: + return torch.nn.functional.linear( + input, self.get_weight(), self.get_bias() + ) + + +class ScaledConv1d(nn.Conv1d): + # See docs for ScaledLinear + def __init__( + self, + *args, + initial_scale: float = 1.0, + initial_speed: float = 1.0, + **kwargs + ): + super(ScaledConv1d, self).__init__(*args, **kwargs) + initial_scale = torch.tensor(initial_scale).log() + self.weight_scale = nn.Parameter(initial_scale.clone().detach()) + if self.bias is not None: + self.bias_scale = nn.Parameter(initial_scale.clone().detach()) + else: + self.register_parameter("bias_scale", None) + self._reset_parameters( + initial_speed + ) # Overrides the reset_parameters in base class + + def _reset_parameters(self, initial_speed: float): + std = 0.1 / initial_speed + a = (3 ** 0.5) * std + nn.init.uniform_(self.weight, -a, a) + if self.bias is not None: + nn.init.constant_(self.bias, 0.0) + fan_in = self.weight.shape[1] * self.weight[0][0].numel() + scale = fan_in ** -0.5 # 1/sqrt(fan_in) + with torch.no_grad(): + self.weight_scale += torch.tensor(scale / std).log() + + def get_weight(self): + return self.weight * self.weight_scale.exp() + + def get_bias(self): + return None if self.bias is None else self.bias * self.bias_scale.exp() + + def forward(self, input: Tensor) -> Tensor: + F = torch.nn.functional + if self.padding_mode != "zeros": + return F.conv1d( + F.pad( + input, + self._reversed_padding_repeated_twice, + mode=self.padding_mode, + ), + self.get_weight(), + self.get_bias(), + self.stride, + _single(0), + self.dilation, + self.groups, + ) + return F.conv1d( + input, + self.get_weight(), + self.get_bias(), + self.stride, + self.padding, + self.dilation, + self.groups, + ) + + +class ScaledConv2d(nn.Conv2d): + # See docs for ScaledLinear + def __init__( + self, + *args, + initial_scale: float = 1.0, + initial_speed: float = 1.0, + **kwargs + ): + super(ScaledConv2d, self).__init__(*args, **kwargs) + initial_scale = torch.tensor(initial_scale).log() + self.weight_scale = nn.Parameter(initial_scale.clone().detach()) + if self.bias is not None: + self.bias_scale = nn.Parameter(initial_scale.clone().detach()) + else: + self.register_parameter("bias_scale", None) + self._reset_parameters( + initial_speed + ) # Overrides the reset_parameters in base class + + def _reset_parameters(self, initial_speed: float): + std = 0.1 / initial_speed + a = (3 ** 0.5) * std + nn.init.uniform_(self.weight, -a, a) + if self.bias is not None: + nn.init.constant_(self.bias, 0.0) + fan_in = self.weight.shape[1] * self.weight[0][0].numel() + scale = fan_in ** -0.5 # 1/sqrt(fan_in) + with torch.no_grad(): + self.weight_scale += torch.tensor(scale / std).log() + + def get_weight(self): + return self.weight * self.weight_scale.exp() + + def get_bias(self): + return None if self.bias is None else self.bias * self.bias_scale.exp() + + def _conv_forward(self, input, weight): + F = torch.nn.functional + if self.padding_mode != "zeros": + return F.conv2d( + F.pad( + input, + self._reversed_padding_repeated_twice, + mode=self.padding_mode, + ), + weight, + self.get_bias(), + self.stride, + _pair(0), + self.dilation, + self.groups, + ) + return F.conv2d( + input, + weight, + self.get_bias(), + self.stride, + self.padding, + self.dilation, + self.groups, + ) + + def forward(self, input: Tensor) -> Tensor: + return self._conv_forward(input, self.get_weight()) + + +class ActivationBalancer(torch.nn.Module): + """ + Modifies the backpropped derivatives of a function to try to encourage, for + each channel, that it is positive at least a proportion `threshold` of the + time. It does this by multiplying negative derivative values by up to + (1+max_factor), and positive derivative values by up to (1-max_factor), + interpolated from 1 at the threshold to those extremal values when none + of the inputs are positive. + + + Args: + channel_dim: the dimension/axis corresponding to the channel, e.g. + -1, 0, 1, 2; will be interpreted as an offset from x.ndim if negative. + min_positive: the minimum, per channel, of the proportion of the time + that (x > 0), below which we start to modify the derivatives. + max_positive: the maximum, per channel, of the proportion of the time + that (x > 0), above which we start to modify the derivatives. + max_factor: the maximum factor by which we modify the derivatives for + either the sign constraint or the magnitude constraint; + e.g. with max_factor=0.02, the the derivatives would be multiplied by + values in the range [0.98..1.02]. + min_abs: the minimum average-absolute-value per channel, which + we allow, before we start to modify the derivatives to prevent + this. + max_abs: the maximum average-absolute-value per channel, which + we allow, before we start to modify the derivatives to prevent + this. + """ + + def __init__( + self, + channel_dim: int, + min_positive: float = 0.05, + max_positive: float = 0.95, + max_factor: float = 0.01, + min_abs: float = 0.2, + max_abs: float = 100.0, + ): + super(ActivationBalancer, self).__init__() + self.channel_dim = channel_dim + self.min_positive = min_positive + self.max_positive = max_positive + self.max_factor = max_factor + self.min_abs = min_abs + self.max_abs = max_abs + + def forward(self, x: Tensor) -> Tensor: + return ActivationBalancerFunction.apply( + x, + self.channel_dim, + self.min_positive, + self.max_positive, + self.max_factor, + self.min_abs, + self.max_abs, + ) + + +class DoubleSwishFunction(torch.autograd.Function): + """ + double_swish(x) = x * torch.sigmoid(x-1) + This is a definition, originally motivated by its close numerical + similarity to swish(swish(x)), where swish(x) = x * sigmoid(x). + + Memory-efficient derivative computation: + double_swish(x) = x * s, where s(x) = torch.sigmoid(x-1) + double_swish'(x) = d/dx double_swish(x) = x * s'(x) + x' * s(x) = x * s'(x) + s(x). + Now, s'(x) = s(x) * (1-s(x)). + double_swish'(x) = x * s'(x) + s(x). + = x * s(x) * (1-s(x)) + s(x). + = double_swish(x) * (1-s(x)) + s(x) + ... so we just need to remember s(x) but not x itself. + """ + + @staticmethod + def forward(ctx, x: Tensor) -> Tensor: + x = x.detach() + s = torch.sigmoid(x - 1.0) + y = x * s + ctx.save_for_backward(s, y) + return y + + @staticmethod + def backward(ctx, y_grad: Tensor) -> Tensor: + s, y = ctx.saved_tensors + return (y * (1 - s) + s) * y_grad + + +class DoubleSwish(torch.nn.Module): + def forward(self, x: Tensor) -> Tensor: + """Return double-swish activation function which is an approximation to Swish(Swish(x)), + that we approximate closely with x * sigmoid(x-1). + """ + return DoubleSwishFunction.apply(x) + + +class ScaledEmbedding(nn.Module): + r"""This is a modified version of nn.Embedding that introduces a learnable scale + on the parameters. Note: due to how we initialize it, it's best used with + schedulers like Noam that have a warmup period. + + It is a simple lookup table that stores embeddings of a fixed dictionary and size. + + This module is often used to store word embeddings and retrieve them using indices. + The input to the module is a list of indices, and the output is the corresponding + word embeddings. + + Args: + num_embeddings (int): size of the dictionary of embeddings + embedding_dim (int): the size of each embedding vector + padding_idx (int, optional): If given, pads the output with the embedding vector at :attr:`padding_idx` + (initialized to zeros) whenever it encounters the index. + max_norm (float, optional): If given, each embedding vector with norm larger than :attr:`max_norm` + is renormalized to have norm :attr:`max_norm`. + norm_type (float, optional): The p of the p-norm to compute for the :attr:`max_norm` option. Default ``2``. + scale_grad_by_freq (boolean, optional): If given, this will scale gradients by the inverse of frequency of + the words in the mini-batch. Default ``False``. + sparse (bool, optional): If ``True``, gradient w.r.t. :attr:`weight` matrix will be a sparse tensor. + See Notes for more details regarding sparse gradients. + + initial_speed (float, optional): This affects how fast the parameter will + learn near the start of training; you can set it to a value less than + one if you suspect that a module is contributing to instability near + the start of training. Nnote: regardless of the use of this option, + it's best to use schedulers like Noam that have a warm-up period. + Alternatively you can set it to more than 1 if you want it to + initially train faster. Must be greater than 0. + + + Attributes: + weight (Tensor): the learnable weights of the module of shape (num_embeddings, embedding_dim) + initialized from :math:`\mathcal{N}(0, 1)` + + Shape: + - Input: :math:`(*)`, LongTensor of arbitrary shape containing the indices to extract + - Output: :math:`(*, H)`, where `*` is the input shape and :math:`H=\text{embedding\_dim}` + + .. note:: + Keep in mind that only a limited number of optimizers support + sparse gradients: currently it's :class:`optim.SGD` (`CUDA` and `CPU`), + :class:`optim.SparseAdam` (`CUDA` and `CPU`) and :class:`optim.Adagrad` (`CPU`) + + .. note:: + With :attr:`padding_idx` set, the embedding vector at + :attr:`padding_idx` is initialized to all zeros. However, note that this + vector can be modified afterwards, e.g., using a customized + initialization method, and thus changing the vector used to pad the + output. The gradient for this vector from :class:`~torch.nn.Embedding` + is always zero. + + Examples:: + + >>> # an Embedding module containing 10 tensors of size 3 + >>> embedding = nn.Embedding(10, 3) + >>> # a batch of 2 samples of 4 indices each + >>> input = torch.LongTensor([[1,2,4,5],[4,3,2,9]]) + >>> embedding(input) + tensor([[[-0.0251, -1.6902, 0.7172], + [-0.6431, 0.0748, 0.6969], + [ 1.4970, 1.3448, -0.9685], + [-0.3677, -2.7265, -0.1685]], + + [[ 1.4970, 1.3448, -0.9685], + [ 0.4362, -0.4004, 0.9400], + [-0.6431, 0.0748, 0.6969], + [ 0.9124, -2.3616, 1.1151]]]) + + + >>> # example with padding_idx + >>> embedding = nn.Embedding(10, 3, padding_idx=0) + >>> input = torch.LongTensor([[0,2,0,5]]) + >>> embedding(input) + tensor([[[ 0.0000, 0.0000, 0.0000], + [ 0.1535, -2.0309, 0.9315], + [ 0.0000, 0.0000, 0.0000], + [-0.1655, 0.9897, 0.0635]]]) + + """ + __constants__ = [ + "num_embeddings", + "embedding_dim", + "padding_idx", + "scale_grad_by_freq", + "sparse", + ] + + num_embeddings: int + embedding_dim: int + padding_idx: int + scale_grad_by_freq: bool + weight: Tensor + sparse: bool + + def __init__( + self, + num_embeddings: int, + embedding_dim: int, + padding_idx: Optional[int] = None, + scale_grad_by_freq: bool = False, + sparse: bool = False, + initial_speed: float = 1.0, + ) -> None: + super(ScaledEmbedding, self).__init__() + self.num_embeddings = num_embeddings + self.embedding_dim = embedding_dim + if padding_idx is not None: + if padding_idx > 0: + assert ( + padding_idx < self.num_embeddings + ), "Padding_idx must be within num_embeddings" + elif padding_idx < 0: + assert ( + padding_idx >= -self.num_embeddings + ), "Padding_idx must be within num_embeddings" + padding_idx = self.num_embeddings + padding_idx + self.padding_idx = padding_idx + self.scale_grad_by_freq = scale_grad_by_freq + + self.scale = nn.Parameter(torch.zeros(())) # see reset_parameters() + self.sparse = sparse + + self.weight = nn.Parameter(torch.Tensor(num_embeddings, embedding_dim)) + self.reset_parameters(initial_speed) + + def reset_parameters(self, initial_speed: float = 1.0) -> None: + std = 0.1 / initial_speed + nn.init.normal_(self.weight, std=std) + nn.init.constant_(self.scale, torch.tensor(1.0 / std).log()) + + if self.padding_idx is not None: + with torch.no_grad(): + self.weight[self.padding_idx].fill_(0) + + def forward(self, input: Tensor) -> Tensor: + F = torch.nn.functional + scale = self.scale.exp() + if input.numel() < self.num_embeddings: + return ( + F.embedding( + input, + self.weight, + self.padding_idx, + None, + 2.0, # None, 2.0 relate to normalization + self.scale_grad_by_freq, + self.sparse, + ) + * scale + ) + else: + return F.embedding( + input, + self.weight * scale, + self.padding_idx, + None, + 2.0, # None, 2.0 relates to normalization + self.scale_grad_by_freq, + self.sparse, + ) + + def extra_repr(self) -> str: + s = "{num_embeddings}, {embedding_dim}, scale={scale}" + if self.padding_idx is not None: + s += ", padding_idx={padding_idx}" + if self.scale_grad_by_freq is not False: + s += ", scale_grad_by_freq={scale_grad_by_freq}" + if self.sparse is not False: + s += ", sparse=True" + return s.format(**self.__dict__) + + +def _test_activation_balancer_sign(): + probs = torch.arange(0, 1, 0.01) + N = 1000 + x = 1.0 * (torch.rand(probs.numel(), N) < probs.unsqueeze(-1)) + x = x.detach() + x.requires_grad = True + m = ActivationBalancer( + channel_dim=0, + min_positive=0.05, + max_positive=0.95, + max_factor=0.2, + min_abs=0.0, + ) + + y_grad = torch.sign(torch.randn(probs.numel(), N)) + + y = m(x) + y.backward(gradient=y_grad) + print("_test_activation_balancer_sign: x = ", x) + print("_test_activation_balancer_sign: y grad = ", y_grad) + print("_test_activation_balancer_sign: x grad = ", x.grad) + + +def _test_activation_balancer_magnitude(): + magnitudes = torch.arange(0, 1, 0.01) + N = 1000 + x = torch.sign(torch.randn(magnitudes.numel(), N)) * magnitudes.unsqueeze( + -1 + ) + x = x.detach() + x.requires_grad = True + m = ActivationBalancer( + channel_dim=0, + min_positive=0.0, + max_positive=1.0, + max_factor=0.2, + min_abs=0.2, + max_abs=0.8, + ) + + y_grad = torch.sign(torch.randn(magnitudes.numel(), N)) + + y = m(x) + y.backward(gradient=y_grad) + print("_test_activation_balancer_magnitude: x = ", x) + print("_test_activation_balancer_magnitude: y grad = ", y_grad) + print("_test_activation_balancer_magnitude: x grad = ", x.grad) + + +def _test_basic_norm(): + num_channels = 128 + m = BasicNorm(num_channels=num_channels, channel_dim=1) + + x = torch.randn(500, num_channels) + + y = m(x) + + assert y.shape == x.shape + x_rms = (x ** 2).mean().sqrt() + y_rms = (y ** 2).mean().sqrt() + print("x rms = ", x_rms) + print("y rms = ", y_rms) + assert y_rms < x_rms + assert y_rms > 0.5 * x_rms + + +def _test_double_swish_deriv(): + x = torch.randn(10, 12, dtype=torch.double) * 0.5 + x.requires_grad = True + m = DoubleSwish() + torch.autograd.gradcheck(m, x) + + +if __name__ == "__main__": + _test_activation_balancer_sign() + _test_activation_balancer_magnitude() + _test_basic_norm() + _test_double_swish_deriv() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py new file mode 100755 index 000000000..4421ce2aa --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py @@ -0,0 +1,977 @@ +#!/usr/bin/env python3 +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang, +# Wei Kang +# Mingshuang Luo) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +""" +Usage: + +export CUDA_VISIBLE_DEVICES="0,1,2,3" + +./pruned_transducer_stateless2/train.py \ + --world-size 4 \ + --num-epochs 30 \ + --start-epoch 0 \ + --exp-dir pruned_transducer_stateless2/exp \ + --full-libri 1 \ + --max-duration 300 + +# For mix precision training: + +./pruned_transducer_stateless2/train.py \ + --world-size 4 \ + --num-epochs 30 \ + --start-epoch 0 \ + --use_fp16 1 \ + --exp-dir pruned_transducer_stateless2/exp \ + --full-libri 1 \ + --max-duration 550 + +""" + + +import argparse +import logging +import warnings +from pathlib import Path +from shutil import copyfile +from typing import Any, Dict, Optional, Tuple, Union + +import k2 +import optim +import sentencepiece as spm +import torch +import torch.multiprocessing as mp +import torch.nn as nn +from asr_datamodule import GigaSpeechAsrDataModule +from conformer import Conformer +from decoder import Decoder +from joiner import Joiner +from lhotse.dataset.sampling.base import CutSampler +from lhotse.utils import fix_random_seed +from model import Transducer +from optim import Eden, Eve +from torch import Tensor +from torch.cuda.amp import GradScaler +from torch.nn.parallel import DistributedDataParallel as DDP +from torch.utils.tensorboard import SummaryWriter + +from icefall import diagnostics +from icefall.checkpoint import load_checkpoint, remove_checkpoints +from icefall.checkpoint import save_checkpoint as save_checkpoint_impl +from icefall.checkpoint import save_checkpoint_with_global_batch_idx +from icefall.dist import cleanup_dist, setup_dist +from icefall.env import get_env_info +from icefall.utils import AttributeDict, MetricsTracker, setup_logger, str2bool + +LRSchedulerType = Union[ + torch.optim.lr_scheduler._LRScheduler, optim.LRScheduler +] + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--world-size", + type=int, + default=1, + help="Number of GPUs for DDP training.", + ) + + parser.add_argument( + "--master-port", + type=int, + default=12354, + help="Master port to use for DDP training.", + ) + + parser.add_argument( + "--tensorboard", + type=str2bool, + default=True, + help="Should various information be logged in tensorboard.", + ) + + parser.add_argument( + "--num-epochs", + type=int, + default=30, + help="Number of epochs to train.", + ) + + parser.add_argument( + "--start-epoch", + type=int, + default=0, + help="""Resume training from from this epoch. + If it is positive, it will load checkpoint from + transducer_stateless2/exp/epoch-{start_epoch-1}.pt + """, + ) + + parser.add_argument( + "--start-batch", + type=int, + default=0, + help="""If positive, --start-epoch is ignored and + it loads the checkpoint from exp-dir/checkpoint-{start_batch}.pt + """, + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="""The experiment dir. + It specifies the directory where all training related + files, e.g., checkpoints, log, etc, are saved + """, + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--initial-lr", + type=float, + default=0.003, + help="The initial learning rate. This value should not need to be changed.", + ) + + parser.add_argument( + "--lr-batches", + type=float, + default=5000, + help="""Number of steps that affects how rapidly the learning rate decreases. + We suggest not to change this.""", + ) + + parser.add_argument( + "--lr-epochs", + type=float, + default=6, + help="""Number of epochs that affects how rapidly the learning rate decreases. + """, + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + + parser.add_argument( + "--prune-range", + type=int, + default=5, + help="The prune range for rnnt loss, it means how many symbols(context)" + "we are using to compute the loss", + ) + + parser.add_argument( + "--lm-scale", + type=float, + default=0.25, + help="The scale to smooth the loss with lm " + "(output of prediction network) part.", + ) + + parser.add_argument( + "--am-scale", + type=float, + default=0.0, + help="The scale to smooth the loss with am (output of encoder network)" + "part.", + ) + + parser.add_argument( + "--simple-loss-scale", + type=float, + default=0.5, + help="To get pruning ranges, we will calculate a simple version" + "loss(joiner is just addition), this simple loss also uses for" + "training (as a regularization item). We will scale the simple loss" + "with this parameter before adding to the final loss.", + ) + + parser.add_argument( + "--seed", + type=int, + default=42, + help="The seed for random generators intended for reproducibility", + ) + + parser.add_argument( + "--print-diagnostics", + type=str2bool, + default=False, + help="Accumulate stats on activations, print them and exit.", + ) + + parser.add_argument( + "--save-every-n", + type=int, + default=8000, + help="""Save checkpoint after processing this number of batches" + periodically. We save checkpoint to exp-dir/ whenever + params.batch_idx_train % save_every_n == 0. The checkpoint filename + has the form: f'exp-dir/checkpoint-{params.batch_idx_train}.pt' + Note: It also saves checkpoint to `exp-dir/epoch-xxx.pt` at the + end of each epoch where `xxx` is the epoch number counting from 0. + """, + ) + + parser.add_argument( + "--keep-last-k", + type=int, + default=20, + help="""Only keep this number of checkpoints on disk. + For instance, if it is 3, there are only 3 checkpoints + in the exp-dir with filenames `checkpoint-xxx.pt`. + It does not affect checkpoints with name `epoch-xxx.pt`. + """, + ) + + parser.add_argument( + "--use-fp16", + type=str2bool, + default=False, + help="Whether to use half precision training.", + ) + + return parser + + +def get_params() -> AttributeDict: + """Return a dict containing training parameters. + + All training related parameters that are not passed from the commandline + are saved in the variable `params`. + + Commandline options are merged into `params` after they are parsed, so + you can also access them via `params`. + + Explanation of options saved in `params`: + + - best_train_loss: Best training loss so far. It is used to select + the model that has the lowest training loss. It is + updated during the training. + + - best_valid_loss: Best validation loss so far. It is used to select + the model that has the lowest validation loss. It is + updated during the training. + + - best_train_epoch: It is the epoch that has the best training loss. + + - best_valid_epoch: It is the epoch that has the best validation loss. + + - batch_idx_train: Used to writing statistics to tensorboard. It + contains number of batches trained so far across + epochs. + + - log_interval: Print training loss if batch_idx % log_interval` is 0 + + - reset_interval: Reset statistics if batch_idx % reset_interval is 0 + + - valid_interval: Run validation if batch_idx % valid_interval is 0 + + - feature_dim: The model input dim. It has to match the one used + in computing features. + + - subsampling_factor: The subsampling factor for the model. + + - encoder_dim: Hidden dim for multi-head attention model. + + - num_decoder_layers: Number of decoder layer of transformer decoder. + + - warm_step: The warm_step for Noam optimizer. + """ + params = AttributeDict( + { + "best_train_loss": float("inf"), + "best_valid_loss": float("inf"), + "best_train_epoch": -1, + "best_valid_epoch": -1, + "batch_idx_train": 0, + "log_interval": 500, + "reset_interval": 2000, + "valid_interval": 20000, + # parameters for conformer + "feature_dim": 80, + "subsampling_factor": 4, + "encoder_dim": 512, + "nhead": 8, + "dim_feedforward": 2048, + "num_encoder_layers": 12, + # parameters for decoder + "decoder_dim": 512, + # parameters for joiner + "joiner_dim": 512, + # parameters for Noam + "model_warm_step": 20000, # arg given to model, not for lrate + "env_info": get_env_info(), + } + ) + + return params + + +def get_encoder_model(params: AttributeDict) -> nn.Module: + # TODO: We can add an option to switch between Conformer and Transformer + encoder = Conformer( + num_features=params.feature_dim, + subsampling_factor=params.subsampling_factor, + d_model=params.encoder_dim, + nhead=params.nhead, + dim_feedforward=params.dim_feedforward, + num_encoder_layers=params.num_encoder_layers, + ) + return encoder + + +def get_decoder_model(params: AttributeDict) -> nn.Module: + decoder = Decoder( + vocab_size=params.vocab_size, + decoder_dim=params.decoder_dim, + blank_id=params.blank_id, + context_size=params.context_size, + ) + return decoder + + +def get_joiner_model(params: AttributeDict) -> nn.Module: + joiner = Joiner( + encoder_dim=params.encoder_dim, + decoder_dim=params.decoder_dim, + joiner_dim=params.joiner_dim, + vocab_size=params.vocab_size, + ) + return joiner + + +def get_transducer_model(params: AttributeDict) -> nn.Module: + encoder = get_encoder_model(params) + decoder = get_decoder_model(params) + joiner = get_joiner_model(params) + + model = Transducer( + encoder=encoder, + decoder=decoder, + joiner=joiner, + encoder_dim=params.encoder_dim, + decoder_dim=params.decoder_dim, + joiner_dim=params.joiner_dim, + vocab_size=params.vocab_size, + ) + return model + + +def load_checkpoint_if_available( + params: AttributeDict, + model: nn.Module, + optimizer: Optional[torch.optim.Optimizer] = None, + scheduler: Optional[LRSchedulerType] = None, +) -> Optional[Dict[str, Any]]: + """Load checkpoint from file. + + If params.start_batch is positive, it will load the checkpoint from + `params.exp_dir/checkpoint-{params.start_batch}.pt`. Otherwise, if + params.start_epoch is positive, it will load the checkpoint from + `params.start_epoch - 1`. + + Apart from loading state dict for `model` and `optimizer` it also updates + `best_train_epoch`, `best_train_loss`, `best_valid_epoch`, + and `best_valid_loss` in `params`. + + Args: + params: + The return value of :func:`get_params`. + model: + The training model. + optimizer: + The optimizer that we are using. + scheduler: + The scheduler that we are using. + Returns: + Return a dict containing previously saved training info. + """ + if params.start_batch > 0: + filename = params.exp_dir / f"checkpoint-{params.start_batch}.pt" + elif params.start_epoch > 0: + filename = params.exp_dir / f"epoch-{params.start_epoch-1}.pt" + else: + return None + + assert filename.is_file(), f"{filename} does not exist!" + + saved_params = load_checkpoint( + filename, + model=model, + optimizer=optimizer, + scheduler=scheduler, + ) + + keys = [ + "best_train_epoch", + "best_valid_epoch", + "batch_idx_train", + "best_train_loss", + "best_valid_loss", + ] + for k in keys: + params[k] = saved_params[k] + + if params.start_batch > 0: + if "cur_epoch" in saved_params: + params["start_epoch"] = saved_params["cur_epoch"] + + if "cur_batch_idx" in saved_params: + params["cur_batch_idx"] = saved_params["cur_batch_idx"] + + return saved_params + + +def save_checkpoint( + params: AttributeDict, + model: nn.Module, + optimizer: Optional[torch.optim.Optimizer] = None, + scheduler: Optional[LRSchedulerType] = None, + sampler: Optional[CutSampler] = None, + scaler: Optional[GradScaler] = None, + rank: int = 0, +) -> None: + """Save model, optimizer, scheduler and training stats to file. + + Args: + params: + It is returned by :func:`get_params`. + model: + The training model. + optimizer: + The optimizer used in the training. + sampler: + The sampler for the training dataset. + scaler: + The scaler used for mix precision training. + """ + if rank != 0: + return + filename = params.exp_dir / f"epoch-{params.cur_epoch}.pt" + save_checkpoint_impl( + filename=filename, + model=model, + params=params, + optimizer=optimizer, + scheduler=scheduler, + sampler=sampler, + scaler=scaler, + rank=rank, + ) + + if params.best_train_epoch == params.cur_epoch: + best_train_filename = params.exp_dir / "best-train-loss.pt" + copyfile(src=filename, dst=best_train_filename) + + if params.best_valid_epoch == params.cur_epoch: + best_valid_filename = params.exp_dir / "best-valid-loss.pt" + copyfile(src=filename, dst=best_valid_filename) + + +def compute_loss( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + batch: dict, + is_training: bool, + warmup: float = 1.0, +) -> Tuple[Tensor, MetricsTracker]: + """ + Compute CTC loss given the model and its inputs. + + Args: + params: + Parameters for training. See :func:`get_params`. + model: + The model for training. It is an instance of Conformer in our case. + batch: + A batch of data. See `lhotse.dataset.K2SpeechRecognitionDataset()` + for the content in it. + is_training: + True for training. False for validation. When it is True, this + function enables autograd during computation; when it is False, it + disables autograd. + warmup: a floating point value which increases throughout training; + values >= 1.0 are fully warmed up and have all modules present. + """ + device = model.device + feature = batch["inputs"] + # at entry, feature is (N, T, C) + assert feature.ndim == 3 + feature = feature.to(device) + + supervisions = batch["supervisions"] + feature_lens = supervisions["num_frames"].to(device) + + texts = batch["supervisions"]["text"] + y = sp.encode(texts, out_type=int) + y = k2.RaggedTensor(y).to(device) + + with torch.set_grad_enabled(is_training): + simple_loss, pruned_loss = model( + x=feature, + x_lens=feature_lens, + y=y, + prune_range=params.prune_range, + am_scale=params.am_scale, + lm_scale=params.lm_scale, + warmup=warmup, + ) + # after the main warmup step, we keep pruned_loss_scale small + # for the same amount of time (model_warm_step), to avoid + # overwhelming the simple_loss and causing it to diverge, + # in case it had not fully learned the alignment yet. + pruned_loss_scale = ( + 0.0 + if warmup < 1.0 + else (0.1 if warmup > 1.0 and warmup < 2.0 else 1.0) + ) + loss = ( + params.simple_loss_scale * simple_loss + + pruned_loss_scale * pruned_loss + ) + + assert loss.requires_grad == is_training + + info = MetricsTracker() + with warnings.catch_warnings(): + warnings.simplefilter("ignore") + info["frames"] = ( + (feature_lens // params.subsampling_factor).sum().item() + ) + + # Note: We use reduction=sum while computing the loss. + info["loss"] = loss.detach().cpu().item() + info["simple_loss"] = simple_loss.detach().cpu().item() + info["pruned_loss"] = pruned_loss.detach().cpu().item() + + return loss, info + + +def compute_validation_loss( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + valid_dl: torch.utils.data.DataLoader, + world_size: int = 1, +) -> MetricsTracker: + """Run the validation process.""" + model.eval() + + tot_loss = MetricsTracker() + + for batch_idx, batch in enumerate(valid_dl): + loss, loss_info = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=False, + ) + assert loss.requires_grad is False + tot_loss = tot_loss + loss_info + + if world_size > 1: + tot_loss.reduce(loss.device) + + loss_value = tot_loss["loss"] / tot_loss["frames"] + if loss_value < params.best_valid_loss: + params.best_valid_epoch = params.cur_epoch + params.best_valid_loss = loss_value + + return tot_loss + + +def train_one_epoch( + params: AttributeDict, + model: nn.Module, + optimizer: torch.optim.Optimizer, + scheduler: LRSchedulerType, + sp: spm.SentencePieceProcessor, + train_dl: torch.utils.data.DataLoader, + valid_dl: torch.utils.data.DataLoader, + scaler: GradScaler, + tb_writer: Optional[SummaryWriter] = None, + world_size: int = 1, + rank: int = 0, +) -> None: + """Train the model for one epoch. + + The training loss from the mean of all frames is saved in + `params.train_loss`. It runs the validation process every + `params.valid_interval` batches. + + Args: + params: + It is returned by :func:`get_params`. + model: + The model for training. + optimizer: + The optimizer we are using. + scheduler: + The learning rate scheduler, we call step() every step. + train_dl: + Dataloader for the training dataset. + valid_dl: + Dataloader for the validation dataset. + scaler: + The scaler used for mix precision training. + tb_writer: + Writer to write log messages to tensorboard. + world_size: + Number of nodes in DDP training. If it is 1, DDP is disabled. + rank: + The rank of the node in DDP training. If no DDP is used, it should + be set to 0. + """ + model.train() + + tot_loss = MetricsTracker() + + cur_batch_idx = params.get("cur_batch_idx", 0) + + for batch_idx, batch in enumerate(train_dl): + if batch_idx < cur_batch_idx: + continue + cur_batch_idx = batch_idx + + params.batch_idx_train += 1 + batch_size = len(batch["supervisions"]["text"]) + + with torch.cuda.amp.autocast(enabled=params.use_fp16): + loss, loss_info = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=True, + warmup=(params.batch_idx_train / params.model_warm_step), + ) + # summary stats + tot_loss = (tot_loss * (1 - 1 / params.reset_interval)) + loss_info + + # NOTE: We use reduction==sum and loss is computed over utterances + # in the batch and there is no normalization to it so far. + scaler.scale(loss).backward() + scheduler.step_batch(params.batch_idx_train) + scaler.step(optimizer) + scaler.update() + optimizer.zero_grad() + + if params.print_diagnostics and batch_idx == 5: + return + + if ( + params.batch_idx_train > 0 + and params.batch_idx_train % params.save_every_n == 0 + ): + params.cur_batch_idx = batch_idx + save_checkpoint_with_global_batch_idx( + out_dir=params.exp_dir, + global_batch_idx=params.batch_idx_train, + model=model, + params=params, + optimizer=optimizer, + scheduler=scheduler, + sampler=train_dl.sampler, + scaler=scaler, + rank=rank, + ) + del params.cur_batch_idx + remove_checkpoints( + out_dir=params.exp_dir, + topk=params.keep_last_k, + rank=rank, + ) + + if batch_idx % params.log_interval == 0: + cur_lr = scheduler.get_last_lr()[0] + logging.info( + f"Epoch {params.cur_epoch}, " + f"batch {batch_idx}, loss[{loss_info}], " + f"tot_loss[{tot_loss}], batch size: {batch_size}, " + f"lr: {cur_lr:.2e}" + ) + + if tb_writer is not None: + tb_writer.add_scalar( + "train/learning_rate", cur_lr, params.batch_idx_train + ) + + loss_info.write_summary( + tb_writer, "train/current_", params.batch_idx_train + ) + tot_loss.write_summary( + tb_writer, "train/tot_", params.batch_idx_train + ) + + if batch_idx > 0 and batch_idx % params.valid_interval == 0: + logging.info("Computing validation loss") + valid_info = compute_validation_loss( + params=params, + model=model, + sp=sp, + valid_dl=valid_dl, + world_size=world_size, + ) + model.train() + logging.info(f"Epoch {params.cur_epoch}, validation: {valid_info}") + if tb_writer is not None: + valid_info.write_summary( + tb_writer, "train/valid_", params.batch_idx_train + ) + + loss_value = tot_loss["loss"] / tot_loss["frames"] + params.train_loss = loss_value + if params.train_loss < params.best_train_loss: + params.best_train_epoch = params.cur_epoch + params.best_train_loss = params.train_loss + + +def run(rank, world_size, args): + """ + Args: + rank: + It is a value between 0 and `world_size-1`, which is + passed automatically by `mp.spawn()` in :func:`main`. + The node with rank 0 is responsible for saving checkpoint. + world_size: + Number of GPUs for DDP training. + args: + The return value of get_parser().parse_args() + """ + params = get_params() + params.update(vars(args)) + + fix_random_seed(params.seed) + if world_size > 1: + setup_dist(rank, world_size, params.master_port) + + setup_logger(f"{params.exp_dir}/log/log-train") + logging.info("Training started") + + if args.tensorboard and rank == 0: + tb_writer = SummaryWriter(log_dir=f"{params.exp_dir}/tensorboard") + else: + tb_writer = None + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", rank) + logging.info(f"Device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + logging.info(f"Number of model parameters: {num_param}") + + checkpoints = load_checkpoint_if_available(params=params, model=model) + + model.to(device) + if world_size > 1: + logging.info("Using DDP") + model = DDP(model, device_ids=[rank]) + model.device = device + + optimizer = Eve(model.parameters(), lr=params.initial_lr) + + scheduler = Eden(optimizer, params.lr_batches, params.lr_epochs) + + if checkpoints and "optimizer" in checkpoints: + logging.info("Loading optimizer state dict") + optimizer.load_state_dict(checkpoints["optimizer"]) + + if ( + checkpoints + and "scheduler" in checkpoints + and checkpoints["scheduler"] is not None + ): + logging.info("Loading scheduler state dict") + scheduler.load_state_dict(checkpoints["scheduler"]) + + if params.print_diagnostics: + opts = diagnostics.TensorDiagnosticOptions( + 2 ** 22 + ) # allow 4 megabytes per sub-module + diagnostic = diagnostics.attach_diagnostics(model, opts) + + gigaspeech = GigaSpeechAsrDataModule(args) + + train_cuts = gigaspeech.train_cuts() + + if params.start_batch > 0 and checkpoints and "sampler" in checkpoints: + # We only load the sampler's state dict when it loads a checkpoint + # saved in the middle of an epoch + sampler_state_dict = checkpoints["sampler"] + else: + sampler_state_dict = None + + train_dl = gigaspeech.train_dataloaders( + train_cuts, sampler_state_dict=sampler_state_dict + ) + + valid_cuts = gigaspeech.dev_cuts() + valid_dl = gigaspeech.valid_dataloaders(valid_cuts) + + if not params.print_diagnostics: + scan_pessimistic_batches_for_oom( + model=model, + train_dl=train_dl, + optimizer=optimizer, + sp=sp, + params=params, + ) + + scaler = GradScaler(enabled=params.use_fp16) + if checkpoints and "grad_scaler" in checkpoints: + logging.info("Loading grad scaler state dict") + scaler.load_state_dict(checkpoints["grad_scaler"]) + + for epoch in range(params.start_epoch, params.num_epochs): + scheduler.step_epoch(epoch) + fix_random_seed(params.seed + epoch) + train_dl.sampler.set_epoch(epoch) + + if tb_writer is not None: + tb_writer.add_scalar("train/epoch", epoch, params.batch_idx_train) + + params.cur_epoch = epoch + + train_one_epoch( + params=params, + model=model, + optimizer=optimizer, + scheduler=scheduler, + sp=sp, + train_dl=train_dl, + valid_dl=valid_dl, + scaler=scaler, + tb_writer=tb_writer, + world_size=world_size, + rank=rank, + ) + + if params.print_diagnostics: + diagnostic.print_diagnostics() + break + + save_checkpoint( + params=params, + model=model, + optimizer=optimizer, + scheduler=scheduler, + sampler=train_dl.sampler, + scaler=scaler, + rank=rank, + ) + + logging.info("Done!") + + if world_size > 1: + torch.distributed.barrier() + cleanup_dist() + + +def scan_pessimistic_batches_for_oom( + model: nn.Module, + train_dl: torch.utils.data.DataLoader, + optimizer: torch.optim.Optimizer, + sp: spm.SentencePieceProcessor, + params: AttributeDict, +): + from lhotse.dataset import find_pessimistic_batches + + logging.info( + "Sanity check -- see if any of the batches in epoch 0 would cause OOM." + ) + batches, crit_values = find_pessimistic_batches(train_dl.sampler) + for criterion, cuts in batches.items(): + batch = train_dl.dataset[cuts] + try: + # warmup = 0.0 is so that the derivs for the pruned loss stay zero + # (i.e. are not remembered by the decaying-average in adam), because + # we want to avoid these params being subject to shrinkage in adam. + with torch.cuda.amp.autocast(enabled=params.use_fp16): + loss, _ = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=True, + warmup=0.0, + ) + loss.backward() + optimizer.step() + optimizer.zero_grad() + except RuntimeError as e: + if "CUDA out of memory" in str(e): + logging.error( + "Your GPU ran out of memory with the current " + "max_duration setting. We recommend decreasing " + "max_duration and trying again.\n" + f"Failing criterion: {criterion} " + f"(={crit_values[criterion]}) ..." + ) + raise + + +def main(): + parser = get_parser() + GigaSpeechAsrDataModule.add_arguments(parser) + args = parser.parse_args() + args.exp_dir = Path(args.exp_dir) + + world_size = args.world_size + assert world_size >= 1 + if world_size > 1: + mp.spawn(run, args=(world_size, args), nprocs=world_size, join=True) + else: + run(rank=0, world_size=1, args=args) + + +torch.set_num_threads(1) +torch.set_num_interop_threads(1) + +if __name__ == "__main__": + main() From e30e042c39746454f0aa54223bcdb8e7e1d58b09 Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Fri, 13 May 2022 13:03:16 +0800 Subject: [PATCH 03/16] Update decoding script for gigaspeech and remove duplicate files. (#361) --- README.md | 21 +- egs/gigaspeech/ASR/RESULTS.md | 8 +- .../beam_search.py | 767 +----------- .../pruned_transducer_stateless2/conformer.py | 1039 +---------------- .../pruned_transducer_stateless2/decode.py | 8 +- .../pruned_transducer_stateless2/decoder.py | 104 +- .../encoder_interface.py | 44 +- .../pruned_transducer_stateless2/joiner.py | 68 +- .../ASR/pruned_transducer_stateless2/model.py | 194 +-- .../ASR/pruned_transducer_stateless2/optim.py | 332 +----- .../pruned_transducer_stateless2/scaling.py | 703 +---------- 11 files changed, 37 insertions(+), 3251 deletions(-) mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py mode change 100644 => 120000 egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py diff --git a/README.md b/README.md index af4a22706..c4dad6aaf 100644 --- a/README.md +++ b/README.md @@ -12,13 +12,14 @@ for installation. Please refer to for more information. -We provide four recipes at present: +We provide 6 recipes at present: - [yesno][yesno] - [LibriSpeech][librispeech] - [Aishell][aishell] - [TIMIT][timit] - [TED-LIUM3][tedlium3] + - [GigaSpeech][gigaspeech] ### yesno @@ -197,6 +198,23 @@ The best WER using modified beam search with beam size 4 is: We provide a Colab notebook to run a pre-trained Pruned Transducer Stateless model: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1je_1zGrOkGVVd4WLzgkXRHxl-I27yWtz?usp=sharing) +### GigaSpeech + +#### Conformer CTC + +| | Dev | Test | +|-----|-------|-------| +| WER | 10.47 | 10.58 | + +#### Pruned stateless RNN-T + +| | Dev | Test | +|----------------------|-------|-------| +| greedy search | 10.59 | 10.87 | +| fast beam search | 10.56 | 10.80 | +| modified beam search | 10.52 | 10.62 | + + ## Deployment with C++ Once you have trained a model in icefall, you may want to deploy it with C++, @@ -225,4 +243,5 @@ Please see: [![Open In Colab](https://colab.research.google.com/assets/colab-bad [aishell]: egs/aishell/ASR [timit]: egs/timit/ASR [tedlium3]: egs/tedlium3/ASR +[gigaspeech]: egs/gigaspeech/ASR [k2]: https://github.com/k2-fsa/k2 diff --git a/egs/gigaspeech/ASR/RESULTS.md b/egs/gigaspeech/ASR/RESULTS.md index eaed9b9a0..de7b84202 100644 --- a/egs/gigaspeech/ASR/RESULTS.md +++ b/egs/gigaspeech/ASR/RESULTS.md @@ -5,10 +5,10 @@ #### Conformer encoder + embedding decoder -Conformer encoder + non-recurrent decoder. The encoder is a -reworked version of the conformer encoder, with many changes. The -decoder contains only an embedding layer, a Conv1d (with kernel -size 2) and a linear layer (to transform tensor dim). k2 pruned +Conformer encoder + non-recurrent decoder. The encoder is a +reworked version of the conformer encoder, with many changes. The +decoder contains only an embedding layer, a Conv1d (with kernel +size 2) and a linear layer (to transform tensor dim). k2 pruned RNN-T loss is used. Results are: diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py deleted file mode 100644 index 2e9bf3e0b..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py +++ /dev/null @@ -1,766 +0,0 @@ -# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -from dataclasses import dataclass -from typing import Dict, List, Optional - -import k2 -import torch -from model import Transducer - -from icefall.decode import one_best_decoding -from icefall.utils import get_texts - - -def fast_beam_search( - model: Transducer, - decoding_graph: k2.Fsa, - encoder_out: torch.Tensor, - encoder_out_lens: torch.Tensor, - beam: float, - max_states: int, - max_contexts: int, -) -> List[List[int]]: - """It limits the maximum number of symbols per frame to 1. - - Args: - model: - An instance of `Transducer`. - decoding_graph: - Decoding graph used for decoding, may be a TrivialGraph or a HLG. - encoder_out: - A tensor of shape (N, T, C) from the encoder. - encoder_out_lens: - A tensor of shape (N,) containing the number of frames in `encoder_out` - before padding. - beam: - Beam value, similar to the beam used in Kaldi.. - max_states: - Max states per stream per frame. - max_contexts: - Max contexts pre stream per frame. - Returns: - Return the decoded result. - """ - assert encoder_out.ndim == 3 - - context_size = model.decoder.context_size - vocab_size = model.decoder.vocab_size - - B, T, C = encoder_out.shape - - config = k2.RnntDecodingConfig( - vocab_size=vocab_size, - decoder_history_len=context_size, - beam=beam, - max_contexts=max_contexts, - max_states=max_states, - ) - individual_streams = [] - for i in range(B): - individual_streams.append(k2.RnntDecodingStream(decoding_graph)) - decoding_streams = k2.RnntDecodingStreams(individual_streams, config) - - encoder_out = model.joiner.encoder_proj(encoder_out) - - for t in range(T): - # shape is a RaggedShape of shape (B, context) - # contexts is a Tensor of shape (shape.NumElements(), context_size) - shape, contexts = decoding_streams.get_contexts() - # `nn.Embedding()` in torch below v1.7.1 supports only torch.int64 - contexts = contexts.to(torch.int64) - # decoder_out is of shape (shape.NumElements(), 1, decoder_out_dim) - decoder_out = model.decoder(contexts, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - # current_encoder_out is of shape - # (shape.NumElements(), 1, joiner_dim) - # fmt: off - current_encoder_out = torch.index_select( - encoder_out[:, t:t + 1, :], 0, shape.row_ids(1).to(torch.int64) - ) - # fmt: on - logits = model.joiner( - current_encoder_out.unsqueeze(2), - decoder_out.unsqueeze(1), - project_input=False, - ) - logits = logits.squeeze(1).squeeze(1) - log_probs = logits.log_softmax(dim=-1) - decoding_streams.advance(log_probs) - decoding_streams.terminate_and_flush_to_streams() - lattice = decoding_streams.format_output(encoder_out_lens.tolist()) - - best_path = one_best_decoding(lattice) - hyps = get_texts(best_path) - return hyps - - -def greedy_search( - model: Transducer, encoder_out: torch.Tensor, max_sym_per_frame: int -) -> List[int]: - """Greedy search for a single utterance. - Args: - model: - An instance of `Transducer`. - encoder_out: - A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. - max_sym_per_frame: - Maximum number of symbols per frame. If it is set to 0, the WER - would be 100%. - Returns: - Return the decoded result. - """ - assert encoder_out.ndim == 3 - - # support only batch_size == 1 for now - assert encoder_out.size(0) == 1, encoder_out.size(0) - - blank_id = model.decoder.blank_id - context_size = model.decoder.context_size - - device = model.device - - decoder_input = torch.tensor( - [blank_id] * context_size, device=device, dtype=torch.int64 - ).reshape(1, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - - encoder_out = model.joiner.encoder_proj(encoder_out) - - T = encoder_out.size(1) - t = 0 - hyp = [blank_id] * context_size - - # Maximum symbols per utterance. - max_sym_per_utt = 1000 - - # symbols per frame - sym_per_frame = 0 - - # symbols per utterance decoded so far - sym_per_utt = 0 - - while t < T and sym_per_utt < max_sym_per_utt: - if sym_per_frame >= max_sym_per_frame: - sym_per_frame = 0 - t += 1 - continue - - # fmt: off - current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) - # fmt: on - logits = model.joiner( - current_encoder_out, decoder_out.unsqueeze(1), project_input=False - ) - # logits is (1, 1, 1, vocab_size) - - y = logits.argmax().item() - if y != blank_id: - hyp.append(y) - decoder_input = torch.tensor( - [hyp[-context_size:]], device=device - ).reshape(1, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - - sym_per_utt += 1 - sym_per_frame += 1 - else: - sym_per_frame = 0 - t += 1 - hyp = hyp[context_size:] # remove blanks - - return hyp - - -def greedy_search_batch( - model: Transducer, encoder_out: torch.Tensor -) -> List[List[int]]: - """Greedy search in batch mode. It hardcodes --max-sym-per-frame=1. - Args: - model: - The transducer model. - encoder_out: - Output from the encoder. Its shape is (N, T, C), where N >= 1. - Returns: - Return a list-of-list of token IDs containing the decoded results. - len(ans) equals to encoder_out.size(0). - """ - assert encoder_out.ndim == 3 - assert encoder_out.size(0) >= 1, encoder_out.size(0) - - device = model.device - - batch_size = encoder_out.size(0) - T = encoder_out.size(1) - - blank_id = model.decoder.blank_id - context_size = model.decoder.context_size - - hyps = [[blank_id] * context_size for _ in range(batch_size)] - - decoder_input = torch.tensor( - hyps, - device=device, - dtype=torch.int64, - ) # (batch_size, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - encoder_out = model.joiner.encoder_proj(encoder_out) - - # decoder_out: (batch_size, 1, decoder_out_dim) - for t in range(T): - current_encoder_out = encoder_out[:, t : t + 1, :].unsqueeze(2) # noqa - # current_encoder_out's shape: (batch_size, 1, 1, encoder_out_dim) - logits = model.joiner( - current_encoder_out, decoder_out.unsqueeze(1), project_input=False - ) - # logits'shape (batch_size, 1, 1, vocab_size) - - logits = logits.squeeze(1).squeeze(1) # (batch_size, vocab_size) - assert logits.ndim == 2, logits.shape - y = logits.argmax(dim=1).tolist() - emitted = False - for i, v in enumerate(y): - if v != blank_id: - hyps[i].append(v) - emitted = True - if emitted: - # update decoder output - decoder_input = [h[-context_size:] for h in hyps] - decoder_input = torch.tensor( - decoder_input, - device=device, - dtype=torch.int64, - ) - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - - ans = [h[context_size:] for h in hyps] - return ans - - -@dataclass -class Hypothesis: - # The predicted tokens so far. - # Newly predicted tokens are appended to `ys`. - ys: List[int] - - # The log prob of ys. - # It contains only one entry. - log_prob: torch.Tensor - - @property - def key(self) -> str: - """Return a string representation of self.ys""" - return "_".join(map(str, self.ys)) - - -class HypothesisList(object): - def __init__(self, data: Optional[Dict[str, Hypothesis]] = None) -> None: - """ - Args: - data: - A dict of Hypotheses. Its key is its `value.key`. - """ - if data is None: - self._data = {} - else: - self._data = data - - @property - def data(self) -> Dict[str, Hypothesis]: - return self._data - - def add(self, hyp: Hypothesis) -> None: - """Add a Hypothesis to `self`. - - If `hyp` already exists in `self`, its probability is updated using - `log-sum-exp` with the existed one. - - Args: - hyp: - The hypothesis to be added. - """ - key = hyp.key - if key in self: - old_hyp = self._data[key] # shallow copy - torch.logaddexp( - old_hyp.log_prob, hyp.log_prob, out=old_hyp.log_prob - ) - else: - self._data[key] = hyp - - def get_most_probable(self, length_norm: bool = False) -> Hypothesis: - """Get the most probable hypothesis, i.e., the one with - the largest `log_prob`. - - Args: - length_norm: - If True, the `log_prob` of a hypothesis is normalized by the - number of tokens in it. - Returns: - Return the hypothesis that has the largest `log_prob`. - """ - if length_norm: - return max( - self._data.values(), key=lambda hyp: hyp.log_prob / len(hyp.ys) - ) - else: - return max(self._data.values(), key=lambda hyp: hyp.log_prob) - - def remove(self, hyp: Hypothesis) -> None: - """Remove a given hypothesis. - - Caution: - `self` is modified **in-place**. - - Args: - hyp: - The hypothesis to be removed from `self`. - Note: It must be contained in `self`. Otherwise, - an exception is raised. - """ - key = hyp.key - assert key in self, f"{key} does not exist" - del self._data[key] - - def filter(self, threshold: torch.Tensor) -> "HypothesisList": - """Remove all Hypotheses whose log_prob is less than threshold. - - Caution: - `self` is not modified. Instead, a new HypothesisList is returned. - - Returns: - Return a new HypothesisList containing all hypotheses from `self` - with `log_prob` being greater than the given `threshold`. - """ - ans = HypothesisList() - for _, hyp in self._data.items(): - if hyp.log_prob > threshold: - ans.add(hyp) # shallow copy - return ans - - def topk(self, k: int) -> "HypothesisList": - """Return the top-k hypothesis.""" - hyps = list(self._data.items()) - - hyps = sorted(hyps, key=lambda h: h[1].log_prob, reverse=True)[:k] - - ans = HypothesisList(dict(hyps)) - return ans - - def __contains__(self, key: str): - return key in self._data - - def __iter__(self): - return iter(self._data.values()) - - def __len__(self) -> int: - return len(self._data) - - def __str__(self) -> str: - s = [] - for key in self: - s.append(key) - return ", ".join(s) - - -def _get_hyps_shape(hyps: List[HypothesisList]) -> k2.RaggedShape: - """Return a ragged shape with axes [utt][num_hyps]. - - Args: - hyps: - len(hyps) == batch_size. It contains the current hypothesis for - each utterance in the batch. - Returns: - Return a ragged shape with 2 axes [utt][num_hyps]. Note that - the shape is on CPU. - """ - num_hyps = [len(h) for h in hyps] - - # torch.cumsum() is inclusive sum, so we put a 0 at the beginning - # to get exclusive sum later. - num_hyps.insert(0, 0) - - num_hyps = torch.tensor(num_hyps) - row_splits = torch.cumsum(num_hyps, dim=0, dtype=torch.int32) - ans = k2.ragged.create_ragged_shape2( - row_splits=row_splits, cached_tot_size=row_splits[-1].item() - ) - return ans - - -def modified_beam_search( - model: Transducer, - encoder_out: torch.Tensor, - beam: int = 4, -) -> List[List[int]]: - """Beam search in batch mode with --max-sym-per-frame=1 being hardcoded. - - Args: - model: - The transducer model. - encoder_out: - Output from the encoder. Its shape is (N, T, C). - beam: - Number of active paths during the beam search. - Returns: - Return a list-of-list of token IDs. ans[i] is the decoding results - for the i-th utterance. - """ - assert encoder_out.ndim == 3, encoder_out.shape - - batch_size = encoder_out.size(0) - T = encoder_out.size(1) - - blank_id = model.decoder.blank_id - context_size = model.decoder.context_size - device = model.device - B = [HypothesisList() for _ in range(batch_size)] - for i in range(batch_size): - B[i].add( - Hypothesis( - ys=[blank_id] * context_size, - log_prob=torch.zeros(1, dtype=torch.float32, device=device), - ) - ) - - encoder_out = model.joiner.encoder_proj(encoder_out) - - for t in range(T): - current_encoder_out = encoder_out[:, t : t + 1, :].unsqueeze(2) # noqa - # current_encoder_out's shape is (batch_size, 1, 1, encoder_out_dim) - - hyps_shape = _get_hyps_shape(B).to(device) - - A = [list(b) for b in B] - B = [HypothesisList() for _ in range(batch_size)] - - ys_log_probs = torch.cat( - [hyp.log_prob.reshape(1, 1) for hyps in A for hyp in hyps] - ) # (num_hyps, 1) - - decoder_input = torch.tensor( - [hyp.ys[-context_size:] for hyps in A for hyp in hyps], - device=device, - dtype=torch.int64, - ) # (num_hyps, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False).unsqueeze(1) - decoder_out = model.joiner.decoder_proj(decoder_out) - # decoder_out is of shape (num_hyps, 1, 1, joiner_dim) - - # Note: For torch 1.7.1 and below, it requires a torch.int64 tensor - # as index, so we use `to(torch.int64)` below. - current_encoder_out = torch.index_select( - current_encoder_out, - dim=0, - index=hyps_shape.row_ids(1).to(torch.int64), - ) # (num_hyps, 1, 1, encoder_out_dim) - - logits = model.joiner( - current_encoder_out, - decoder_out, - project_input=False, - ) # (num_hyps, 1, 1, vocab_size) - - logits = logits.squeeze(1).squeeze(1) # (num_hyps, vocab_size) - - log_probs = logits.log_softmax(dim=-1) # (num_hyps, vocab_size) - - log_probs.add_(ys_log_probs) - - vocab_size = log_probs.size(-1) - - log_probs = log_probs.reshape(-1) - - row_splits = hyps_shape.row_splits(1) * vocab_size - log_probs_shape = k2.ragged.create_ragged_shape2( - row_splits=row_splits, cached_tot_size=log_probs.numel() - ) - ragged_log_probs = k2.RaggedTensor( - shape=log_probs_shape, value=log_probs - ) - - for i in range(batch_size): - topk_log_probs, topk_indexes = ragged_log_probs[i].topk(beam) - - topk_hyp_indexes = (topk_indexes // vocab_size).tolist() - topk_token_indexes = (topk_indexes % vocab_size).tolist() - - for k in range(len(topk_hyp_indexes)): - hyp_idx = topk_hyp_indexes[k] - hyp = A[i][hyp_idx] - - new_ys = hyp.ys[:] - new_token = topk_token_indexes[k] - if new_token != blank_id: - new_ys.append(new_token) - - new_log_prob = topk_log_probs[k] - new_hyp = Hypothesis(ys=new_ys, log_prob=new_log_prob) - B[i].add(new_hyp) - - best_hyps = [b.get_most_probable(length_norm=True) for b in B] - ans = [h.ys[context_size:] for h in best_hyps] - - return ans - - -def _deprecated_modified_beam_search( - model: Transducer, - encoder_out: torch.Tensor, - beam: int = 4, -) -> List[int]: - """It limits the maximum number of symbols per frame to 1. - - It decodes only one utterance at a time. We keep it only for reference. - The function :func:`modified_beam_search` should be preferred as it - supports batch decoding. - - - Args: - model: - An instance of `Transducer`. - encoder_out: - A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. - beam: - Beam size. - Returns: - Return the decoded result. - """ - - assert encoder_out.ndim == 3 - - # support only batch_size == 1 for now - assert encoder_out.size(0) == 1, encoder_out.size(0) - blank_id = model.decoder.blank_id - context_size = model.decoder.context_size - - device = model.device - - T = encoder_out.size(1) - - B = HypothesisList() - B.add( - Hypothesis( - ys=[blank_id] * context_size, - log_prob=torch.zeros(1, dtype=torch.float32, device=device), - ) - ) - encoder_out = model.joiner.encoder_proj(encoder_out) - - for t in range(T): - # fmt: off - current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) - # current_encoder_out is of shape (1, 1, 1, encoder_out_dim) - # fmt: on - A = list(B) - B = HypothesisList() - - ys_log_probs = torch.cat([hyp.log_prob.reshape(1, 1) for hyp in A]) - # ys_log_probs is of shape (num_hyps, 1) - - decoder_input = torch.tensor( - [hyp.ys[-context_size:] for hyp in A], - device=device, - dtype=torch.int64, - ) - # decoder_input is of shape (num_hyps, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False).unsqueeze(1) - decoder_out = model.joiner.decoder_proj(decoder_out) - # decoder_output is of shape (num_hyps, 1, 1, joiner_dim) - - current_encoder_out = current_encoder_out.expand( - decoder_out.size(0), 1, 1, -1 - ) # (num_hyps, 1, 1, encoder_out_dim) - - logits = model.joiner( - current_encoder_out, - decoder_out, - project_input=False, - ) - # logits is of shape (num_hyps, 1, 1, vocab_size) - logits = logits.squeeze(1).squeeze(1) - - # now logits is of shape (num_hyps, vocab_size) - log_probs = logits.log_softmax(dim=-1) - - log_probs.add_(ys_log_probs) - - log_probs = log_probs.reshape(-1) - topk_log_probs, topk_indexes = log_probs.topk(beam) - - # topk_hyp_indexes are indexes into `A` - topk_hyp_indexes = topk_indexes // logits.size(-1) - topk_token_indexes = topk_indexes % logits.size(-1) - - topk_hyp_indexes = topk_hyp_indexes.tolist() - topk_token_indexes = topk_token_indexes.tolist() - - for i in range(len(topk_hyp_indexes)): - hyp = A[topk_hyp_indexes[i]] - new_ys = hyp.ys[:] - new_token = topk_token_indexes[i] - if new_token != blank_id: - new_ys.append(new_token) - new_log_prob = topk_log_probs[i] - new_hyp = Hypothesis(ys=new_ys, log_prob=new_log_prob) - B.add(new_hyp) - - best_hyp = B.get_most_probable(length_norm=True) - ys = best_hyp.ys[context_size:] # [context_size:] to remove blanks - - return ys - - -def beam_search( - model: Transducer, - encoder_out: torch.Tensor, - beam: int = 4, -) -> List[int]: - """ - It implements Algorithm 1 in https://arxiv.org/pdf/1211.3711.pdf - - espnet/nets/beam_search_transducer.py#L247 is used as a reference. - - Args: - model: - An instance of `Transducer`. - encoder_out: - A tensor of shape (N, T, C) from the encoder. Support only N==1 for now. - beam: - Beam size. - Returns: - Return the decoded result. - """ - assert encoder_out.ndim == 3 - - # support only batch_size == 1 for now - assert encoder_out.size(0) == 1, encoder_out.size(0) - blank_id = model.decoder.blank_id - context_size = model.decoder.context_size - - device = model.device - - decoder_input = torch.tensor( - [blank_id] * context_size, - device=device, - dtype=torch.int64, - ).reshape(1, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - - encoder_out = model.joiner.encoder_proj(encoder_out) - - T = encoder_out.size(1) - t = 0 - - B = HypothesisList() - B.add(Hypothesis(ys=[blank_id] * context_size, log_prob=0.0)) - - max_sym_per_utt = 20000 - - sym_per_utt = 0 - - decoder_cache: Dict[str, torch.Tensor] = {} - - while t < T and sym_per_utt < max_sym_per_utt: - # fmt: off - current_encoder_out = encoder_out[:, t:t+1, :].unsqueeze(2) - # fmt: on - A = B - B = HypothesisList() - - joint_cache: Dict[str, torch.Tensor] = {} - - # TODO(fangjun): Implement prefix search to update the `log_prob` - # of hypotheses in A - - while True: - y_star = A.get_most_probable() - A.remove(y_star) - - cached_key = y_star.key - - if cached_key not in decoder_cache: - decoder_input = torch.tensor( - [y_star.ys[-context_size:]], - device=device, - dtype=torch.int64, - ).reshape(1, context_size) - - decoder_out = model.decoder(decoder_input, need_pad=False) - decoder_out = model.joiner.decoder_proj(decoder_out) - decoder_cache[cached_key] = decoder_out - else: - decoder_out = decoder_cache[cached_key] - - cached_key += f"-t-{t}" - if cached_key not in joint_cache: - logits = model.joiner( - current_encoder_out, - decoder_out.unsqueeze(1), - project_input=False, - ) - - # TODO(fangjun): Scale the blank posterior - log_prob = logits.log_softmax(dim=-1) - # log_prob is (1, 1, 1, vocab_size) - log_prob = log_prob.squeeze() - # Now log_prob is (vocab_size,) - joint_cache[cached_key] = log_prob - else: - log_prob = joint_cache[cached_key] - - # First, process the blank symbol - skip_log_prob = log_prob[blank_id] - new_y_star_log_prob = y_star.log_prob + skip_log_prob - - # ys[:] returns a copy of ys - B.add(Hypothesis(ys=y_star.ys[:], log_prob=new_y_star_log_prob)) - - # Second, process other non-blank labels - values, indices = log_prob.topk(beam + 1) - for i, v in zip(indices.tolist(), values.tolist()): - if i == blank_id: - continue - new_ys = y_star.ys + [i] - new_log_prob = y_star.log_prob + v - A.add(Hypothesis(ys=new_ys, log_prob=new_log_prob)) - - # Check whether B contains more than "beam" elements more probable - # than the most probable in A - A_most_probable = A.get_most_probable() - - kept_B = B.filter(A_most_probable.log_prob) - - if len(kept_B) >= beam: - B = kept_B.topk(beam) - break - - t += 1 - - best_hyp = B.get_most_probable(length_norm=True) - ys = best_hyp.ys[context_size:] # [context_size:] to remove blanks - return ys diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py new file mode 120000 index 000000000..e24eca39f --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/beam_search.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/beam_search.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py deleted file mode 100644 index 257936b59..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py +++ /dev/null @@ -1,1038 +0,0 @@ -#!/usr/bin/env python3 -# Copyright (c) 2021 University of Chinese Academy of Sciences (author: Han Zhu) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -import copy -import math -import warnings -from typing import Optional, Tuple - -import torch -from encoder_interface import EncoderInterface -from scaling import ( - ActivationBalancer, - BasicNorm, - DoubleSwish, - ScaledConv1d, - ScaledConv2d, - ScaledLinear, -) -from torch import Tensor, nn - -from icefall.utils import make_pad_mask - - -class Conformer(EncoderInterface): - """ - Args: - num_features (int): Number of input features - subsampling_factor (int): subsampling factor of encoder (the convolution layers before transformers) - d_model (int): attention dimension, also the output dimension - nhead (int): number of head - dim_feedforward (int): feedforward dimention - num_encoder_layers (int): number of encoder layers - dropout (float): dropout rate - layer_dropout (float): layer-dropout rate. - cnn_module_kernel (int): Kernel size of convolution module - vgg_frontend (bool): whether to use vgg frontend. - """ - - def __init__( - self, - num_features: int, - subsampling_factor: int = 4, - d_model: int = 256, - nhead: int = 4, - dim_feedforward: int = 2048, - num_encoder_layers: int = 12, - dropout: float = 0.1, - layer_dropout: float = 0.075, - cnn_module_kernel: int = 31, - ) -> None: - super(Conformer, self).__init__() - - self.num_features = num_features - self.subsampling_factor = subsampling_factor - if subsampling_factor != 4: - raise NotImplementedError("Support only 'subsampling_factor=4'.") - - # self.encoder_embed converts the input of shape (N, T, num_features) - # to the shape (N, T//subsampling_factor, d_model). - # That is, it does two things simultaneously: - # (1) subsampling: T -> T//subsampling_factor - # (2) embedding: num_features -> d_model - self.encoder_embed = Conv2dSubsampling(num_features, d_model) - - self.encoder_pos = RelPositionalEncoding(d_model, dropout) - - encoder_layer = ConformerEncoderLayer( - d_model, - nhead, - dim_feedforward, - dropout, - layer_dropout, - cnn_module_kernel, - ) - self.encoder = ConformerEncoder(encoder_layer, num_encoder_layers) - - def forward( - self, x: torch.Tensor, x_lens: torch.Tensor, warmup: float = 1.0 - ) -> Tuple[torch.Tensor, torch.Tensor]: - """ - Args: - x: - The input tensor. Its shape is (batch_size, seq_len, feature_dim). - x_lens: - A tensor of shape (batch_size,) containing the number of frames in - `x` before padding. - warmup: - A floating point value that gradually increases from 0 throughout - training; when it is >= 1.0 we are "fully warmed up". It is used - to turn modules on sequentially. - Returns: - Return a tuple containing 2 tensors: - - embeddings: its shape is (batch_size, output_seq_len, d_model) - - lengths, a tensor of shape (batch_size,) containing the number - of frames in `embeddings` before padding. - """ - x = self.encoder_embed(x) - x, pos_emb = self.encoder_pos(x) - x = x.permute(1, 0, 2) # (N, T, C) -> (T, N, C) - - with warnings.catch_warnings(): - warnings.simplefilter("ignore") - # Caution: We assume the subsampling factor is 4! - lengths = ((x_lens - 1) // 2 - 1) // 2 - assert x.size(0) == lengths.max().item() - mask = make_pad_mask(lengths) - - x = self.encoder( - x, pos_emb, src_key_padding_mask=mask, warmup=warmup - ) # (T, N, C) - - x = x.permute(1, 0, 2) # (T, N, C) ->(N, T, C) - - return x, lengths - - -class ConformerEncoderLayer(nn.Module): - """ - ConformerEncoderLayer is made up of self-attn, feedforward and convolution networks. - See: "Conformer: Convolution-augmented Transformer for Speech Recognition" - - Args: - d_model: the number of expected features in the input (required). - nhead: the number of heads in the multiheadattention models (required). - dim_feedforward: the dimension of the feedforward network model (default=2048). - dropout: the dropout value (default=0.1). - cnn_module_kernel (int): Kernel size of convolution module. - - Examples:: - >>> encoder_layer = ConformerEncoderLayer(d_model=512, nhead=8) - >>> src = torch.rand(10, 32, 512) - >>> pos_emb = torch.rand(32, 19, 512) - >>> out = encoder_layer(src, pos_emb) - """ - - def __init__( - self, - d_model: int, - nhead: int, - dim_feedforward: int = 2048, - dropout: float = 0.1, - layer_dropout: float = 0.075, - cnn_module_kernel: int = 31, - ) -> None: - super(ConformerEncoderLayer, self).__init__() - - self.layer_dropout = layer_dropout - - self.d_model = d_model - - self.self_attn = RelPositionMultiheadAttention( - d_model, nhead, dropout=0.0 - ) - - self.feed_forward = nn.Sequential( - ScaledLinear(d_model, dim_feedforward), - ActivationBalancer(channel_dim=-1), - DoubleSwish(), - nn.Dropout(dropout), - ScaledLinear(dim_feedforward, d_model, initial_scale=0.25), - ) - - self.feed_forward_macaron = nn.Sequential( - ScaledLinear(d_model, dim_feedforward), - ActivationBalancer(channel_dim=-1), - DoubleSwish(), - nn.Dropout(dropout), - ScaledLinear(dim_feedforward, d_model, initial_scale=0.25), - ) - - self.conv_module = ConvolutionModule(d_model, cnn_module_kernel) - - self.norm_final = BasicNorm(d_model) - - # try to ensure the output is close to zero-mean (or at least, zero-median). - self.balancer = ActivationBalancer( - channel_dim=-1, min_positive=0.45, max_positive=0.55, max_abs=6.0 - ) - - self.dropout = nn.Dropout(dropout) - - def forward( - self, - src: Tensor, - pos_emb: Tensor, - src_mask: Optional[Tensor] = None, - src_key_padding_mask: Optional[Tensor] = None, - warmup: float = 1.0, - ) -> Tensor: - """ - Pass the input through the encoder layer. - - Args: - src: the sequence to the encoder layer (required). - pos_emb: Positional embedding tensor (required). - src_mask: the mask for the src sequence (optional). - src_key_padding_mask: the mask for the src keys per batch (optional). - warmup: controls selective bypass of of layers; if < 1.0, we will - bypass layers more frequently. - - Shape: - src: (S, N, E). - pos_emb: (N, 2*S-1, E) - src_mask: (S, S). - src_key_padding_mask: (N, S). - S is the source sequence length, N is the batch size, E is the feature number - """ - src_orig = src - - warmup_scale = min(0.1 + warmup, 1.0) - # alpha = 1.0 means fully use this encoder layer, 0.0 would mean - # completely bypass it. - if self.training: - alpha = ( - warmup_scale - if torch.rand(()).item() <= (1.0 - self.layer_dropout) - else 0.1 - ) - else: - alpha = 1.0 - - # macaron style feed forward module - src = src + self.dropout(self.feed_forward_macaron(src)) - - # multi-headed self-attention module - src_att = self.self_attn( - src, - src, - src, - pos_emb=pos_emb, - attn_mask=src_mask, - key_padding_mask=src_key_padding_mask, - )[0] - src = src + self.dropout(src_att) - - # convolution module - src = src + self.dropout(self.conv_module(src)) - - # feed forward module - src = src + self.dropout(self.feed_forward(src)) - - src = self.norm_final(self.balancer(src)) - - if alpha != 1.0: - src = alpha * src + (1 - alpha) * src_orig - - return src - - -class ConformerEncoder(nn.Module): - r"""ConformerEncoder is a stack of N encoder layers - - Args: - encoder_layer: an instance of the ConformerEncoderLayer() class (required). - num_layers: the number of sub-encoder-layers in the encoder (required). - - Examples:: - >>> encoder_layer = ConformerEncoderLayer(d_model=512, nhead=8) - >>> conformer_encoder = ConformerEncoder(encoder_layer, num_layers=6) - >>> src = torch.rand(10, 32, 512) - >>> pos_emb = torch.rand(32, 19, 512) - >>> out = conformer_encoder(src, pos_emb) - """ - - def __init__(self, encoder_layer: nn.Module, num_layers: int) -> None: - super().__init__() - self.layers = nn.ModuleList( - [copy.deepcopy(encoder_layer) for i in range(num_layers)] - ) - self.num_layers = num_layers - - def forward( - self, - src: Tensor, - pos_emb: Tensor, - mask: Optional[Tensor] = None, - src_key_padding_mask: Optional[Tensor] = None, - warmup: float = 1.0, - ) -> Tensor: - r"""Pass the input through the encoder layers in turn. - - Args: - src: the sequence to the encoder (required). - pos_emb: Positional embedding tensor (required). - mask: the mask for the src sequence (optional). - src_key_padding_mask: the mask for the src keys per batch (optional). - - Shape: - src: (S, N, E). - pos_emb: (N, 2*S-1, E) - mask: (S, S). - src_key_padding_mask: (N, S). - S is the source sequence length, T is the target sequence length, N is the batch size, E is the feature number - - """ - output = src - - for i, mod in enumerate(self.layers): - output = mod( - output, - pos_emb, - src_mask=mask, - src_key_padding_mask=src_key_padding_mask, - warmup=warmup, - ) - - return output - - -class RelPositionalEncoding(torch.nn.Module): - """Relative positional encoding module. - - See : Appendix B in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" - Modified from https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/transformer/embedding.py - - Args: - d_model: Embedding dimension. - dropout_rate: Dropout rate. - max_len: Maximum input length. - - """ - - def __init__( - self, d_model: int, dropout_rate: float, max_len: int = 5000 - ) -> None: - """Construct an PositionalEncoding object.""" - super(RelPositionalEncoding, self).__init__() - self.d_model = d_model - self.dropout = torch.nn.Dropout(p=dropout_rate) - self.pe = None - self.extend_pe(torch.tensor(0.0).expand(1, max_len)) - - def extend_pe(self, x: Tensor) -> None: - """Reset the positional encodings.""" - if self.pe is not None: - # self.pe contains both positive and negative parts - # the length of self.pe is 2 * input_len - 1 - if self.pe.size(1) >= x.size(1) * 2 - 1: - # Note: TorchScript doesn't implement operator== for torch.Device - if self.pe.dtype != x.dtype or str(self.pe.device) != str( - x.device - ): - self.pe = self.pe.to(dtype=x.dtype, device=x.device) - return - # Suppose `i` means to the position of query vecotr and `j` means the - # position of key vector. We use position relative positions when keys - # are to the left (i>j) and negative relative positions otherwise (i Tuple[Tensor, Tensor]: - """Add positional encoding. - - Args: - x (torch.Tensor): Input tensor (batch, time, `*`). - - Returns: - torch.Tensor: Encoded tensor (batch, time, `*`). - torch.Tensor: Encoded tensor (batch, 2*time-1, `*`). - - """ - self.extend_pe(x) - pos_emb = self.pe[ - :, - self.pe.size(1) // 2 - - x.size(1) - + 1 : self.pe.size(1) // 2 # noqa E203 - + x.size(1), - ] - return self.dropout(x), self.dropout(pos_emb) - - -class RelPositionMultiheadAttention(nn.Module): - r"""Multi-Head Attention layer with relative position encoding - - See reference: "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" - - Args: - embed_dim: total dimension of the model. - num_heads: parallel attention heads. - dropout: a Dropout layer on attn_output_weights. Default: 0.0. - - Examples:: - - >>> rel_pos_multihead_attn = RelPositionMultiheadAttention(embed_dim, num_heads) - >>> attn_output, attn_output_weights = multihead_attn(query, key, value, pos_emb) - """ - - def __init__( - self, - embed_dim: int, - num_heads: int, - dropout: float = 0.0, - ) -> None: - super(RelPositionMultiheadAttention, self).__init__() - self.embed_dim = embed_dim - self.num_heads = num_heads - self.dropout = dropout - self.head_dim = embed_dim // num_heads - assert ( - self.head_dim * num_heads == self.embed_dim - ), "embed_dim must be divisible by num_heads" - - self.in_proj = ScaledLinear(embed_dim, 3 * embed_dim, bias=True) - self.out_proj = ScaledLinear( - embed_dim, embed_dim, bias=True, initial_scale=0.25 - ) - - # linear transformation for positional encoding. - self.linear_pos = ScaledLinear(embed_dim, embed_dim, bias=False) - # these two learnable bias are used in matrix c and matrix d - # as described in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" Section 3.3 - self.pos_bias_u = nn.Parameter(torch.Tensor(num_heads, self.head_dim)) - self.pos_bias_v = nn.Parameter(torch.Tensor(num_heads, self.head_dim)) - self.pos_bias_u_scale = nn.Parameter(torch.zeros(()).detach()) - self.pos_bias_v_scale = nn.Parameter(torch.zeros(()).detach()) - self._reset_parameters() - - def _pos_bias_u(self): - return self.pos_bias_u * self.pos_bias_u_scale.exp() - - def _pos_bias_v(self): - return self.pos_bias_v * self.pos_bias_v_scale.exp() - - def _reset_parameters(self) -> None: - nn.init.normal_(self.pos_bias_u, std=0.01) - nn.init.normal_(self.pos_bias_v, std=0.01) - - def forward( - self, - query: Tensor, - key: Tensor, - value: Tensor, - pos_emb: Tensor, - key_padding_mask: Optional[Tensor] = None, - need_weights: bool = True, - attn_mask: Optional[Tensor] = None, - ) -> Tuple[Tensor, Optional[Tensor]]: - r""" - Args: - query, key, value: map a query and a set of key-value pairs to an output. - pos_emb: Positional embedding tensor - key_padding_mask: if provided, specified padding elements in the key will - be ignored by the attention. When given a binary mask and a value is True, - the corresponding value on the attention layer will be ignored. When given - a byte mask and a value is non-zero, the corresponding value on the attention - layer will be ignored - need_weights: output attn_output_weights. - attn_mask: 2D or 3D mask that prevents attention to certain positions. A 2D mask will be broadcasted for all - the batches while a 3D mask allows to specify a different mask for the entries of each batch. - - Shape: - - Inputs: - - query: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, E is - the embedding dimension. - - key: :math:`(S, N, E)`, where S is the source sequence length, N is the batch size, E is - the embedding dimension. - - value: :math:`(S, N, E)` where S is the source sequence length, N is the batch size, E is - the embedding dimension. - - pos_emb: :math:`(N, 2*L-1, E)` where L is the target sequence length, N is the batch size, E is - the embedding dimension. - - key_padding_mask: :math:`(N, S)` where N is the batch size, S is the source sequence length. - If a ByteTensor is provided, the non-zero positions will be ignored while the position - with the zero positions will be unchanged. If a BoolTensor is provided, the positions with the - value of ``True`` will be ignored while the position with the value of ``False`` will be unchanged. - - attn_mask: 2D mask :math:`(L, S)` where L is the target sequence length, S is the source sequence length. - 3D mask :math:`(N*num_heads, L, S)` where N is the batch size, L is the target sequence length, - S is the source sequence length. attn_mask ensure that position i is allowed to attend the unmasked - positions. If a ByteTensor is provided, the non-zero positions are not allowed to attend - while the zero positions will be unchanged. If a BoolTensor is provided, positions with ``True`` - is not allowed to attend while ``False`` values will be unchanged. If a FloatTensor - is provided, it will be added to the attention weight. - - - Outputs: - - attn_output: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, - E is the embedding dimension. - - attn_output_weights: :math:`(N, L, S)` where N is the batch size, - L is the target sequence length, S is the source sequence length. - """ - return self.multi_head_attention_forward( - query, - key, - value, - pos_emb, - self.embed_dim, - self.num_heads, - self.in_proj.get_weight(), - self.in_proj.get_bias(), - self.dropout, - self.out_proj.get_weight(), - self.out_proj.get_bias(), - training=self.training, - key_padding_mask=key_padding_mask, - need_weights=need_weights, - attn_mask=attn_mask, - ) - - def rel_shift(self, x: Tensor) -> Tensor: - """Compute relative positional encoding. - - Args: - x: Input tensor (batch, head, time1, 2*time1-1). - time1 means the length of query vector. - - Returns: - Tensor: tensor of shape (batch, head, time1, time2) - (note: time2 has the same value as time1, but it is for - the key, while time1 is for the query). - """ - (batch_size, num_heads, time1, n) = x.shape - assert n == 2 * time1 - 1 - # Note: TorchScript requires explicit arg for stride() - batch_stride = x.stride(0) - head_stride = x.stride(1) - time1_stride = x.stride(2) - n_stride = x.stride(3) - return x.as_strided( - (batch_size, num_heads, time1, time1), - (batch_stride, head_stride, time1_stride - n_stride, n_stride), - storage_offset=n_stride * (time1 - 1), - ) - - def multi_head_attention_forward( - self, - query: Tensor, - key: Tensor, - value: Tensor, - pos_emb: Tensor, - embed_dim_to_check: int, - num_heads: int, - in_proj_weight: Tensor, - in_proj_bias: Tensor, - dropout_p: float, - out_proj_weight: Tensor, - out_proj_bias: Tensor, - training: bool = True, - key_padding_mask: Optional[Tensor] = None, - need_weights: bool = True, - attn_mask: Optional[Tensor] = None, - ) -> Tuple[Tensor, Optional[Tensor]]: - r""" - Args: - query, key, value: map a query and a set of key-value pairs to an output. - pos_emb: Positional embedding tensor - embed_dim_to_check: total dimension of the model. - num_heads: parallel attention heads. - in_proj_weight, in_proj_bias: input projection weight and bias. - dropout_p: probability of an element to be zeroed. - out_proj_weight, out_proj_bias: the output projection weight and bias. - training: apply dropout if is ``True``. - key_padding_mask: if provided, specified padding elements in the key will - be ignored by the attention. This is an binary mask. When the value is True, - the corresponding value on the attention layer will be filled with -inf. - need_weights: output attn_output_weights. - attn_mask: 2D or 3D mask that prevents attention to certain positions. A 2D mask will be broadcasted for all - the batches while a 3D mask allows to specify a different mask for the entries of each batch. - - Shape: - Inputs: - - query: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, E is - the embedding dimension. - - key: :math:`(S, N, E)`, where S is the source sequence length, N is the batch size, E is - the embedding dimension. - - value: :math:`(S, N, E)` where S is the source sequence length, N is the batch size, E is - the embedding dimension. - - pos_emb: :math:`(N, 2*L-1, E)` or :math:`(1, 2*L-1, E)` where L is the target sequence - length, N is the batch size, E is the embedding dimension. - - key_padding_mask: :math:`(N, S)` where N is the batch size, S is the source sequence length. - If a ByteTensor is provided, the non-zero positions will be ignored while the zero positions - will be unchanged. If a BoolTensor is provided, the positions with the - value of ``True`` will be ignored while the position with the value of ``False`` will be unchanged. - - attn_mask: 2D mask :math:`(L, S)` where L is the target sequence length, S is the source sequence length. - 3D mask :math:`(N*num_heads, L, S)` where N is the batch size, L is the target sequence length, - S is the source sequence length. attn_mask ensures that position i is allowed to attend the unmasked - positions. If a ByteTensor is provided, the non-zero positions are not allowed to attend - while the zero positions will be unchanged. If a BoolTensor is provided, positions with ``True`` - are not allowed to attend while ``False`` values will be unchanged. If a FloatTensor - is provided, it will be added to the attention weight. - - Outputs: - - attn_output: :math:`(L, N, E)` where L is the target sequence length, N is the batch size, - E is the embedding dimension. - - attn_output_weights: :math:`(N, L, S)` where N is the batch size, - L is the target sequence length, S is the source sequence length. - """ - - tgt_len, bsz, embed_dim = query.size() - assert embed_dim == embed_dim_to_check - assert key.size(0) == value.size(0) and key.size(1) == value.size(1) - - head_dim = embed_dim // num_heads - assert ( - head_dim * num_heads == embed_dim - ), "embed_dim must be divisible by num_heads" - - scaling = float(head_dim) ** -0.5 - - if torch.equal(query, key) and torch.equal(key, value): - # self-attention - q, k, v = nn.functional.linear( - query, in_proj_weight, in_proj_bias - ).chunk(3, dim=-1) - - elif torch.equal(key, value): - # encoder-decoder attention - # This is inline in_proj function with in_proj_weight and in_proj_bias - _b = in_proj_bias - _start = 0 - _end = embed_dim - _w = in_proj_weight[_start:_end, :] - if _b is not None: - _b = _b[_start:_end] - q = nn.functional.linear(query, _w, _b) - - # This is inline in_proj function with in_proj_weight and in_proj_bias - _b = in_proj_bias - _start = embed_dim - _end = None - _w = in_proj_weight[_start:, :] - if _b is not None: - _b = _b[_start:] - k, v = nn.functional.linear(key, _w, _b).chunk(2, dim=-1) - - else: - # This is inline in_proj function with in_proj_weight and in_proj_bias - _b = in_proj_bias - _start = 0 - _end = embed_dim - _w = in_proj_weight[_start:_end, :] - if _b is not None: - _b = _b[_start:_end] - q = nn.functional.linear(query, _w, _b) - - # This is inline in_proj function with in_proj_weight and in_proj_bias - _b = in_proj_bias - _start = embed_dim - _end = embed_dim * 2 - _w = in_proj_weight[_start:_end, :] - if _b is not None: - _b = _b[_start:_end] - k = nn.functional.linear(key, _w, _b) - - # This is inline in_proj function with in_proj_weight and in_proj_bias - _b = in_proj_bias - _start = embed_dim * 2 - _end = None - _w = in_proj_weight[_start:, :] - if _b is not None: - _b = _b[_start:] - v = nn.functional.linear(value, _w, _b) - - if attn_mask is not None: - assert ( - attn_mask.dtype == torch.float32 - or attn_mask.dtype == torch.float64 - or attn_mask.dtype == torch.float16 - or attn_mask.dtype == torch.uint8 - or attn_mask.dtype == torch.bool - ), "Only float, byte, and bool types are supported for attn_mask, not {}".format( - attn_mask.dtype - ) - if attn_mask.dtype == torch.uint8: - warnings.warn( - "Byte tensor for attn_mask is deprecated. Use bool tensor instead." - ) - attn_mask = attn_mask.to(torch.bool) - - if attn_mask.dim() == 2: - attn_mask = attn_mask.unsqueeze(0) - if list(attn_mask.size()) != [1, query.size(0), key.size(0)]: - raise RuntimeError( - "The size of the 2D attn_mask is not correct." - ) - elif attn_mask.dim() == 3: - if list(attn_mask.size()) != [ - bsz * num_heads, - query.size(0), - key.size(0), - ]: - raise RuntimeError( - "The size of the 3D attn_mask is not correct." - ) - else: - raise RuntimeError( - "attn_mask's dimension {} is not supported".format( - attn_mask.dim() - ) - ) - # attn_mask's dim is 3 now. - - # convert ByteTensor key_padding_mask to bool - if ( - key_padding_mask is not None - and key_padding_mask.dtype == torch.uint8 - ): - warnings.warn( - "Byte tensor for key_padding_mask is deprecated. Use bool tensor instead." - ) - key_padding_mask = key_padding_mask.to(torch.bool) - - q = (q * scaling).contiguous().view(tgt_len, bsz, num_heads, head_dim) - k = k.contiguous().view(-1, bsz, num_heads, head_dim) - v = v.contiguous().view(-1, bsz * num_heads, head_dim).transpose(0, 1) - - src_len = k.size(0) - - if key_padding_mask is not None: - assert key_padding_mask.size(0) == bsz, "{} == {}".format( - key_padding_mask.size(0), bsz - ) - assert key_padding_mask.size(1) == src_len, "{} == {}".format( - key_padding_mask.size(1), src_len - ) - - q = q.transpose(0, 1) # (batch, time1, head, d_k) - - pos_emb_bsz = pos_emb.size(0) - assert pos_emb_bsz in (1, bsz) # actually it is 1 - p = self.linear_pos(pos_emb).view(pos_emb_bsz, -1, num_heads, head_dim) - p = p.transpose(1, 2) # (batch, head, 2*time1-1, d_k) - - q_with_bias_u = (q + self._pos_bias_u()).transpose( - 1, 2 - ) # (batch, head, time1, d_k) - - q_with_bias_v = (q + self._pos_bias_v()).transpose( - 1, 2 - ) # (batch, head, time1, d_k) - - # compute attention score - # first compute matrix a and matrix c - # as described in "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" Section 3.3 - k = k.permute(1, 2, 3, 0) # (batch, head, d_k, time2) - matrix_ac = torch.matmul( - q_with_bias_u, k - ) # (batch, head, time1, time2) - - # compute matrix b and matrix d - matrix_bd = torch.matmul( - q_with_bias_v, p.transpose(-2, -1) - ) # (batch, head, time1, 2*time1-1) - matrix_bd = self.rel_shift(matrix_bd) - - attn_output_weights = ( - matrix_ac + matrix_bd - ) # (batch, head, time1, time2) - - attn_output_weights = attn_output_weights.view( - bsz * num_heads, tgt_len, -1 - ) - - assert list(attn_output_weights.size()) == [ - bsz * num_heads, - tgt_len, - src_len, - ] - - if attn_mask is not None: - if attn_mask.dtype == torch.bool: - attn_output_weights.masked_fill_(attn_mask, float("-inf")) - else: - attn_output_weights += attn_mask - - if key_padding_mask is not None: - attn_output_weights = attn_output_weights.view( - bsz, num_heads, tgt_len, src_len - ) - attn_output_weights = attn_output_weights.masked_fill( - key_padding_mask.unsqueeze(1).unsqueeze(2), - float("-inf"), - ) - attn_output_weights = attn_output_weights.view( - bsz * num_heads, tgt_len, src_len - ) - - attn_output_weights = nn.functional.softmax(attn_output_weights, dim=-1) - attn_output_weights = nn.functional.dropout( - attn_output_weights, p=dropout_p, training=training - ) - - attn_output = torch.bmm(attn_output_weights, v) - assert list(attn_output.size()) == [bsz * num_heads, tgt_len, head_dim] - attn_output = ( - attn_output.transpose(0, 1) - .contiguous() - .view(tgt_len, bsz, embed_dim) - ) - attn_output = nn.functional.linear( - attn_output, out_proj_weight, out_proj_bias - ) - - if need_weights: - # average attention weights over heads - attn_output_weights = attn_output_weights.view( - bsz, num_heads, tgt_len, src_len - ) - return attn_output, attn_output_weights.sum(dim=1) / num_heads - else: - return attn_output, None - - -class ConvolutionModule(nn.Module): - """ConvolutionModule in Conformer model. - Modified from https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/conformer/convolution.py - - Args: - channels (int): The number of channels of conv layers. - kernel_size (int): Kernerl size of conv layers. - bias (bool): Whether to use bias in conv layers (default=True). - - """ - - def __init__( - self, channels: int, kernel_size: int, bias: bool = True - ) -> None: - """Construct an ConvolutionModule object.""" - super(ConvolutionModule, self).__init__() - # kernerl_size should be a odd number for 'SAME' padding - assert (kernel_size - 1) % 2 == 0 - - self.pointwise_conv1 = ScaledConv1d( - channels, - 2 * channels, - kernel_size=1, - stride=1, - padding=0, - bias=bias, - ) - - # after pointwise_conv1 we put x through a gated linear unit (nn.functional.glu). - # For most layers the normal rms value of channels of x seems to be in the range 1 to 4, - # but sometimes, for some reason, for layer 0 the rms ends up being very large, - # between 50 and 100 for different channels. This will cause very peaky and - # sparse derivatives for the sigmoid gating function, which will tend to make - # the loss function not learn effectively. (for most layers the average absolute values - # are in the range 0.5..9.0, and the average p(x>0), i.e. positive proportion, - # at the output of pointwise_conv1.output is around 0.35 to 0.45 for different - # layers, which likely breaks down as 0.5 for the "linear" half and - # 0.2 to 0.3 for the part that goes into the sigmoid. The idea is that if we - # constrain the rms values to a reasonable range via a constraint of max_abs=10.0, - # it will be in a better position to start learning something, i.e. to latch onto - # the correct range. - self.deriv_balancer1 = ActivationBalancer( - channel_dim=1, max_abs=10.0, min_positive=0.05, max_positive=1.0 - ) - - self.depthwise_conv = ScaledConv1d( - channels, - channels, - kernel_size, - stride=1, - padding=(kernel_size - 1) // 2, - groups=channels, - bias=bias, - ) - - self.deriv_balancer2 = ActivationBalancer( - channel_dim=1, min_positive=0.05, max_positive=1.0 - ) - - self.activation = DoubleSwish() - - self.pointwise_conv2 = ScaledConv1d( - channels, - channels, - kernel_size=1, - stride=1, - padding=0, - bias=bias, - initial_scale=0.25, - ) - - def forward(self, x: Tensor) -> Tensor: - """Compute convolution module. - - Args: - x: Input tensor (#time, batch, channels). - - Returns: - Tensor: Output tensor (#time, batch, channels). - - """ - # exchange the temporal dimension and the feature dimension - x = x.permute(1, 2, 0) # (#batch, channels, time). - - # GLU mechanism - x = self.pointwise_conv1(x) # (batch, 2*channels, time) - - x = self.deriv_balancer1(x) - x = nn.functional.glu(x, dim=1) # (batch, channels, time) - - # 1D Depthwise Conv - x = self.depthwise_conv(x) - - x = self.deriv_balancer2(x) - x = self.activation(x) - - x = self.pointwise_conv2(x) # (batch, channel, time) - - return x.permute(2, 0, 1) - - -class Conv2dSubsampling(nn.Module): - """Convolutional 2D subsampling (to 1/4 length). - - Convert an input of shape (N, T, idim) to an output - with shape (N, T', odim), where - T' = ((T-1)//2 - 1)//2, which approximates T' == T//4 - - It is based on - https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/transformer/subsampling.py # noqa - """ - - def __init__( - self, - in_channels: int, - out_channels: int, - layer1_channels: int = 8, - layer2_channels: int = 32, - layer3_channels: int = 128, - ) -> None: - """ - Args: - in_channels: - Number of channels in. The input shape is (N, T, in_channels). - Caution: It requires: T >=7, in_channels >=7 - out_channels - Output dim. The output shape is (N, ((T-1)//2 - 1)//2, out_channels) - layer1_channels: - Number of channels in layer1 - layer1_channels: - Number of channels in layer2 - """ - assert in_channels >= 7 - super().__init__() - - self.conv = nn.Sequential( - ScaledConv2d( - in_channels=1, - out_channels=layer1_channels, - kernel_size=3, - padding=1, - ), - ActivationBalancer(channel_dim=1), - DoubleSwish(), - ScaledConv2d( - in_channels=layer1_channels, - out_channels=layer2_channels, - kernel_size=3, - stride=2, - ), - ActivationBalancer(channel_dim=1), - DoubleSwish(), - ScaledConv2d( - in_channels=layer2_channels, - out_channels=layer3_channels, - kernel_size=3, - stride=2, - ), - ActivationBalancer(channel_dim=1), - DoubleSwish(), - ) - self.out = ScaledLinear( - layer3_channels * (((in_channels - 1) // 2 - 1) // 2), out_channels - ) - # set learn_eps=False because out_norm is preceded by `out`, and `out` - # itself has learned scale, so the extra degree of freedom is not - # needed. - self.out_norm = BasicNorm(out_channels, learn_eps=False) - # constrain median of output to be close to zero. - self.out_balancer = ActivationBalancer( - channel_dim=-1, min_positive=0.45, max_positive=0.55 - ) - - def forward(self, x: torch.Tensor) -> torch.Tensor: - """Subsample x. - - Args: - x: - Its shape is (N, T, idim). - - Returns: - Return a tensor of shape (N, ((T-1)//2 - 1)//2, odim) - """ - # On entry, x is (N, T, idim) - x = x.unsqueeze(1) # (N, T, idim) -> (N, 1, T, idim) i.e., (N, C, H, W) - x = self.conv(x) - # Now x is of shape (N, odim, ((T-1)//2 - 1)//2, ((idim-1)//2 - 1)//2) - b, c, t, f = x.size() - x = self.out(x.transpose(1, 2).contiguous().view(b, t, c * f)) - # Now x is of shape (N, ((T-1)//2 - 1))//2, odim) - x = self.out_norm(x) - x = self.out_balancer(x) - return x - - -if __name__ == "__main__": - feature_dim = 50 - c = Conformer(num_features=feature_dim, d_model=128, nhead=4) - batch_size = 5 - seq_len = 20 - # Just make sure the forward pass runs. - f = c( - torch.randn(batch_size, seq_len, feature_dim), - torch.full((batch_size,), seq_len, dtype=torch.int64), - warmup=0.5, - ) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py new file mode 120000 index 000000000..a65957180 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/conformer.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/conformer.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py index 076feb7a9..92a5b0b28 100755 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py @@ -69,7 +69,7 @@ import torch.nn as nn from asr_datamodule import GigaSpeechAsrDataModule from beam_search import ( beam_search, - fast_beam_search, + fast_beam_search_one_best, greedy_search, greedy_search_batch, modified_beam_search, @@ -263,7 +263,7 @@ def decode_one_batch( hyps = [] if params.decoding_method == "fast_beam_search": - hyp_tokens = fast_beam_search( + hyp_tokens = fast_beam_search_one_best( model=model, decoding_graph=decoding_graph, encoder_out=encoder_out, @@ -281,6 +281,7 @@ def decode_one_batch( hyp_tokens = greedy_search_batch( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, ) for hyp in sp.decode(hyp_tokens): hyps.append(hyp.split()) @@ -288,6 +289,7 @@ def decode_one_batch( hyp_tokens = modified_beam_search( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, beam=params.beam_size, ) for hyp in sp.decode(hyp_tokens): @@ -366,7 +368,7 @@ def decode_dataset( except TypeError: num_batches = "?" - log_interval = 100 + log_interval = 20 results = defaultdict(list) for batch_idx, batch in enumerate(dl): diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py deleted file mode 100644 index b6d94aaf1..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py +++ /dev/null @@ -1,103 +0,0 @@ -# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -import torch -import torch.nn as nn -import torch.nn.functional as F -from scaling import ScaledConv1d, ScaledEmbedding - - -class Decoder(nn.Module): - """This class modifies the stateless decoder from the following paper: - - RNN-transducer with stateless prediction network - https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9054419 - - It removes the recurrent connection from the decoder, i.e., the prediction - network. Different from the above paper, it adds an extra Conv1d - right after the embedding layer. - - TODO: Implement https://arxiv.org/pdf/2109.07513.pdf - """ - - def __init__( - self, - vocab_size: int, - decoder_dim: int, - blank_id: int, - context_size: int, - ): - """ - Args: - vocab_size: - Number of tokens of the modeling unit including blank. - decoder_dim: - Dimension of the input embedding, and of the decoder output. - blank_id: - The ID of the blank symbol. - context_size: - Number of previous words to use to predict the next word. - 1 means bigram; 2 means trigram. n means (n+1)-gram. - """ - super().__init__() - - self.embedding = ScaledEmbedding( - num_embeddings=vocab_size, - embedding_dim=decoder_dim, - padding_idx=blank_id, - ) - self.blank_id = blank_id - - assert context_size >= 1, context_size - self.context_size = context_size - self.vocab_size = vocab_size - if context_size > 1: - self.conv = ScaledConv1d( - in_channels=decoder_dim, - out_channels=decoder_dim, - kernel_size=context_size, - padding=0, - groups=decoder_dim, - bias=False, - ) - - def forward(self, y: torch.Tensor, need_pad: bool = True) -> torch.Tensor: - """ - Args: - y: - A 2-D tensor of shape (N, U). - need_pad: - True to left pad the input. Should be True during training. - False to not pad the input. Should be False during inference. - Returns: - Return a tensor of shape (N, U, decoder_dim). - """ - y = y.to(torch.int64) - embedding_out = self.embedding(y) - if self.context_size > 1: - embedding_out = embedding_out.permute(0, 2, 1) - if need_pad is True: - embedding_out = F.pad( - embedding_out, pad=(self.context_size - 1, 0) - ) - else: - # During inference time, there is no need to do extra padding - # as we only need one output - assert embedding_out.size(-1) == self.context_size - embedding_out = self.conv(embedding_out) - embedding_out = embedding_out.permute(0, 2, 1) - embedding_out = F.relu(embedding_out) - return embedding_out diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py new file mode 120000 index 000000000..722e1c894 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decoder.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/decoder.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py deleted file mode 100644 index 257facce4..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py +++ /dev/null @@ -1,43 +0,0 @@ -# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -from typing import Tuple - -import torch -import torch.nn as nn - - -class EncoderInterface(nn.Module): - def forward( - self, x: torch.Tensor, x_lens: torch.Tensor - ) -> Tuple[torch.Tensor, torch.Tensor]: - """ - Args: - x: - A tensor of shape (batch_size, input_seq_len, num_features) - containing the input features. - x_lens: - A tensor of shape (batch_size,) containing the number of frames - in `x` before padding. - Returns: - Return a tuple containing two tensors: - - encoder_out, a tensor of (batch_size, out_seq_len, output_dim) - containing unnormalized probabilities, i.e., the output of a - linear layer. - - encoder_out_lens, a tensor of shape (batch_size,) containing - the number of frames in `encoder_out` before padding. - """ - raise NotImplementedError("Please implement it in a subclass") diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py new file mode 120000 index 000000000..f58253127 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/encoder_interface.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/encoder_interface.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py deleted file mode 100644 index 35f75ed2a..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py +++ /dev/null @@ -1,67 +0,0 @@ -# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -import torch -import torch.nn as nn -from scaling import ScaledLinear - - -class Joiner(nn.Module): - def __init__( - self, - encoder_dim: int, - decoder_dim: int, - joiner_dim: int, - vocab_size: int, - ): - super().__init__() - - self.encoder_proj = ScaledLinear(encoder_dim, joiner_dim) - self.decoder_proj = ScaledLinear(decoder_dim, joiner_dim) - self.output_linear = ScaledLinear(joiner_dim, vocab_size) - - def forward( - self, - encoder_out: torch.Tensor, - decoder_out: torch.Tensor, - project_input: bool = True, - ) -> torch.Tensor: - """ - Args: - encoder_out: - Output from the encoder. Its shape is (N, T, s_range, C). - decoder_out: - Output from the decoder. Its shape is (N, T, s_range, C). - project_input: - If true, apply input projections encoder_proj and decoder_proj. - If this is false, it is the user's responsibility to do this - manually. - Returns: - Return a tensor of shape (N, T, s_range, C). - """ - assert encoder_out.ndim == decoder_out.ndim == 4 - assert encoder_out.shape[:-1] == decoder_out.shape[:-1] - - if project_input: - logit = self.encoder_proj(encoder_out) + self.decoder_proj( - decoder_out - ) - else: - logit = encoder_out + decoder_out - - logit = self.output_linear(torch.tanh(logit)) - - return logit diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py new file mode 120000 index 000000000..9052f3cbb --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/joiner.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/joiner.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py deleted file mode 100644 index 599bf2506..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py +++ /dev/null @@ -1,193 +0,0 @@ -# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang, Wei Kang) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - - -import k2 -import torch -import torch.nn as nn -from encoder_interface import EncoderInterface -from scaling import ScaledLinear - -from icefall.utils import add_sos - - -class Transducer(nn.Module): - """It implements https://arxiv.org/pdf/1211.3711.pdf - "Sequence Transduction with Recurrent Neural Networks" - """ - - def __init__( - self, - encoder: EncoderInterface, - decoder: nn.Module, - joiner: nn.Module, - encoder_dim: int, - decoder_dim: int, - joiner_dim: int, - vocab_size: int, - ): - """ - Args: - encoder: - It is the transcription network in the paper. Its accepts - two inputs: `x` of (N, T, encoder_dim) and `x_lens` of shape (N,). - It returns two tensors: `logits` of shape (N, T, encoder_dm) and - `logit_lens` of shape (N,). - decoder: - It is the prediction network in the paper. Its input shape - is (N, U) and its output shape is (N, U, decoder_dim). - It should contain one attribute: `blank_id`. - joiner: - It has two inputs with shapes: (N, T, encoder_dim) and (N, U, decoder_dim). - Its output shape is (N, T, U, vocab_size). Note that its output contains - unnormalized probs, i.e., not processed by log-softmax. - """ - super().__init__() - assert isinstance(encoder, EncoderInterface), type(encoder) - assert hasattr(decoder, "blank_id") - - self.encoder = encoder - self.decoder = decoder - self.joiner = joiner - - self.simple_am_proj = ScaledLinear( - encoder_dim, vocab_size, initial_speed=0.5 - ) - self.simple_lm_proj = ScaledLinear(decoder_dim, vocab_size) - - def forward( - self, - x: torch.Tensor, - x_lens: torch.Tensor, - y: k2.RaggedTensor, - prune_range: int = 5, - am_scale: float = 0.0, - lm_scale: float = 0.0, - warmup: float = 1.0, - ) -> torch.Tensor: - """ - Args: - x: - A 3-D tensor of shape (N, T, C). - x_lens: - A 1-D tensor of shape (N,). It contains the number of frames in `x` - before padding. - y: - A ragged tensor with 2 axes [utt][label]. It contains labels of each - utterance. - prune_range: - The prune range for rnnt loss, it means how many symbols(context) - we are considering for each frame to compute the loss. - am_scale: - The scale to smooth the loss with am (output of encoder network) - part - lm_scale: - The scale to smooth the loss with lm (output of predictor network) - part - warmup: - A value warmup >= 0 that determines which modules are active, values - warmup > 1 "are fully warmed up" and all modules will be active. - Returns: - Return the transducer loss. - - Note: - Regarding am_scale & lm_scale, it will make the loss-function one of - the form: - lm_scale * lm_probs + am_scale * am_probs + - (1-lm_scale-am_scale) * combined_probs - """ - assert x.ndim == 3, x.shape - assert x_lens.ndim == 1, x_lens.shape - assert y.num_axes == 2, y.num_axes - - assert x.size(0) == x_lens.size(0) == y.dim0 - - encoder_out, x_lens = self.encoder(x, x_lens, warmup=warmup) - assert torch.all(x_lens > 0) - - # Now for the decoder, i.e., the prediction network - row_splits = y.shape.row_splits(1) - y_lens = row_splits[1:] - row_splits[:-1] - - blank_id = self.decoder.blank_id - sos_y = add_sos(y, sos_id=blank_id) - - # sos_y_padded: [B, S + 1], start with SOS. - sos_y_padded = sos_y.pad(mode="constant", padding_value=blank_id) - - # decoder_out: [B, S + 1, decoder_dim] - decoder_out = self.decoder(sos_y_padded) - - # Note: y does not start with SOS - # y_padded : [B, S] - y_padded = y.pad(mode="constant", padding_value=0) - - y_padded = y_padded.to(torch.int64) - boundary = torch.zeros( - (x.size(0), 4), dtype=torch.int64, device=x.device - ) - boundary[:, 2] = y_lens - boundary[:, 3] = x_lens - - lm = self.simple_lm_proj(decoder_out) - am = self.simple_am_proj(encoder_out) - - with torch.cuda.amp.autocast(enabled=False): - simple_loss, (px_grad, py_grad) = k2.rnnt_loss_smoothed( - lm=lm.float(), - am=am.float(), - symbols=y_padded, - termination_symbol=blank_id, - lm_only_scale=lm_scale, - am_only_scale=am_scale, - boundary=boundary, - reduction="sum", - return_grad=True, - ) - - # ranges : [B, T, prune_range] - ranges = k2.get_rnnt_prune_ranges( - px_grad=px_grad, - py_grad=py_grad, - boundary=boundary, - s_range=prune_range, - ) - - # am_pruned : [B, T, prune_range, encoder_dim] - # lm_pruned : [B, T, prune_range, decoder_dim] - am_pruned, lm_pruned = k2.do_rnnt_pruning( - am=self.joiner.encoder_proj(encoder_out), - lm=self.joiner.decoder_proj(decoder_out), - ranges=ranges, - ) - - # logits : [B, T, prune_range, vocab_size] - - # project_input=False since we applied the decoder's input projections - # prior to do_rnnt_pruning (this is an optimization for speed). - logits = self.joiner(am_pruned, lm_pruned, project_input=False) - - with torch.cuda.amp.autocast(enabled=False): - pruned_loss = k2.rnnt_loss_pruned( - logits=logits.float(), - symbols=y_padded, - ranges=ranges, - termination_symbol=blank_id, - boundary=boundary, - reduction="sum", - ) - - return (simple_loss, pruned_loss) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py new file mode 120000 index 000000000..a99e74334 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/model.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/model.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py deleted file mode 100644 index 432bf8220..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py +++ /dev/null @@ -1,331 +0,0 @@ -# Copyright 2022 Xiaomi Corp. (authors: Daniel Povey) -# -# See ../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - - -from typing import List, Optional, Union - -import torch -from torch.optim import Optimizer - - -class Eve(Optimizer): - r""" - Implements Eve algorithm. This is a modified version of AdamW with a special - way of setting the weight-decay / shrinkage-factor, which is designed to make the - rms of the parameters approach a particular target_rms (default: 0.1). This is - for use with networks with 'scaled' versions of modules (see scaling.py), which - will be close to invariant to the absolute scale on the parameter matrix. - - The original Adam algorithm was proposed in `Adam: A Method for Stochastic Optimization`_. - The AdamW variant was proposed in `Decoupled Weight Decay Regularization`_. - Eve is unpublished so far. - - Arguments: - params (iterable): iterable of parameters to optimize or dicts defining - parameter groups - lr (float, optional): learning rate (default: 1e-3) - betas (Tuple[float, float], optional): coefficients used for computing - running averages of gradient and its square (default: (0.9, 0.999)) - eps (float, optional): term added to the denominator to improve - numerical stability (default: 1e-8) - weight_decay (float, optional): weight decay coefficient (default: 3e-4; - this value means that the weight would decay significantly after - about 3k minibatches. Is not multiplied by learning rate, but - is conditional on RMS-value of parameter being > target_rms. - target_rms (float, optional): target root-mean-square value of - parameters, if they fall below this we will stop applying weight decay. - - - .. _Adam\: A Method for Stochastic Optimization: - https://arxiv.org/abs/1412.6980 - .. _Decoupled Weight Decay Regularization: - https://arxiv.org/abs/1711.05101 - .. _On the Convergence of Adam and Beyond: - https://openreview.net/forum?id=ryQu7f-RZ - """ - - def __init__( - self, - params, - lr=1e-3, - betas=(0.9, 0.98), - eps=1e-8, - weight_decay=1e-3, - target_rms=0.1, - ): - - if not 0.0 <= lr: - raise ValueError("Invalid learning rate: {}".format(lr)) - if not 0.0 <= eps: - raise ValueError("Invalid epsilon value: {}".format(eps)) - if not 0.0 <= betas[0] < 1.0: - raise ValueError( - "Invalid beta parameter at index 0: {}".format(betas[0]) - ) - if not 0.0 <= betas[1] < 1.0: - raise ValueError( - "Invalid beta parameter at index 1: {}".format(betas[1]) - ) - if not 0 <= weight_decay <= 0.1: - raise ValueError( - "Invalid weight_decay value: {}".format(weight_decay) - ) - if not 0 < target_rms <= 10.0: - raise ValueError("Invalid target_rms value: {}".format(target_rms)) - defaults = dict( - lr=lr, - betas=betas, - eps=eps, - weight_decay=weight_decay, - target_rms=target_rms, - ) - super(Eve, self).__init__(params, defaults) - - def __setstate__(self, state): - super(Eve, self).__setstate__(state) - - @torch.no_grad() - def step(self, closure=None): - """Performs a single optimization step. - - Arguments: - closure (callable, optional): A closure that reevaluates the model - and returns the loss. - """ - loss = None - if closure is not None: - with torch.enable_grad(): - loss = closure() - - for group in self.param_groups: - for p in group["params"]: - if p.grad is None: - continue - - # Perform optimization step - grad = p.grad - if grad.is_sparse: - raise RuntimeError( - "AdamW does not support sparse gradients" - ) - - state = self.state[p] - - # State initialization - if len(state) == 0: - state["step"] = 0 - # Exponential moving average of gradient values - state["exp_avg"] = torch.zeros_like( - p, memory_format=torch.preserve_format - ) - # Exponential moving average of squared gradient values - state["exp_avg_sq"] = torch.zeros_like( - p, memory_format=torch.preserve_format - ) - - exp_avg, exp_avg_sq = state["exp_avg"], state["exp_avg_sq"] - - beta1, beta2 = group["betas"] - - state["step"] += 1 - bias_correction1 = 1 - beta1 ** state["step"] - bias_correction2 = 1 - beta2 ** state["step"] - - # Decay the first and second moment running average coefficient - exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1) - exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2) - denom = (exp_avg_sq.sqrt() * (bias_correction2 ** -0.5)).add_( - group["eps"] - ) - - step_size = group["lr"] / bias_correction1 - target_rms = group["target_rms"] - weight_decay = group["weight_decay"] - - if p.numel() > 1: - # avoid applying this weight-decay on "scaling factors" - # (which are scalar). - is_above_target_rms = p.norm() > ( - target_rms * (p.numel() ** 0.5) - ) - p.mul_(1 - (weight_decay * is_above_target_rms)) - p.addcdiv_(exp_avg, denom, value=-step_size) - - return loss - - -class LRScheduler(object): - """ - Base-class for learning rate schedulers where the learning-rate depends on both the - batch and the epoch. - """ - - def __init__(self, optimizer: Optimizer, verbose: bool = False): - # Attach optimizer - if not isinstance(optimizer, Optimizer): - raise TypeError( - "{} is not an Optimizer".format(type(optimizer).__name__) - ) - self.optimizer = optimizer - self.verbose = verbose - - for group in optimizer.param_groups: - group.setdefault("initial_lr", group["lr"]) - - self.base_lrs = [ - group["initial_lr"] for group in optimizer.param_groups - ] - - self.epoch = 0 - self.batch = 0 - - def state_dict(self): - """Returns the state of the scheduler as a :class:`dict`. - - It contains an entry for every variable in self.__dict__ which - is not the optimizer. - """ - return { - "base_lrs": self.base_lrs, - "epoch": self.epoch, - "batch": self.batch, - } - - def load_state_dict(self, state_dict): - """Loads the schedulers state. - - Args: - state_dict (dict): scheduler state. Should be an object returned - from a call to :meth:`state_dict`. - """ - self.__dict__.update(state_dict) - - def get_last_lr(self) -> List[float]: - """Return last computed learning rate by current scheduler. Will be a list of float.""" - return self._last_lr - - def get_lr(self): - # Compute list of learning rates from self.epoch and self.batch and - # self.base_lrs; this must be overloaded by the user. - # e.g. return [some_formula(self.batch, self.epoch, base_lr) for base_lr in self.base_lrs ] - raise NotImplementedError - - def step_batch(self, batch: Optional[int] = None) -> None: - # Step the batch index, or just set it. If `batch` is specified, it - # must be the batch index from the start of training, i.e. summed over - # all epochs. - # You can call this in any order; if you don't provide 'batch', it should - # of course be called once per batch. - if batch is not None: - self.batch = batch - else: - self.batch = self.batch + 1 - self._set_lrs() - - def step_epoch(self, epoch: Optional[int] = None): - # Step the epoch index, or just set it. If you provide the 'epoch' arg, - # you should call this at the start of the epoch; if you don't provide the 'epoch' - # arg, you should call it at the end of the epoch. - if epoch is not None: - self.epoch = epoch - else: - self.epoch = self.epoch + 1 - self._set_lrs() - - def _set_lrs(self): - values = self.get_lr() - assert len(values) == len(self.optimizer.param_groups) - - for i, data in enumerate(zip(self.optimizer.param_groups, values)): - param_group, lr = data - param_group["lr"] = lr - self.print_lr(self.verbose, i, lr) - self._last_lr = [group["lr"] for group in self.optimizer.param_groups] - - def print_lr(self, is_verbose, group, lr): - """Display the current learning rate.""" - if is_verbose: - print( - f"Epoch={self.epoch}, batch={self.batch}: adjusting learning rate" - f" of group {group} to {lr:.4e}." - ) - - -class Eden(LRScheduler): - """ - Eden scheduler. - lr = initial_lr * (((batch**2 + lr_batches**2) / lr_batches**2) ** -0.25 * - (((epoch**2 + lr_epochs**2) / lr_epochs**2) ** -0.25)) - - E.g. suggest initial-lr = 0.003 (passed to optimizer). - - Args: - optimizer: the optimizer to change the learning rates on - lr_batches: the number of batches after which we start significantly - decreasing the learning rate, suggest 5000. - lr_epochs: the number of epochs after which we start significantly - decreasing the learning rate, suggest 6 if you plan to do e.g. - 20 to 40 epochs, but may need smaller number if dataset is huge - and you will do few epochs. - """ - - def __init__( - self, - optimizer: Optimizer, - lr_batches: Union[int, float], - lr_epochs: Union[int, float], - verbose: bool = False, - ): - super(Eden, self).__init__(optimizer, verbose) - self.lr_batches = lr_batches - self.lr_epochs = lr_epochs - - def get_lr(self): - factor = ( - (self.batch ** 2 + self.lr_batches ** 2) / self.lr_batches ** 2 - ) ** -0.25 * ( - ((self.epoch ** 2 + self.lr_epochs ** 2) / self.lr_epochs ** 2) - ** -0.25 - ) - return [x * factor for x in self.base_lrs] - - -def _test_eden(): - m = torch.nn.Linear(100, 100) - optim = Eve(m.parameters(), lr=0.003) - - scheduler = Eden(optim, lr_batches=30, lr_epochs=2, verbose=True) - - for epoch in range(10): - scheduler.step_epoch(epoch) # sets epoch to `epoch` - - for step in range(20): - x = torch.randn(200, 100).detach() - x.requires_grad = True - y = m(x) - dy = torch.randn(200, 100).detach() - f = (y * dy).sum() - f.backward() - - optim.step() - scheduler.step_batch() - optim.zero_grad() - print("last lr = ", scheduler.get_last_lr()) - print("state dict = ", scheduler.state_dict()) - - -if __name__ == "__main__": - _test_eden() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py new file mode 120000 index 000000000..0a2f285aa --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/optim.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/optim.py \ No newline at end of file diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py deleted file mode 100644 index f89d2963e..000000000 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py +++ /dev/null @@ -1,702 +0,0 @@ -# Copyright 2022 Xiaomi Corp. (authors: Daniel Povey) -# -# See ../../../../LICENSE for clarification regarding multiple authors -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - - -import collections -from itertools import repeat -from typing import Optional, Tuple - -import torch -import torch.nn as nn -from torch import Tensor - - -def _ntuple(n): - def parse(x): - if isinstance(x, collections.Iterable): - return x - return tuple(repeat(x, n)) - - return parse - - -_single = _ntuple(1) -_pair = _ntuple(2) - - -class ActivationBalancerFunction(torch.autograd.Function): - @staticmethod - def forward( - ctx, - x: Tensor, - channel_dim: int, - min_positive: float, # e.g. 0.05 - max_positive: float, # e.g. 0.95 - max_factor: float, # e.g. 0.01 - min_abs: float, # e.g. 0.2 - max_abs: float, # e.g. 100.0 - ) -> Tensor: - if x.requires_grad: - if channel_dim < 0: - channel_dim += x.ndim - sum_dims = [d for d in range(x.ndim) if d != channel_dim] - xgt0 = x > 0 - proportion_positive = torch.mean( - xgt0.to(x.dtype), dim=sum_dims, keepdim=True - ) - factor1 = ( - (min_positive - proportion_positive).relu() - * (max_factor / min_positive) - if min_positive != 0.0 - else 0.0 - ) - factor2 = ( - (proportion_positive - max_positive).relu() - * (max_factor / (max_positive - 1.0)) - if max_positive != 1.0 - else 0.0 - ) - factor = factor1 + factor2 - if isinstance(factor, float): - factor = torch.zeros_like(proportion_positive) - - mean_abs = torch.mean(x.abs(), dim=sum_dims, keepdim=True) - below_threshold = mean_abs < min_abs - above_threshold = mean_abs > max_abs - - ctx.save_for_backward( - factor, xgt0, below_threshold, above_threshold - ) - ctx.max_factor = max_factor - ctx.sum_dims = sum_dims - return x - - @staticmethod - def backward( - ctx, x_grad: Tensor - ) -> Tuple[Tensor, None, None, None, None, None, None]: - factor, xgt0, below_threshold, above_threshold = ctx.saved_tensors - dtype = x_grad.dtype - scale_factor = ( - (below_threshold.to(dtype) - above_threshold.to(dtype)) - * (xgt0.to(dtype) - 0.5) - * (ctx.max_factor * 2.0) - ) - - neg_delta_grad = x_grad.abs() * (factor + scale_factor) - return x_grad - neg_delta_grad, None, None, None, None, None, None - - -class BasicNorm(torch.nn.Module): - """ - This is intended to be a simpler, and hopefully cheaper, replacement for - LayerNorm. The observation this is based on, is that Transformer-type - networks, especially with pre-norm, sometimes seem to set one of the - feature dimensions to a large constant value (e.g. 50), which "defeats" - the LayerNorm because the output magnitude is then not strongly dependent - on the other (useful) features. Presumably the weight and bias of the - LayerNorm are required to allow it to do this. - - So the idea is to introduce this large constant value as an explicit - parameter, that takes the role of the "eps" in LayerNorm, so the network - doesn't have to do this trick. We make the "eps" learnable. - - Args: - num_channels: the number of channels, e.g. 512. - channel_dim: the axis/dimension corresponding to the channel, - interprted as an offset from the input's ndim if negative. - shis is NOT the num_channels; it should typically be one of - {-2, -1, 0, 1, 2, 3}. - eps: the initial "epsilon" that we add as ballast in: - scale = ((input_vec**2).mean() + epsilon)**-0.5 - Note: our epsilon is actually large, but we keep the name - to indicate the connection with conventional LayerNorm. - learn_eps: if true, we learn epsilon; if false, we keep it - at the initial value. - """ - - def __init__( - self, - num_channels: int, - channel_dim: int = -1, # CAUTION: see documentation. - eps: float = 0.25, - learn_eps: bool = True, - ) -> None: - super(BasicNorm, self).__init__() - self.num_channels = num_channels - self.channel_dim = channel_dim - if learn_eps: - self.eps = nn.Parameter(torch.tensor(eps).log().detach()) - else: - self.register_buffer("eps", torch.tensor(eps).log().detach()) - - def forward(self, x: Tensor) -> Tensor: - assert x.shape[self.channel_dim] == self.num_channels - scales = ( - torch.mean(x ** 2, dim=self.channel_dim, keepdim=True) - + self.eps.exp() - ) ** -0.5 - return x * scales - - -class ScaledLinear(nn.Linear): - """ - A modified version of nn.Linear where the parameters are scaled before - use, via: - weight = self.weight * self.weight_scale.exp() - bias = self.bias * self.bias_scale.exp() - - Args: - Accepts the standard args and kwargs that nn.Linear accepts - e.g. in_features, out_features, bias=False. - - initial_scale: you can override this if you want to increase - or decrease the initial magnitude of the module's output - (affects the initialization of weight_scale and bias_scale). - Another option, if you want to do something like this, is - to re-initialize the parameters. - initial_speed: this affects how fast the parameter will - learn near the start of training; you can set it to a - value less than one if you suspect that a module - is contributing to instability near the start of training. - Nnote: regardless of the use of this option, it's best to - use schedulers like Noam that have a warm-up period. - Alternatively you can set it to more than 1 if you want it to - initially train faster. Must be greater than 0. - """ - - def __init__( - self, - *args, - initial_scale: float = 1.0, - initial_speed: float = 1.0, - **kwargs - ): - super(ScaledLinear, self).__init__(*args, **kwargs) - initial_scale = torch.tensor(initial_scale).log() - self.weight_scale = nn.Parameter(initial_scale.clone().detach()) - if self.bias is not None: - self.bias_scale = nn.Parameter(initial_scale.clone().detach()) - else: - self.register_parameter("bias_scale", None) - - self._reset_parameters( - initial_speed - ) # Overrides the reset_parameters in nn.Linear - - def _reset_parameters(self, initial_speed: float): - std = 0.1 / initial_speed - a = (3 ** 0.5) * std - nn.init.uniform_(self.weight, -a, a) - if self.bias is not None: - nn.init.constant_(self.bias, 0.0) - fan_in = self.weight.shape[1] * self.weight[0][0].numel() - scale = fan_in ** -0.5 # 1/sqrt(fan_in) - with torch.no_grad(): - self.weight_scale += torch.tensor(scale / std).log() - - def get_weight(self): - return self.weight * self.weight_scale.exp() - - def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() - - def forward(self, input: Tensor) -> Tensor: - return torch.nn.functional.linear( - input, self.get_weight(), self.get_bias() - ) - - -class ScaledConv1d(nn.Conv1d): - # See docs for ScaledLinear - def __init__( - self, - *args, - initial_scale: float = 1.0, - initial_speed: float = 1.0, - **kwargs - ): - super(ScaledConv1d, self).__init__(*args, **kwargs) - initial_scale = torch.tensor(initial_scale).log() - self.weight_scale = nn.Parameter(initial_scale.clone().detach()) - if self.bias is not None: - self.bias_scale = nn.Parameter(initial_scale.clone().detach()) - else: - self.register_parameter("bias_scale", None) - self._reset_parameters( - initial_speed - ) # Overrides the reset_parameters in base class - - def _reset_parameters(self, initial_speed: float): - std = 0.1 / initial_speed - a = (3 ** 0.5) * std - nn.init.uniform_(self.weight, -a, a) - if self.bias is not None: - nn.init.constant_(self.bias, 0.0) - fan_in = self.weight.shape[1] * self.weight[0][0].numel() - scale = fan_in ** -0.5 # 1/sqrt(fan_in) - with torch.no_grad(): - self.weight_scale += torch.tensor(scale / std).log() - - def get_weight(self): - return self.weight * self.weight_scale.exp() - - def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() - - def forward(self, input: Tensor) -> Tensor: - F = torch.nn.functional - if self.padding_mode != "zeros": - return F.conv1d( - F.pad( - input, - self._reversed_padding_repeated_twice, - mode=self.padding_mode, - ), - self.get_weight(), - self.get_bias(), - self.stride, - _single(0), - self.dilation, - self.groups, - ) - return F.conv1d( - input, - self.get_weight(), - self.get_bias(), - self.stride, - self.padding, - self.dilation, - self.groups, - ) - - -class ScaledConv2d(nn.Conv2d): - # See docs for ScaledLinear - def __init__( - self, - *args, - initial_scale: float = 1.0, - initial_speed: float = 1.0, - **kwargs - ): - super(ScaledConv2d, self).__init__(*args, **kwargs) - initial_scale = torch.tensor(initial_scale).log() - self.weight_scale = nn.Parameter(initial_scale.clone().detach()) - if self.bias is not None: - self.bias_scale = nn.Parameter(initial_scale.clone().detach()) - else: - self.register_parameter("bias_scale", None) - self._reset_parameters( - initial_speed - ) # Overrides the reset_parameters in base class - - def _reset_parameters(self, initial_speed: float): - std = 0.1 / initial_speed - a = (3 ** 0.5) * std - nn.init.uniform_(self.weight, -a, a) - if self.bias is not None: - nn.init.constant_(self.bias, 0.0) - fan_in = self.weight.shape[1] * self.weight[0][0].numel() - scale = fan_in ** -0.5 # 1/sqrt(fan_in) - with torch.no_grad(): - self.weight_scale += torch.tensor(scale / std).log() - - def get_weight(self): - return self.weight * self.weight_scale.exp() - - def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() - - def _conv_forward(self, input, weight): - F = torch.nn.functional - if self.padding_mode != "zeros": - return F.conv2d( - F.pad( - input, - self._reversed_padding_repeated_twice, - mode=self.padding_mode, - ), - weight, - self.get_bias(), - self.stride, - _pair(0), - self.dilation, - self.groups, - ) - return F.conv2d( - input, - weight, - self.get_bias(), - self.stride, - self.padding, - self.dilation, - self.groups, - ) - - def forward(self, input: Tensor) -> Tensor: - return self._conv_forward(input, self.get_weight()) - - -class ActivationBalancer(torch.nn.Module): - """ - Modifies the backpropped derivatives of a function to try to encourage, for - each channel, that it is positive at least a proportion `threshold` of the - time. It does this by multiplying negative derivative values by up to - (1+max_factor), and positive derivative values by up to (1-max_factor), - interpolated from 1 at the threshold to those extremal values when none - of the inputs are positive. - - - Args: - channel_dim: the dimension/axis corresponding to the channel, e.g. - -1, 0, 1, 2; will be interpreted as an offset from x.ndim if negative. - min_positive: the minimum, per channel, of the proportion of the time - that (x > 0), below which we start to modify the derivatives. - max_positive: the maximum, per channel, of the proportion of the time - that (x > 0), above which we start to modify the derivatives. - max_factor: the maximum factor by which we modify the derivatives for - either the sign constraint or the magnitude constraint; - e.g. with max_factor=0.02, the the derivatives would be multiplied by - values in the range [0.98..1.02]. - min_abs: the minimum average-absolute-value per channel, which - we allow, before we start to modify the derivatives to prevent - this. - max_abs: the maximum average-absolute-value per channel, which - we allow, before we start to modify the derivatives to prevent - this. - """ - - def __init__( - self, - channel_dim: int, - min_positive: float = 0.05, - max_positive: float = 0.95, - max_factor: float = 0.01, - min_abs: float = 0.2, - max_abs: float = 100.0, - ): - super(ActivationBalancer, self).__init__() - self.channel_dim = channel_dim - self.min_positive = min_positive - self.max_positive = max_positive - self.max_factor = max_factor - self.min_abs = min_abs - self.max_abs = max_abs - - def forward(self, x: Tensor) -> Tensor: - return ActivationBalancerFunction.apply( - x, - self.channel_dim, - self.min_positive, - self.max_positive, - self.max_factor, - self.min_abs, - self.max_abs, - ) - - -class DoubleSwishFunction(torch.autograd.Function): - """ - double_swish(x) = x * torch.sigmoid(x-1) - This is a definition, originally motivated by its close numerical - similarity to swish(swish(x)), where swish(x) = x * sigmoid(x). - - Memory-efficient derivative computation: - double_swish(x) = x * s, where s(x) = torch.sigmoid(x-1) - double_swish'(x) = d/dx double_swish(x) = x * s'(x) + x' * s(x) = x * s'(x) + s(x). - Now, s'(x) = s(x) * (1-s(x)). - double_swish'(x) = x * s'(x) + s(x). - = x * s(x) * (1-s(x)) + s(x). - = double_swish(x) * (1-s(x)) + s(x) - ... so we just need to remember s(x) but not x itself. - """ - - @staticmethod - def forward(ctx, x: Tensor) -> Tensor: - x = x.detach() - s = torch.sigmoid(x - 1.0) - y = x * s - ctx.save_for_backward(s, y) - return y - - @staticmethod - def backward(ctx, y_grad: Tensor) -> Tensor: - s, y = ctx.saved_tensors - return (y * (1 - s) + s) * y_grad - - -class DoubleSwish(torch.nn.Module): - def forward(self, x: Tensor) -> Tensor: - """Return double-swish activation function which is an approximation to Swish(Swish(x)), - that we approximate closely with x * sigmoid(x-1). - """ - return DoubleSwishFunction.apply(x) - - -class ScaledEmbedding(nn.Module): - r"""This is a modified version of nn.Embedding that introduces a learnable scale - on the parameters. Note: due to how we initialize it, it's best used with - schedulers like Noam that have a warmup period. - - It is a simple lookup table that stores embeddings of a fixed dictionary and size. - - This module is often used to store word embeddings and retrieve them using indices. - The input to the module is a list of indices, and the output is the corresponding - word embeddings. - - Args: - num_embeddings (int): size of the dictionary of embeddings - embedding_dim (int): the size of each embedding vector - padding_idx (int, optional): If given, pads the output with the embedding vector at :attr:`padding_idx` - (initialized to zeros) whenever it encounters the index. - max_norm (float, optional): If given, each embedding vector with norm larger than :attr:`max_norm` - is renormalized to have norm :attr:`max_norm`. - norm_type (float, optional): The p of the p-norm to compute for the :attr:`max_norm` option. Default ``2``. - scale_grad_by_freq (boolean, optional): If given, this will scale gradients by the inverse of frequency of - the words in the mini-batch. Default ``False``. - sparse (bool, optional): If ``True``, gradient w.r.t. :attr:`weight` matrix will be a sparse tensor. - See Notes for more details regarding sparse gradients. - - initial_speed (float, optional): This affects how fast the parameter will - learn near the start of training; you can set it to a value less than - one if you suspect that a module is contributing to instability near - the start of training. Nnote: regardless of the use of this option, - it's best to use schedulers like Noam that have a warm-up period. - Alternatively you can set it to more than 1 if you want it to - initially train faster. Must be greater than 0. - - - Attributes: - weight (Tensor): the learnable weights of the module of shape (num_embeddings, embedding_dim) - initialized from :math:`\mathcal{N}(0, 1)` - - Shape: - - Input: :math:`(*)`, LongTensor of arbitrary shape containing the indices to extract - - Output: :math:`(*, H)`, where `*` is the input shape and :math:`H=\text{embedding\_dim}` - - .. note:: - Keep in mind that only a limited number of optimizers support - sparse gradients: currently it's :class:`optim.SGD` (`CUDA` and `CPU`), - :class:`optim.SparseAdam` (`CUDA` and `CPU`) and :class:`optim.Adagrad` (`CPU`) - - .. note:: - With :attr:`padding_idx` set, the embedding vector at - :attr:`padding_idx` is initialized to all zeros. However, note that this - vector can be modified afterwards, e.g., using a customized - initialization method, and thus changing the vector used to pad the - output. The gradient for this vector from :class:`~torch.nn.Embedding` - is always zero. - - Examples:: - - >>> # an Embedding module containing 10 tensors of size 3 - >>> embedding = nn.Embedding(10, 3) - >>> # a batch of 2 samples of 4 indices each - >>> input = torch.LongTensor([[1,2,4,5],[4,3,2,9]]) - >>> embedding(input) - tensor([[[-0.0251, -1.6902, 0.7172], - [-0.6431, 0.0748, 0.6969], - [ 1.4970, 1.3448, -0.9685], - [-0.3677, -2.7265, -0.1685]], - - [[ 1.4970, 1.3448, -0.9685], - [ 0.4362, -0.4004, 0.9400], - [-0.6431, 0.0748, 0.6969], - [ 0.9124, -2.3616, 1.1151]]]) - - - >>> # example with padding_idx - >>> embedding = nn.Embedding(10, 3, padding_idx=0) - >>> input = torch.LongTensor([[0,2,0,5]]) - >>> embedding(input) - tensor([[[ 0.0000, 0.0000, 0.0000], - [ 0.1535, -2.0309, 0.9315], - [ 0.0000, 0.0000, 0.0000], - [-0.1655, 0.9897, 0.0635]]]) - - """ - __constants__ = [ - "num_embeddings", - "embedding_dim", - "padding_idx", - "scale_grad_by_freq", - "sparse", - ] - - num_embeddings: int - embedding_dim: int - padding_idx: int - scale_grad_by_freq: bool - weight: Tensor - sparse: bool - - def __init__( - self, - num_embeddings: int, - embedding_dim: int, - padding_idx: Optional[int] = None, - scale_grad_by_freq: bool = False, - sparse: bool = False, - initial_speed: float = 1.0, - ) -> None: - super(ScaledEmbedding, self).__init__() - self.num_embeddings = num_embeddings - self.embedding_dim = embedding_dim - if padding_idx is not None: - if padding_idx > 0: - assert ( - padding_idx < self.num_embeddings - ), "Padding_idx must be within num_embeddings" - elif padding_idx < 0: - assert ( - padding_idx >= -self.num_embeddings - ), "Padding_idx must be within num_embeddings" - padding_idx = self.num_embeddings + padding_idx - self.padding_idx = padding_idx - self.scale_grad_by_freq = scale_grad_by_freq - - self.scale = nn.Parameter(torch.zeros(())) # see reset_parameters() - self.sparse = sparse - - self.weight = nn.Parameter(torch.Tensor(num_embeddings, embedding_dim)) - self.reset_parameters(initial_speed) - - def reset_parameters(self, initial_speed: float = 1.0) -> None: - std = 0.1 / initial_speed - nn.init.normal_(self.weight, std=std) - nn.init.constant_(self.scale, torch.tensor(1.0 / std).log()) - - if self.padding_idx is not None: - with torch.no_grad(): - self.weight[self.padding_idx].fill_(0) - - def forward(self, input: Tensor) -> Tensor: - F = torch.nn.functional - scale = self.scale.exp() - if input.numel() < self.num_embeddings: - return ( - F.embedding( - input, - self.weight, - self.padding_idx, - None, - 2.0, # None, 2.0 relate to normalization - self.scale_grad_by_freq, - self.sparse, - ) - * scale - ) - else: - return F.embedding( - input, - self.weight * scale, - self.padding_idx, - None, - 2.0, # None, 2.0 relates to normalization - self.scale_grad_by_freq, - self.sparse, - ) - - def extra_repr(self) -> str: - s = "{num_embeddings}, {embedding_dim}, scale={scale}" - if self.padding_idx is not None: - s += ", padding_idx={padding_idx}" - if self.scale_grad_by_freq is not False: - s += ", scale_grad_by_freq={scale_grad_by_freq}" - if self.sparse is not False: - s += ", sparse=True" - return s.format(**self.__dict__) - - -def _test_activation_balancer_sign(): - probs = torch.arange(0, 1, 0.01) - N = 1000 - x = 1.0 * (torch.rand(probs.numel(), N) < probs.unsqueeze(-1)) - x = x.detach() - x.requires_grad = True - m = ActivationBalancer( - channel_dim=0, - min_positive=0.05, - max_positive=0.95, - max_factor=0.2, - min_abs=0.0, - ) - - y_grad = torch.sign(torch.randn(probs.numel(), N)) - - y = m(x) - y.backward(gradient=y_grad) - print("_test_activation_balancer_sign: x = ", x) - print("_test_activation_balancer_sign: y grad = ", y_grad) - print("_test_activation_balancer_sign: x grad = ", x.grad) - - -def _test_activation_balancer_magnitude(): - magnitudes = torch.arange(0, 1, 0.01) - N = 1000 - x = torch.sign(torch.randn(magnitudes.numel(), N)) * magnitudes.unsqueeze( - -1 - ) - x = x.detach() - x.requires_grad = True - m = ActivationBalancer( - channel_dim=0, - min_positive=0.0, - max_positive=1.0, - max_factor=0.2, - min_abs=0.2, - max_abs=0.8, - ) - - y_grad = torch.sign(torch.randn(magnitudes.numel(), N)) - - y = m(x) - y.backward(gradient=y_grad) - print("_test_activation_balancer_magnitude: x = ", x) - print("_test_activation_balancer_magnitude: y grad = ", y_grad) - print("_test_activation_balancer_magnitude: x grad = ", x.grad) - - -def _test_basic_norm(): - num_channels = 128 - m = BasicNorm(num_channels=num_channels, channel_dim=1) - - x = torch.randn(500, num_channels) - - y = m(x) - - assert y.shape == x.shape - x_rms = (x ** 2).mean().sqrt() - y_rms = (y ** 2).mean().sqrt() - print("x rms = ", x_rms) - print("y rms = ", y_rms) - assert y_rms < x_rms - assert y_rms > 0.5 * x_rms - - -def _test_double_swish_deriv(): - x = torch.randn(10, 12, dtype=torch.double) * 0.5 - x.requires_grad = True - m = DoubleSwish() - torch.autograd.gradcheck(m, x) - - -if __name__ == "__main__": - _test_activation_balancer_sign() - _test_activation_balancer_magnitude() - _test_basic_norm() - _test_double_swish_deriv() diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py new file mode 120000 index 000000000..c10cdfe12 --- /dev/null +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/scaling.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/scaling.py \ No newline at end of file From 0f180b3ce20c3e704ef46972aa61816fd05a42ac Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Fri, 13 May 2022 14:00:35 +0800 Subject: [PATCH 04/16] Validate that there are no OOV tokens in BPE-based lexicons. (#359) * Validate that there are no OOV tokens in BPE-based lexicons. * Typo fixes. --- egs/librispeech/ASR/local/prepare_lang_bpe.py | 9 ++- .../ASR/local/validate_bpe_lexicon.py | 77 +++++++++++++++++++ egs/librispeech/ASR/prepare.sh | 15 +++- 3 files changed, 96 insertions(+), 5 deletions(-) create mode 100755 egs/librispeech/ASR/local/validate_bpe_lexicon.py diff --git a/egs/librispeech/ASR/local/prepare_lang_bpe.py b/egs/librispeech/ASR/local/prepare_lang_bpe.py index cf32f308d..dec8a7442 100755 --- a/egs/librispeech/ASR/local/prepare_lang_bpe.py +++ b/egs/librispeech/ASR/local/prepare_lang_bpe.py @@ -145,7 +145,14 @@ def generate_lexicon( sp = spm.SentencePieceProcessor() sp.load(str(model_file)) - words_pieces: List[List[str]] = sp.encode(words, out_type=str) + # Convert word to word piece IDs instead of word piece strings + # to avoid OOV tokens. + words_pieces_ids: List[List[int]] = sp.encode(words, out_type=int) + + # Now convert word piece IDs back to word piece strings. + words_pieces: List[List[str]] = [ + sp.id_to_piece(ids) for ids in words_pieces_ids + ] lexicon = [] for word, pieces in zip(words, words_pieces): diff --git a/egs/librispeech/ASR/local/validate_bpe_lexicon.py b/egs/librispeech/ASR/local/validate_bpe_lexicon.py new file mode 100755 index 000000000..c542f2fab --- /dev/null +++ b/egs/librispeech/ASR/local/validate_bpe_lexicon.py @@ -0,0 +1,77 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +""" +This script checks that there are no OOV tokens in the BPE-based lexicon. + +Usage example: + + python3 ./local/validate_bpe_lexicon.py \ + --lexicon /path/to/lexicon.txt \ + --bpe-model /path/to/bpe.model +""" + +import argparse +from pathlib import Path +from typing import List, Tuple + +import sentencepiece as spm + +from icefall.lexicon import read_lexicon + +# Map word to word pieces +Lexicon = List[Tuple[str, List[str]]] + + +def get_args(): + parser = argparse.ArgumentParser() + + parser.add_argument( + "--lexicon", + required=True, + type=Path, + help="Path to lexicon.txt", + ) + + parser.add_argument( + "--bpe-model", + required=True, + type=Path, + help="Path to bpe.model", + ) + + return parser.parse_args() + + +def main(): + args = get_args() + assert args.lexicon.is_file(), args.lexicon + assert args.bpe_model.is_file(), args.bpe_model + + lexicon = read_lexicon(args.lexicon) + + sp = spm.SentencePieceProcessor() + sp.load(str(args.bpe_model)) + + word_pieces = set(sp.id_to_piece(list(range(sp.vocab_size())))) + for word, pieces in lexicon: + for p in pieces: + if p not in word_pieces: + raise ValueError(f"The word {word} contains an OOV token {p}") + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/prepare.sh b/egs/librispeech/ASR/prepare.sh index 33d298a7a..8cfb046c8 100755 --- a/egs/librispeech/ASR/prepare.sh +++ b/egs/librispeech/ASR/prepare.sh @@ -184,13 +184,20 @@ if [ $stage -le 6 ] && [ $stop_stage -ge 6 ]; then done > $lang_dir/transcript_words.txt fi - ./local/train_bpe_model.py \ - --lang-dir $lang_dir \ - --vocab-size $vocab_size \ - --transcript $lang_dir/transcript_words.txt + if [ ! -f $lang_dir/bpe.model ]; then + ./local/train_bpe_model.py \ + --lang-dir $lang_dir \ + --vocab-size $vocab_size \ + --transcript $lang_dir/transcript_words.txt + fi if [ ! -f $lang_dir/L_disambig.pt ]; then ./local/prepare_lang_bpe.py --lang-dir $lang_dir + + log "Validating $lang_dir/lexicon.txt" + ./local/validate_bpe_lexicon.py \ + --lexicon $lang_dir/lexicon.txt \ + --bpe-model $lang_dir/bpe.model fi done fi From 2d7096dfc61675eadeec522446e8636ef17c4188 Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Sat, 14 May 2022 08:53:22 +0800 Subject: [PATCH 05/16] Decode gigaspeech in GitHub actions (#362) * Add CI for gigaspeech. --- .../download-gigaspeech-dev-test-dataset.sh | 15 +++ ...pruned-transducer-stateless2-2022-05-12.sh | 49 +++++++ .../workflows/run-gigaspeech-2022-05-13.yml | 120 ++++++++++++++++++ 3 files changed, 184 insertions(+) create mode 100755 .github/scripts/download-gigaspeech-dev-test-dataset.sh create mode 100755 .github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh create mode 100644 .github/workflows/run-gigaspeech-2022-05-13.yml diff --git a/.github/scripts/download-gigaspeech-dev-test-dataset.sh b/.github/scripts/download-gigaspeech-dev-test-dataset.sh new file mode 100755 index 000000000..b9464de9f --- /dev/null +++ b/.github/scripts/download-gigaspeech-dev-test-dataset.sh @@ -0,0 +1,15 @@ +#!/usr/bin/env bash + +# This script downloads the pre-computed fbank features for +# dev and test datasets of GigaSpeech. +# +# You will find directories `~/tmp/giga-dev-dataset-fbank` after running +# this script. + +mkdir -p ~/tmp +cd ~/tmp + +git lfs install +git clone https://huggingface.co/csukuangfj/giga-dev-dataset-fbank + +ls -lh giga-dev-dataset-fbank/data/fbank diff --git a/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh b/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh new file mode 100755 index 000000000..25bac2fdb --- /dev/null +++ b/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh @@ -0,0 +1,49 @@ +#!/usr/bin/env bash + +log() { + # This function is from espnet + local fname=${BASH_SOURCE[1]##*/} + echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*" +} + +cd egs/gigaspeech/ASR + +repo_url=https://huggingface.co/wgb14/icefall-asr-gigaspeech-pruned-transducer-stateless2 + +log "Downloading pre-trained model from $repo_url" +git lfs install +git clone $repo_url +repo=$(basename $repo_url) + +echo "GITHUB_EVENT_NAME: ${GITHUB_EVENT_NAME}" +echo "GITHUB_EVENT_LABEL_NAME: ${GITHUB_EVENT_LABEL_NAME}" +if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" == x"run-decode" ]]; then + mkdir -p pruned_transducer_stateless2/exp + ln -s $PWD/$repo/exp/pretrained-epoch-29-avg-11.pt pruned_transducer_stateless2/exp/epoch-999.pt + ln -s $PWD/$repo/data/lang_bpe_500 data/ + + ls -lh data + ls -lh data/lang_bpe_500 + ls -lh data/fbank + ls -lh pruned_transducer_stateless2/exp + + log "Decoding dev and test" + + # use a small value for decoding with CPU + max_duration=100 + + # Test only greedy_search to reduce CI running time + # for method in greedy_search fast_beam_search modified_beam_search; do + for method in greedy_search; do + log "Decoding with $method" + + ./pruned_transducer_stateless2/decode.py \ + --decoding-method $method \ + --epoch 999 \ + --avg 1 \ + --max-duration $max_duration \ + --exp-dir pruned_transducer_stateless2/exp + done + + rm pruned_transducer_stateless2/exp/*.pt +fi diff --git a/.github/workflows/run-gigaspeech-2022-05-13.yml b/.github/workflows/run-gigaspeech-2022-05-13.yml new file mode 100644 index 000000000..d250b72b0 --- /dev/null +++ b/.github/workflows/run-gigaspeech-2022-05-13.yml @@ -0,0 +1,120 @@ +# Copyright 2021 Fangjun Kuang (csukuangfj@gmail.com) + +# See ../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +name: run-gigaspeech-2022-05-13 +# stateless transducer + k2 pruned rnnt-loss + reworked conformer + +on: + push: + branches: + - master + pull_request: + types: [labeled] + + schedule: + # minute (0-59) + # hour (0-23) + # day of the month (1-31) + # month (1-12) + # day of the week (0-6) + # nightly build at 15:50 UTC time every day + - cron: "50 15 * * *" + +jobs: + run_gigaspeech_2022_05_13: + if: github.event.label.name == 'ready' || github.event.label.name == 'run-decode' || github.event_name == 'push' || github.event_name == 'schedule' + runs-on: ${{ matrix.os }} + strategy: + matrix: + os: [ubuntu-18.04] + python-version: [3.7, 3.8, 3.9] + + fail-fast: false + + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 + + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v2 + with: + python-version: ${{ matrix.python-version }} + cache: 'pip' + cache-dependency-path: '**/requirements-ci.txt' + + - name: Install Python dependencies + run: | + grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install + + - name: Cache kaldifeat + id: my-cache + uses: actions/cache@v2 + with: + path: | + ~/tmp/kaldifeat + key: cache-tmp-${{ matrix.python-version }} + + - name: Install kaldifeat + if: steps.my-cache.outputs.cache-hit != 'true' + shell: bash + run: | + .github/scripts/install-kaldifeat.sh + + - name: Download GigaSpeech dev/test dataset + shell: bash + run: | + sudo apt-get install -y -q git-lfs + + .github/scripts/download-gigaspeech-dev-test-dataset.sh + + - name: Inference with pre-trained model + shell: bash + env: + GITHUB_EVENT_NAME: ${{ github.event_name }} + GITHUB_EVENT_LABEL_NAME: ${{ github.event.label.name }} + run: | + ln -s ~/tmp/giga-dev-dataset-fbank/data egs/gigaspeech/ASR/ + + ls -lh egs/gigaspeech/ASR/data/fbank + + export PYTHONPATH=$PWD:$PYTHONPATH + export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH + export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH + + .github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh + + - name: Display decoding results for gigaspeech pruned_transducer_stateless2 + if: github.event_name == 'schedule' || github.event.label.name == 'run-decode' + shell: bash + run: | + cd egs/gigaspeech/ASR/ + tree ./pruned_transducer_stateless2/exp + + sudo apt-get -qq install tree + + cd pruned_transducer_stateless2 + echo "results for pruned_transducer_stateless2" + echo "===greedy search===" + find exp/greedy_search -name "log-*" -exec grep -n --color "best for dev" {} + | sort -n -k2 + find exp/greedy_search -name "log-*" -exec grep -n --color "best for test" {} + | sort -n -k2 + + - name: Upload decoding results for gigaspeech pruned_transducer_stateless2 + uses: actions/upload-artifact@v2 + if: github.event_name == 'schedule' || github.event.label.name == 'run-decode' + with: + name: torch-${{ matrix.torch }}-python-${{ matrix.python-version }}-ubuntu-18.04-cpu-gigaspeech-pruned_transducer_stateless2-2022-05-12 + path: egs/gigaspeech/ASR/pruned_transducer_stateless2/exp/ From f23dd437196178e11de9f1b283053a3a9565f63d Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Sat, 14 May 2022 21:45:39 +0800 Subject: [PATCH 06/16] Update results for libri+giga multi dataset setup. (#363) * Update results for libri+giga multi dataset setup. --- ...pruned-transducer-stateless3-2022-05-13.sh | 80 ++++++++++ .../workflows/run-librispeech-2022-04-29.yml | 8 +- ...runed-transducer-stateless3-2022-05-13.yml | 151 ++++++++++++++++++ README.md | 2 +- egs/librispeech/ASR/RESULTS.md | 63 +++++++- .../pruned_transducer_stateless/pretrained.py | 49 +++--- .../pretrained.py | 82 +++++++--- .../pretrained.py | 82 +++++++--- .../ASR/pruned_transducer_stateless3/train.py | 1 + 9 files changed, 451 insertions(+), 67 deletions(-) create mode 100755 .github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh create mode 100644 .github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml diff --git a/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh b/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh new file mode 100755 index 000000000..3617bc369 --- /dev/null +++ b/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh @@ -0,0 +1,80 @@ +#!/usr/bin/env bash + +log() { + # This function is from espnet + local fname=${BASH_SOURCE[1]##*/} + echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*" +} + +cd egs/librispeech/ASR + +repo_url=https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13 + +log "Downloading pre-trained model from $repo_url" +git lfs install +git clone $repo_url +repo=$(basename $repo_url) + +log "Display test files" +tree $repo/ +soxi $repo/test_wavs/*.wav +ls -lh $repo/test_wavs/*.wav + +pushd $repo/exp +ln -s pretrained-iter-1224000-avg-14.pt pretrained.pt +popd + +for sym in 1 2 3; do + log "Greedy search with --max-sym-per-frame $sym" + + ./pruned_transducer_stateless3/pretrained.py \ + --method greedy_search \ + --max-sym-per-frame $sym \ + --checkpoint $repo/exp/pretrained.pt \ + --bpe-model $repo/data/lang_bpe_500/bpe.model \ + $repo/test_wavs/1089-134686-0001.wav \ + $repo/test_wavs/1221-135766-0001.wav \ + $repo/test_wavs/1221-135766-0002.wav +done + +for method in modified_beam_search beam_search fast_beam_search; do + log "$method" + + ./pruned_transducer_stateless3/pretrained.py \ + --method $method \ + --beam-size 4 \ + --checkpoint $repo/exp/pretrained.pt \ + --bpe-model $repo/data/lang_bpe_500/bpe.model \ + $repo/test_wavs/1089-134686-0001.wav \ + $repo/test_wavs/1221-135766-0001.wav \ + $repo/test_wavs/1221-135766-0002.wav +done + +echo "GITHUB_EVENT_NAME: ${GITHUB_EVENT_NAME}" +echo "GITHUB_EVENT_LABEL_NAME: ${GITHUB_EVENT_LABEL_NAME}" +if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" == x"run-decode" ]]; then + mkdir -p pruned_transducer_stateless3/exp + ln -s $PWD/$repo/exp/pretrained.pt pruned_transducer_stateless3/exp/epoch-999.pt + ln -s $PWD/$repo/data/lang_bpe_500 data/ + + ls -lh data + ls -lh pruned_transducer_stateless3/exp + + log "Decoding test-clean and test-other" + + # use a small value for decoding with CPU + max_duration=100 + + for method in greedy_search fast_beam_search modified_beam_search; do + log "Decoding with $method" + + ./pruned_transducer_stateless3/decode.py \ + --decoding-method $method \ + --epoch 999 \ + --avg 1 \ + --max-duration $max_duration \ + --exp-dir pruned_transducer_stateless3/exp + done + + rm pruned_transducer_stateless3/exp/*.pt +fi diff --git a/.github/workflows/run-librispeech-2022-04-29.yml b/.github/workflows/run-librispeech-2022-04-29.yml index e3fe3b904..6c8188b48 100644 --- a/.github/workflows/run-librispeech-2022-04-29.yml +++ b/.github/workflows/run-librispeech-2022-04-29.yml @@ -142,8 +142,8 @@ jobs: find fast_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 echo "===modified beam search===" - find exp/modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 - find exp/modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 - name: Display decoding results for pruned_transducer_stateless3 if: github.event_name == 'schedule' || github.event.label.name == 'run-decode' @@ -161,8 +161,8 @@ jobs: find fast_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 echo "===modified beam search===" - find exp/modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 - find exp/modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 - name: Upload decoding results for pruned_transducer_stateless2 uses: actions/upload-artifact@v2 diff --git a/.github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml b/.github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml new file mode 100644 index 000000000..512f1b334 --- /dev/null +++ b/.github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml @@ -0,0 +1,151 @@ +# Copyright 2021 Fangjun Kuang (csukuangfj@gmail.com) + +# See ../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +name: run-librispeech-pruned-transducer-stateless3-2022-05-13 +# stateless pruned transducer (reworked model) + giga speech + +on: + push: + branches: + - master + pull_request: + types: [labeled] + + schedule: + # minute (0-59) + # hour (0-23) + # day of the month (1-31) + # month (1-12) + # day of the week (0-6) + # nightly build at 15:50 UTC time every day + - cron: "50 15 * * *" + +jobs: + run_librispeech_pruned_transducer_stateless3_2022_05_13: + if: github.event.label.name == 'ready' || github.event.label.name == 'run-decode' || github.event_name == 'push' || github.event_name == 'schedule' + runs-on: ${{ matrix.os }} + strategy: + matrix: + os: [ubuntu-18.04] + python-version: [3.7, 3.8, 3.9] + + fail-fast: false + + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 + + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v2 + with: + python-version: ${{ matrix.python-version }} + cache: 'pip' + cache-dependency-path: '**/requirements-ci.txt' + + - name: Install Python dependencies + run: | + grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install + + - name: Cache kaldifeat + id: my-cache + uses: actions/cache@v2 + with: + path: | + ~/tmp/kaldifeat + key: cache-tmp-${{ matrix.python-version }} + + - name: Install kaldifeat + if: steps.my-cache.outputs.cache-hit != 'true' + shell: bash + run: | + .github/scripts/install-kaldifeat.sh + + - name: Cache LibriSpeech test-clean and test-other datasets + id: libri-test-clean-and-test-other-data + uses: actions/cache@v2 + with: + path: | + ~/tmp/download + key: cache-libri-test-clean-and-test-other + + - name: Download LibriSpeech test-clean and test-other + if: steps.libri-test-clean-and-test-other-data.outputs.cache-hit != 'true' + shell: bash + run: | + .github/scripts/download-librispeech-test-clean-and-test-other-dataset.sh + + - name: Prepare manifests for LibriSpeech test-clean and test-other + shell: bash + run: | + .github/scripts/prepare-librispeech-test-clean-and-test-other-manifests.sh + + - name: Cache LibriSpeech test-clean and test-other fbank features + id: libri-test-clean-and-test-other-fbank + uses: actions/cache@v2 + with: + path: | + ~/tmp/fbank-libri + key: cache-libri-fbank-test-clean-and-test-other + + - name: Compute fbank for LibriSpeech test-clean and test-other + if: steps.libri-test-clean-and-test-other-fbank.outputs.cache-hit != 'true' + shell: bash + run: | + .github/scripts/compute-fbank-librispeech-test-clean-and-test-other.sh + + - name: Inference with pre-trained model + shell: bash + env: + GITHUB_EVENT_NAME: ${{ github.event_name }} + GITHUB_EVENT_LABEL_NAME: ${{ github.event.label.name }} + run: | + mkdir -p egs/librispeech/ASR/data + ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank + ls -lh egs/librispeech/ASR/data/* + + sudo apt-get -qq install git-lfs tree sox + export PYTHONPATH=$PWD:$PYTHONPATH + export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH + export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH + + .github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh + + - name: Display decoding results for pruned_transducer_stateless3 + if: github.event_name == 'schedule' || github.event.label.name == 'run-decode' + shell: bash + run: | + cd egs/librispeech/ASR + tree pruned_transducer_stateless3/exp + cd pruned_transducer_stateless3/exp + echo "===greedy search===" + find greedy_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 + find greedy_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 + + echo "===fast_beam_search===" + find fast_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 + find fast_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 + + echo "===modified beam search===" + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2 + find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2 + + - name: Upload decoding results for pruned_transducer_stateless3 + uses: actions/upload-artifact@v2 + if: github.event_name == 'schedule' || github.event.label.name == 'run-decode' + with: + name: torch-${{ matrix.torch }}-python-${{ matrix.python-version }}-ubuntu-18.04-cpu-pruned_transducer_stateless3-2022-04-29 + path: egs/librispeech/ASR/pruned_transducer_stateless3/exp/ diff --git a/README.md b/README.md index c4dad6aaf..47bf0e212 100644 --- a/README.md +++ b/README.md @@ -107,7 +107,7 @@ We provide a Colab notebook to run a pre-trained transducer conformer + stateles | | test-clean | test-other | |-----|------------|------------| -| WER | 2.19 | 4.97 | +| WER | 2.00 | 4.63 | ### Aishell diff --git a/egs/librispeech/ASR/RESULTS.md b/egs/librispeech/ASR/RESULTS.md index 3143fa077..874700f11 100644 --- a/egs/librispeech/ASR/RESULTS.md +++ b/egs/librispeech/ASR/RESULTS.md @@ -1,6 +1,6 @@ ## Results -### LibriSpeech BPE training results (Pruned Transducer 3) +### LibriSpeech BPE training results (Pruned Transducer 3, 2022-04-29) [pruned_transducer_stateless3](./pruned_transducer_stateless3) Same as `Pruned Transducer 2` but using the XL subset from @@ -152,6 +152,67 @@ for epoch in 27; do done ``` +### LibriSpeech BPE training results (Pruned Transducer 3, 2022-05-13) + +Same setup as [pruned_transducer_stateless3](./pruned_transducer_stateless3) (2022-04-29) +but change `--giga-prob` from 0.8 to 0.9. Also use `repeat` on gigaspeech XL +subset so that the gigaspeech dataloader never exhausts. + +| | test-clean | test-other | comment | +|-------------------------------------|------------|------------|---------------------------------------------| +| greedy search (max sym per frame 1) | 2.03 | 4.70 | --iter 1224000 --avg 14 --max-duration 600 | +| modified beam search | 2.00 | 4.63 | --iter 1224000 --avg 14 --max-duration 600 | +| fast beam search | 2.10 | 4.68 | --iter 1224000 --avg 14 --max-duration 600 | + +The training commands are: + +```bash +export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" + +./prepare.sh +./prepare_giga_speech.sh + +./pruned_transducer_stateless3/train.py \ + --world-size 8 \ + --num-epochs 30 \ + --start-epoch 0 \ + --full-libri 1 \ + --exp-dir pruned_transducer_stateless3/exp-0.9 \ + --max-duration 300 \ + --use-fp16 1 \ + --lr-epochs 4 \ + --num-workers 2 \ + --giga-prob 0.9 +``` + +The tensorboard log is available at + + +Decoding commands: + +```bash +for iter in 1224000; do + for avg in 14; do + for method in greedy_search modified_beam_search fast_beam_search ; do + ./pruned_transducer_stateless3/decode.py \ + --iter $iter \ + --avg $avg \ + --exp-dir ./pruned_transducer_stateless3/exp-0.9/ \ + --max-duration 600 \ + --decoding-method $method \ + --max-sym-per-frame 1 \ + --beam 4 \ + --max-contexts 32 + done + done +done +``` + +The pretrained models, training logs, decoding logs, and decoding results +can be found at + + + ### LibriSpeech BPE training results (Pruned Transducer 2) [pruned_transducer_stateless2](./pruned_transducer_stateless2) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless/pretrained.py b/egs/librispeech/ASR/pruned_transducer_stateless/pretrained.py index 148bf7b02..d21a737b8 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless/pretrained.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless/pretrained.py @@ -19,38 +19,38 @@ Usage: (1) greedy search ./pruned_transducer_stateless/pretrained.py \ - --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method greedy_search \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method greedy_search \ + /path/to/foo.wav \ + /path/to/bar.wav (2) beam search ./pruned_transducer_stateless/pretrained.py \ - --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method beam_search \ - --beam-size 4 \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav (3) modified beam search ./pruned_transducer_stateless/pretrained.py \ - --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method modified_beam_search \ - --beam-size 4 \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method modified_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav (4) fast beam search ./pruned_transducer_stateless/pretrained.py \ - --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method fast_beam_search \ - --beam-size 4 \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method fast_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav You can also use `./pruned_transducer_stateless/exp/epoch-xx.pt`. @@ -233,6 +233,9 @@ def main(): logging.info("Creating model") model = get_transducer_model(params) + num_param = sum([p.numel() for p in model.parameters()]) + logging.info(f"Number of model parameters: {num_param}") + checkpoint = torch.load(args.checkpoint, map_location="cpu") model.load_state_dict(checkpoint["model"], strict=False) model.to(device) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/pretrained.py b/egs/librispeech/ASR/pruned_transducer_stateless2/pretrained.py index bcafe68d6..21bcf7cfd 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/pretrained.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/pretrained.py @@ -19,20 +19,38 @@ Usage: (1) greedy search ./pruned_transducer_stateless2/pretrained.py \ - --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method greedy_search \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method greedy_search \ + /path/to/foo.wav \ + /path/to/bar.wav -(1) beam search +(2) beam search ./pruned_transducer_stateless2/pretrained.py \ - --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method beam_search \ - --beam-size 4 \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav + +(3) modified beam search +./pruned_transducer_stateless2/pretrained.py \ + --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method modified_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav + +(4) fast beam search +./pruned_transducer_stateless2/pretrained.py \ + --checkpoint ./pruned_transducer_stateless2/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method fast_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav You can also use `./pruned_transducer_stateless2/exp/epoch-xx.pt`. @@ -79,9 +97,7 @@ def get_parser(): parser.add_argument( "--bpe-model", type=str, - help="""Path to bpe.model. - Used only when method is ctc-decoding. - """, + help="""Path to bpe.model.""", ) parser.add_argument( @@ -117,7 +133,33 @@ def get_parser(): "--beam-size", type=int, default=4, - help="Used only when --method is beam_search and modified_beam_search", + help="""An integer indicating how many candidates we will keep for each + frame. Used only when --method is beam_search or + modified_beam_search.""", + ) + + parser.add_argument( + "--beam", + type=float, + default=4, + help="""A floating point value to calculate the cutoff score during beam + search (i.e., `cutoff = max-score - beam`), which is the same as the + `beam` in Kaldi. + Used only when --method is fast_beam_search""", + ) + + parser.add_argument( + "--max-contexts", + type=int, + default=4, + help="""Used only when --method is fast_beam_search""", + ) + + parser.add_argument( + "--max-states", + type=int, + default=8, + help="""Used only when --method is fast_beam_search""", ) parser.add_argument( @@ -244,9 +286,9 @@ def main(): decoding_graph=decoding_graph, encoder_out=encoder_out, encoder_out_lens=encoder_out_lens, - beam=8.0, - max_contexts=32, - max_states=8, + beam=params.beam, + max_contexts=params.max_contexts, + max_states=params.max_states, ) for hyp in sp.decode(hyp_tokens): hyps.append(hyp.split()) @@ -254,6 +296,7 @@ def main(): hyp_tokens = modified_beam_search( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, beam=params.beam_size, ) @@ -263,6 +306,7 @@ def main(): hyp_tokens = greedy_search_batch( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, ) for hyp in sp.decode(hyp_tokens): hyps.append(hyp.split()) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/pretrained.py b/egs/librispeech/ASR/pruned_transducer_stateless3/pretrained.py index d0fe5d24e..7efa592f9 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/pretrained.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/pretrained.py @@ -19,20 +19,38 @@ Usage: (1) greedy search ./pruned_transducer_stateless3/pretrained.py \ - --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method greedy_search \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method greedy_search \ + /path/to/foo.wav \ + /path/to/bar.wav -(1) beam search +(2) beam search ./pruned_transducer_stateless3/pretrained.py \ - --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --method beam_search \ - --beam-size 4 \ - /path/to/foo.wav \ - /path/to/bar.wav \ + --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav + +(3) modified beam search +./pruned_transducer_stateless3/pretrained.py \ + --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method modified_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav + +(4) fast beam search +./pruned_transducer_stateless3/pretrained.py \ + --checkpoint ./pruned_transducer_stateless3/exp/pretrained.pt \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --method fast_beam_search \ + --beam-size 4 \ + /path/to/foo.wav \ + /path/to/bar.wav You can also use `./pruned_transducer_stateless3/exp/epoch-xx.pt`. @@ -79,9 +97,7 @@ def get_parser(): parser.add_argument( "--bpe-model", type=str, - help="""Path to bpe.model. - Used only when method is ctc-decoding. - """, + help="""Path to bpe.model.""", ) parser.add_argument( @@ -117,7 +133,33 @@ def get_parser(): "--beam-size", type=int, default=4, - help="Used only when --method is beam_search and modified_beam_search", + help="""An integer indicating how many candidates we will keep for each + frame. Used only when --method is beam_search or + modified_beam_search.""", + ) + + parser.add_argument( + "--beam", + type=float, + default=4, + help="""A floating point value to calculate the cutoff score during beam + search (i.e., `cutoff = max-score - beam`), which is the same as the + `beam` in Kaldi. + Used only when --method is fast_beam_search""", + ) + + parser.add_argument( + "--max-contexts", + type=int, + default=4, + help="""Used only when --method is fast_beam_search""", + ) + + parser.add_argument( + "--max-states", + type=int, + default=8, + help="""Used only when --method is fast_beam_search""", ) parser.add_argument( @@ -244,9 +286,9 @@ def main(): decoding_graph=decoding_graph, encoder_out=encoder_out, encoder_out_lens=encoder_out_lens, - beam=8.0, - max_contexts=32, - max_states=8, + beam=params.beam, + max_contexts=params.max_contexts, + max_states=params.max_states, ) for hyp in sp.decode(hyp_tokens): hyps.append(hyp.split()) @@ -254,6 +296,7 @@ def main(): hyp_tokens = modified_beam_search( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, beam=params.beam_size, ) @@ -263,6 +306,7 @@ def main(): hyp_tokens = greedy_search_batch( model=model, encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, ) for hyp in sp.decode(hyp_tokens): hyps.append(hyp.split()) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py index 4966ea57f..037f99bc7 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py @@ -968,6 +968,7 @@ def run(rank, world_size, args): train_giga_cuts = gigaspeech.train_S_cuts() train_giga_cuts = filter_short_and_long_utterances(train_giga_cuts) + train_giga_cuts = train_giga_cuts.repeat(times=None) if args.enable_musan: cuts_musan = load_manifest( From 9630f9a3baca1406247fb5881044551b4d3b7f63 Mon Sep 17 00:00:00 2001 From: Guanbo Wang Date: Sun, 15 May 2022 00:57:40 -0400 Subject: [PATCH 07/16] Update GigaSpeech reults (#364) * Update decode.py * Update export.py * Update results * Update README.md --- README.md | 13 ++-- egs/gigaspeech/ASR/README.md | 2 +- egs/gigaspeech/ASR/RESULTS.md | 29 ++++----- .../pruned_transducer_stateless2/decode.py | 64 ++++++++++++------- .../pruned_transducer_stateless2/export.py | 42 ++++++++++-- 5 files changed, 102 insertions(+), 48 deletions(-) diff --git a/README.md b/README.md index 47bf0e212..8911a4336 100644 --- a/README.md +++ b/README.md @@ -200,19 +200,22 @@ We provide a Colab notebook to run a pre-trained Pruned Transducer Stateless mod ### GigaSpeech +We provide two models for this recipe: [Conformer CTC model][GigaSpeech_conformer_ctc] +and [Pruned stateless RNN-T: Conformer encoder + Embedding decoder + k2 pruned RNN-T loss][GigaSpeech_pruned_transducer_stateless2]. + #### Conformer CTC | | Dev | Test | |-----|-------|-------| | WER | 10.47 | 10.58 | -#### Pruned stateless RNN-T +#### Pruned stateless RNN-T: Conformer encoder + Embedding decoder + k2 pruned RNN-T loss | | Dev | Test | |----------------------|-------|-------| -| greedy search | 10.59 | 10.87 | -| fast beam search | 10.56 | 10.80 | -| modified beam search | 10.52 | 10.62 | +| greedy search | 10.51 | 10.73 | +| fast beam search | 10.50 | 10.69 | +| modified beam search | 10.40 | 10.51 | ## Deployment with C++ @@ -238,6 +241,8 @@ Please see: [![Open In Colab](https://colab.research.google.com/assets/colab-bad [TIMIT_tdnn_ligru_ctc]: egs/timit/ASR/tdnn_ligru_ctc [TED-LIUM3_transducer_stateless]: egs/tedlium3/ASR/transducer_stateless [TED-LIUM3_pruned_transducer_stateless]: egs/tedlium3/ASR/pruned_transducer_stateless +[GigaSpeech_conformer_ctc]: egs/gigaspeech/ASR/conformer_ctc +[GigaSpeech_pruned_transducer_stateless2]: egs/gigaspeech/ASR/pruned_transducer_stateless2 [yesno]: egs/yesno/ASR [librispeech]: egs/librispeech/ASR [aishell]: egs/aishell/ASR diff --git a/egs/gigaspeech/ASR/README.md b/egs/gigaspeech/ASR/README.md index 1fca69e8b..32a0457c6 100644 --- a/egs/gigaspeech/ASR/README.md +++ b/egs/gigaspeech/ASR/README.md @@ -16,6 +16,6 @@ ln -sfv /path/to/GigaSpeech download/GigaSpeech | | Dev | Test | |--------------------------------|-------|-------| | `conformer_ctc` | 10.47 | 10.58 | -| `pruned_transducer_stateless2` | 10.52 | 10.62 | +| `pruned_transducer_stateless2` | 10.40 | 10.51 | See [RESULTS](/egs/gigaspeech/ASR/RESULTS.md) for details. diff --git a/egs/gigaspeech/ASR/RESULTS.md b/egs/gigaspeech/ASR/RESULTS.md index de7b84202..7ab565844 100644 --- a/egs/gigaspeech/ASR/RESULTS.md +++ b/egs/gigaspeech/ASR/RESULTS.md @@ -11,13 +11,15 @@ decoder contains only an embedding layer, a Conv1d (with kernel size 2) and a linear layer (to transform tensor dim). k2 pruned RNN-T loss is used. +The best WER, as of 2022-05-12, for the gigaspeech is below + Results are: | | Dev | Test | |----------------------|-------|-------| -| greedy search | 10.59 | 10.87 | -| fast beam search | 10.56 | 10.80 | -| modified beam search | 10.52 | 10.62 | +| greedy search | 10.51 | 10.73 | +| fast beam search | 10.50 | 10.69 | +| modified beam search | 10.40 | 10.51 | To reproduce the above result, use the following commands for training: @@ -39,33 +41,30 @@ and the following commands for decoding: ```bash # greedy search ./pruned_transducer_stateless2/decode.py \ - --epoch 29 \ - --avg 11 \ + --iter 3488000 \ + --avg 20 \ --decoding-method greedy_search \ --exp-dir pruned_transducer_stateless2/exp \ --bpe-model data/lang_bpe_500/bpe.model \ - --max-duration 20 \ - --num-workers 1 + --max-duration 600 # fast beam search ./pruned_transducer_stateless2/decode.py \ - --epoch 29 \ - --avg 9 \ + --iter 3488000 \ + --avg 20 \ --decoding-method fast_beam_search \ --exp-dir pruned_transducer_stateless2/exp \ --bpe-model data/lang_bpe_500/bpe.model \ - --max-duration 20 \ - --num-workers 1 + --max-duration 600 # modified beam search ./pruned_transducer_stateless2/decode.py \ - --epoch 29 \ - --avg 8 \ + --iter 3488000 \ + --avg 15 \ --decoding-method modified_beam_search \ --exp-dir pruned_transducer_stateless2/exp \ --bpe-model data/lang_bpe_500/bpe.model \ - --max-duration 20 \ - --num-workers 1 + --max-duration 600 ``` Pretrained model is available at diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py index 92a5b0b28..ce5116336 100755 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py @@ -22,7 +22,7 @@ Usage: --epoch 28 \ --avg 15 \ --exp-dir ./pruned_transducer_stateless2/exp \ - --max-duration 100 \ + --max-duration 600 \ --decoding-method greedy_search (2) beam search @@ -30,7 +30,7 @@ Usage: --epoch 28 \ --avg 15 \ --exp-dir ./pruned_transducer_stateless2/exp \ - --max-duration 100 \ + --max-duration 600 \ --decoding-method beam_search \ --beam-size 4 @@ -39,7 +39,7 @@ Usage: --epoch 28 \ --avg 15 \ --exp-dir ./pruned_transducer_stateless2/exp \ - --max-duration 100 \ + --max-duration 600 \ --decoding-method modified_beam_search \ --beam-size 4 @@ -48,7 +48,7 @@ Usage: --epoch 28 \ --avg 15 \ --exp-dir ./pruned_transducer_stateless2/exp \ - --max-duration 1500 \ + --max-duration 600 \ --decoding-method fast_beam_search \ --beam 4 \ --max-contexts 4 \ @@ -99,27 +99,28 @@ def get_parser(): "--epoch", type=int, default=29, - help="It specifies the checkpoint to use for decoding." - "Note: Epoch counts from 0.", + help="""It specifies the checkpoint to use for decoding. + Note: Epoch counts from 0. + You can specify --avg to use more checkpoints for model averaging.""", ) + + parser.add_argument( + "--iter", + type=int, + default=0, + help="""If positive, --epoch is ignored and it + will use the checkpoint exp_dir/checkpoint-iter.pt. + You can specify --avg to use more checkpoints for model averaging. + """, + ) + parser.add_argument( "--avg", type=int, default=8, help="Number of checkpoints to average. Automatically select " "consecutive checkpoints before the checkpoint specified by " - "'--epoch'. ", - ) - - parser.add_argument( - "--avg-last-n", - type=int, - default=0, - help="""If positive, --epoch and --avg are ignored and it - will use the last n checkpoints exp_dir/checkpoint-xxx.pt - where xxx is the number of processed batches while - saving that checkpoint. - """, + "'--epoch' and '--iter'", ) parser.add_argument( @@ -152,7 +153,7 @@ def get_parser(): "--beam-size", type=int, default=4, - help="""An interger indicating how many candidates we will keep for each + help="""An integer indicating how many candidates we will keep for each frame. Used only when --decoding-method is beam_search or modified_beam_search.""", ) @@ -465,7 +466,11 @@ def main(): ) params.res_dir = params.exp_dir / params.decoding_method - params.suffix = f"epoch-{params.epoch}-avg-{params.avg}" + if params.iter > 0: + params.suffix = f"iter-{params.iter}-avg-{params.avg}" + else: + params.suffix = f"epoch-{params.epoch}-avg-{params.avg}" + if "fast_beam_search" in params.decoding_method: params.suffix += f"-beam-{params.beam}" params.suffix += f"-max-contexts-{params.max_contexts}" @@ -488,8 +493,9 @@ def main(): sp = spm.SentencePieceProcessor() sp.load(params.bpe_model) - # is defined in local/train_bpe_model.py + # and is defined in local/train_bpe_model.py params.blank_id = sp.piece_to_id("") + params.unk_id = sp.piece_to_id("") params.vocab_size = sp.get_piece_size() logging.info(params) @@ -497,8 +503,20 @@ def main(): logging.info("About to create model") model = get_transducer_model(params) - if params.avg_last_n > 0: - filenames = find_checkpoints(params.exp_dir)[: params.avg_last_n] + if params.iter > 0: + filenames = find_checkpoints(params.exp_dir, iteration=-params.iter)[ + : params.avg + ] + if len(filenames) == 0: + raise ValueError( + f"No checkpoints found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + elif len(filenames) < params.avg: + raise ValueError( + f"Not enough checkpoints ({len(filenames)}) found for" + f" --iter {params.iter}, --avg {params.avg}" + ) logging.info(f"averaging {filenames}") model.to(device) model.load_state_dict(average_checkpoints(filenames, device=device)) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py index b5757ee8c..6b3a7a9ff 100755 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/export.py @@ -51,7 +51,11 @@ import sentencepiece as spm import torch from train import get_params, get_transducer_model -from icefall.checkpoint import average_checkpoints, load_checkpoint +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) from icefall.utils import str2bool @@ -64,8 +68,19 @@ def get_parser(): "--epoch", type=int, default=28, - help="It specifies the checkpoint to use for decoding." - "Note: Epoch counts from 0.", + help="""It specifies the checkpoint to use for averaging. + Note: Epoch counts from 0. + You can specify --avg to use more checkpoints for model averaging.""", + ) + + parser.add_argument( + "--iter", + type=int, + default=0, + help="""If positive, --epoch is ignored and it + will use the checkpoint exp_dir/checkpoint-iter.pt. + You can specify --avg to use more checkpoints for model averaging. + """, ) parser.add_argument( @@ -74,7 +89,7 @@ def get_parser(): default=15, help="Number of checkpoints to average. Automatically select " "consecutive checkpoints before the checkpoint specified by " - "'--epoch'. ", + "'--epoch' and '--iter'", ) parser.add_argument( @@ -141,7 +156,24 @@ def main(): model.to(device) - if params.avg == 1: + if params.iter > 0: + filenames = find_checkpoints(params.exp_dir, iteration=-params.iter)[ + : params.avg + ] + if len(filenames) == 0: + raise ValueError( + f"No checkpoints found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + elif len(filenames) < params.avg: + raise ValueError( + f"Not enough checkpoints ({len(filenames)}) found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) else: start = params.epoch - params.avg + 1 From 6f7860a0a60b53026216fa4ba19048955951333e Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Sun, 15 May 2022 14:25:35 +0800 Subject: [PATCH 08/16] Fix GitHub CI for decoding GigaSpeech dev/test datasets (#366) --- .../run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh b/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh index 25bac2fdb..528d04cd1 100755 --- a/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh +++ b/.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh @@ -19,7 +19,7 @@ echo "GITHUB_EVENT_NAME: ${GITHUB_EVENT_NAME}" echo "GITHUB_EVENT_LABEL_NAME: ${GITHUB_EVENT_LABEL_NAME}" if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" == x"run-decode" ]]; then mkdir -p pruned_transducer_stateless2/exp - ln -s $PWD/$repo/exp/pretrained-epoch-29-avg-11.pt pruned_transducer_stateless2/exp/epoch-999.pt + ln -s $PWD/$repo/exp/pretrained-iter-3488000-avg-20.pt pruned_transducer_stateless2/exp/epoch-999.pt ln -s $PWD/$repo/data/lang_bpe_500 data/ ls -lh data From 5aafbb970e8b9195d5b6ed3b86da655d9ff2ccc4 Mon Sep 17 00:00:00 2001 From: Desh Raj Date: Mon, 16 May 2022 08:52:14 -0400 Subject: [PATCH 09/16] SPGISpeech recipe (#334) * initial commit for SPGISpeech recipe * add decoding * add spgispeech transducer * remove conformer ctc; minor fixes in RNN-T * add results * add tensorboard * add pretrained model to HF * remove unused scripts and soft link common scripts * remove duplicate files * pre commit hooks * remove change in librispeech * pre commit hook * add CER numbers --- egs/spgispeech/ASR/README.md | 32 + egs/spgispeech/ASR/RESULTS.md | 73 ++ egs/spgispeech/ASR/local/__init__.py | 0 egs/spgispeech/ASR/local/compile_hlg.py | 1 + .../ASR/local/compute_fbank_musan.py | 104 ++ .../ASR/local/compute_fbank_spgispeech.py | 145 +++ egs/spgispeech/ASR/local/prepare_lang.py | 1 + egs/spgispeech/ASR/local/prepare_lang_bpe.py | 1 + egs/spgispeech/ASR/local/prepare_splits.py | 79 ++ egs/spgispeech/ASR/local/train_bpe_model.py | 1 + egs/spgispeech/ASR/prepare.sh | 196 ++++ .../pruned_transducer_stateless2/__init__.py | 0 .../asr_datamodule.py | 366 ++++++ .../beam_search.py | 1 + .../pruned_transducer_stateless2/conformer.py | 1 + .../pruned_transducer_stateless2/decode.py | 594 ++++++++++ .../pruned_transducer_stateless2/decoder.py | 1 + .../encoder_interface.py | 1 + .../pruned_transducer_stateless2/export.py | 201 ++++ .../pruned_transducer_stateless2/joiner.py | 1 + .../ASR/pruned_transducer_stateless2/model.py | 1 + .../ASR/pruned_transducer_stateless2/optim.py | 1 + .../pruned_transducer_stateless2/scaling.py | 1 + .../ASR/pruned_transducer_stateless2/train.py | 1031 +++++++++++++++++ egs/spgispeech/ASR/shared | 1 + 25 files changed, 2834 insertions(+) create mode 100644 egs/spgispeech/ASR/README.md create mode 100644 egs/spgispeech/ASR/RESULTS.md create mode 100644 egs/spgispeech/ASR/local/__init__.py create mode 120000 egs/spgispeech/ASR/local/compile_hlg.py create mode 100755 egs/spgispeech/ASR/local/compute_fbank_musan.py create mode 100755 egs/spgispeech/ASR/local/compute_fbank_spgispeech.py create mode 120000 egs/spgispeech/ASR/local/prepare_lang.py create mode 120000 egs/spgispeech/ASR/local/prepare_lang_bpe.py create mode 100755 egs/spgispeech/ASR/local/prepare_splits.py create mode 120000 egs/spgispeech/ASR/local/train_bpe_model.py create mode 100755 egs/spgispeech/ASR/prepare.sh create mode 100644 egs/spgispeech/ASR/pruned_transducer_stateless2/__init__.py create mode 100644 egs/spgispeech/ASR/pruned_transducer_stateless2/asr_datamodule.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/beam_search.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/conformer.py create mode 100755 egs/spgispeech/ASR/pruned_transducer_stateless2/decode.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/decoder.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/encoder_interface.py create mode 100755 egs/spgispeech/ASR/pruned_transducer_stateless2/export.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/joiner.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/model.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/optim.py create mode 120000 egs/spgispeech/ASR/pruned_transducer_stateless2/scaling.py create mode 100755 egs/spgispeech/ASR/pruned_transducer_stateless2/train.py create mode 120000 egs/spgispeech/ASR/shared diff --git a/egs/spgispeech/ASR/README.md b/egs/spgispeech/ASR/README.md new file mode 100644 index 000000000..f60408cc1 --- /dev/null +++ b/egs/spgispeech/ASR/README.md @@ -0,0 +1,32 @@ +# SPGISpeech + +SPGISpeech consists of 5,000 hours of recorded company earnings calls and their respective +transcriptions. The original calls were split into slices ranging from 5 to 15 seconds in +length to allow easy training for speech recognition systems. Calls represent a broad +cross-section of international business English; SPGISpeech contains approximately 50,000 +speakers, one of the largest numbers of any speech corpus, and offers a variety of L1 and +L2 English accents. The format of each WAV file is single channel, 16kHz, 16 bit audio. + +Transcription text represents the output of several stages of manual post-processing. +As such, the text contains polished English orthography following a detailed style guide, +including proper casing, punctuation, and denormalized non-standard words such as numbers +and acronyms, making SPGISpeech suited for training fully formatted end-to-end models. + +Official reference: + +O’Neill, P.K., Lavrukhin, V., Majumdar, S., Noroozi, V., Zhang, Y., Kuchaiev, O., Balam, +J., Dovzhenko, Y., Freyberg, K., Shulman, M.D., Ginsburg, B., Watanabe, S., & Kucsko, G. +(2021). SPGISpeech: 5, 000 hours of transcribed financial audio for fully formatted +end-to-end speech recognition. ArXiv, abs/2104.02014. + +ArXiv link: https://arxiv.org/abs/2104.02014 + +## Performance Record + +| Decoding method | val WER | val CER | +|---------------------------|------------|---------| +| greedy search | 2.40 | 0.99 | +| modified beam search | 2.24 | 0.91 | +| fast beam search | 2.35 | 0.97 | + +See [RESULTS](/egs/spgispeech/ASR/RESULTS.md) for details. diff --git a/egs/spgispeech/ASR/RESULTS.md b/egs/spgispeech/ASR/RESULTS.md new file mode 100644 index 000000000..de9e35c5a --- /dev/null +++ b/egs/spgispeech/ASR/RESULTS.md @@ -0,0 +1,73 @@ +## Results + +### SPGISpeech BPE training results (Pruned Transducer) + +#### 2022-05-11 + +#### Conformer encoder + embedding decoder + +Conformer encoder + non-current decoder. The decoder +contains only an embedding layer, a Conv1d (with kernel size 2) and a linear +layer (to transform tensor dim). + +The WERs are + +| | dev | val | comment | +|---------------------------|------------|------------|------------------------------------------| +| greedy search | 2.46 | 2.40 | --avg-last-n 10 --max-duration 500 | +| modified beam search | 2.28 | 2.24 | --avg-last-n 10 --max-duration 500 --beam-size 4 | +| fast beam search | 2.38 | 2.35 | --avg-last-n 10 --max-duration 500 --beam-size 4 --max-contexts 4 --max-states 8 | + +**NOTE:** SPGISpeech transcripts can be prepared in `ortho` or `norm` ways, which refer to whether the +transcripts are orthographic or normalized. These WERs correspond to the normalized transcription +scenario. + +The training command for reproducing is given below: + +``` +export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" + +./pruned_transducer_stateless2/train.py \ + --world-size 8 \ + --num-epochs 20 \ + --start-epoch 0 \ + --exp-dir pruned_transducer_stateless2/exp \ + --max-duration 200 \ + --prune-range 5 \ + --lr-factor 5 \ + --lm-scale 0.25 \ + --use-fp16 True +``` + +The decoding command is: +``` +# greedy search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method greedy_search + +# modified beam search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method modified_beam_search \ + --beam-size 4 + +# fast beam search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 1500 \ + --decoding-method fast_beam_search \ + --beam 4 \ + --max-contexts 4 \ + --max-states 8 +``` + +Pretrained model is available at + +The tensorboard training log can be found at + diff --git a/egs/spgispeech/ASR/local/__init__.py b/egs/spgispeech/ASR/local/__init__.py new file mode 100644 index 000000000..e69de29bb diff --git a/egs/spgispeech/ASR/local/compile_hlg.py b/egs/spgispeech/ASR/local/compile_hlg.py new file mode 120000 index 000000000..471aa7fb4 --- /dev/null +++ b/egs/spgispeech/ASR/local/compile_hlg.py @@ -0,0 +1 @@ +../../../librispeech/ASR/local/compile_hlg.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/local/compute_fbank_musan.py b/egs/spgispeech/ASR/local/compute_fbank_musan.py new file mode 100755 index 000000000..b4f409ba6 --- /dev/null +++ b/egs/spgispeech/ASR/local/compute_fbank_musan.py @@ -0,0 +1,104 @@ +#!/usr/bin/env python3 +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +This file computes fbank features of the musan dataset. +It looks for manifests in the directory data/manifests. + +The generated fbank features are saved in data/fbank. +""" + +import logging +from pathlib import Path + +import torch +from lhotse import LilcomChunkyWriter, CutSet, combine +from lhotse.features.kaldifeat import ( + KaldifeatFbank, + KaldifeatFbankConfig, + KaldifeatMelOptions, + KaldifeatFrameOptions, +) +from lhotse.recipes.utils import read_manifests_if_cached + +from icefall.utils import get_executor + +# Torch's multithreaded behavior needs to be disabled or +# it wastes a lot of CPU and slow things down. +# Do this outside of main() in case it needs to take effect +# even when we are not invoking the main (e.g. when spawning subprocesses). +torch.set_num_threads(1) +torch.set_num_interop_threads(1) + + +def compute_fbank_musan(): + src_dir = Path("data/manifests") + output_dir = Path("data/fbank") + + sampling_rate = 16000 + num_mel_bins = 80 + + extractor = KaldifeatFbank( + KaldifeatFbankConfig( + frame_opts=KaldifeatFrameOptions(sampling_rate=sampling_rate), + mel_opts=KaldifeatMelOptions(num_bins=num_mel_bins), + device="cuda", + ) + ) + + dataset_parts = ( + "music", + "speech", + "noise", + ) + manifests = read_manifests_if_cached( + dataset_parts=dataset_parts, output_dir=src_dir + ) + assert manifests is not None + + musan_cuts_path = src_dir / "cuts_musan.jsonl.gz" + + if musan_cuts_path.is_file(): + logging.info(f"{musan_cuts_path} already exists - skipping") + return + + logging.info("Extracting features for Musan") + + # create chunks of Musan with duration 5 - 10 seconds + musan_cuts = ( + CutSet.from_manifests( + recordings=combine(part["recordings"] for part in manifests.values()) + ) + .cut_into_windows(10.0) + .filter(lambda c: c.duration > 5) + .compute_and_store_features_batch( + extractor=extractor, + storage_path=output_dir / f"feats_musan", + manifest_path=src_dir / f"cuts_musan.jsonl.gz", + batch_duration=500, + num_workers=4, + storage_type=LilcomChunkyWriter, + ) + ) + + +if __name__ == "__main__": + formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + + logging.basicConfig(format=formatter, level=logging.INFO) + compute_fbank_musan() diff --git a/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py b/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py new file mode 100755 index 000000000..cc8c8a670 --- /dev/null +++ b/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py @@ -0,0 +1,145 @@ +#!/usr/bin/env python3 +# Copyright 2022 Johns Hopkins University (authors: Desh Raj) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +This file computes fbank features of the SPGISpeech dataset. +It looks for manifests in the directory data/manifests. + +The generated fbank features are saved in data/fbank. +""" +import argparse +import logging +from pathlib import Path +from tqdm import tqdm + +import torch +from lhotse import load_manifest_lazy, LilcomChunkyWriter +from lhotse.features.kaldifeat import ( + KaldifeatFbank, + KaldifeatFbankConfig, + KaldifeatMelOptions, + KaldifeatFrameOptions, +) +from lhotse.manipulation import combine + +# Torch's multithreaded behavior needs to be disabled or +# it wastes a lot of CPU and slow things down. +# Do this outside of main() in case it needs to take effect +# even when we are not invoking the main (e.g. when spawning subprocesses). +torch.set_num_threads(1) +torch.set_num_interop_threads(1) + + +def get_args(): + parser = argparse.ArgumentParser() + parser.add_argument( + "--num-splits", + type=int, + default=20, + help="Number of splits for the train set.", + ) + parser.add_argument( + "--start", + type=int, + default=0, + help="Start index of the train set split.", + ) + parser.add_argument( + "--stop", + type=int, + default=-1, + help="Stop index of the train set split.", + ) + parser.add_argument( + "--test", + action="store_true", + help="If set, only compute features for the dev and val set.", + ) + parser.add_argument( + "--train", + action="store_true", + help="If set, only compute features for the train set.", + ) + + return parser.parse_args() + + +def compute_fbank_spgispeech(args): + assert args.train or args.test, "Either train or test must be set." + + src_dir = Path("data/manifests") + output_dir = Path("data/fbank") + + sampling_rate = 16000 + num_mel_bins = 80 + + extractor = KaldifeatFbank( + KaldifeatFbankConfig( + frame_opts=KaldifeatFrameOptions(sampling_rate=sampling_rate), + mel_opts=KaldifeatMelOptions(num_bins=num_mel_bins), + device="cuda", + ) + ) + + if args.train: + logging.info(f"Processing train") + cut_set = load_manifest_lazy(src_dir / f"cuts_train_raw.jsonl.gz") + chunk_size = len(cut_set) // args.num_splits + cut_sets = cut_set.split_lazy( + output_dir=src_dir / f"cuts_train_raw_split{args.num_splits}", + chunk_size=chunk_size, + ) + start = args.start + stop = min(args.stop, args.num_splits) if args.stop > 0 else args.num_splits + num_digits = len(str(args.num_splits)) + for i in range(start, stop): + idx = f"{i + 1}".zfill(num_digits) + logging.info(f"Processing train split {i}") + cs = cut_sets[i].compute_and_store_features_batch( + extractor=extractor, + storage_path=output_dir / f"feats_train_{idx}", + manifest_path=src_dir / f"cuts_train_{idx}.jsonl.gz", + batch_duration=500, + num_workers=4, + storage_type=LilcomChunkyWriter, + ) + + if args.test: + for partition in ["dev", "val"]: + if (output_dir / f"cuts_{partition}.jsonl.gz").is_file(): + logging.info(f"{partition} already exists - skipping.") + continue + logging.info(f"Processing {partition}") + cut_set = load_manifest_lazy(src_dir / f"cuts_{partition}_raw.jsonl.gz") + cut_set = cut_set.compute_and_store_features_batch( + extractor=extractor, + storage_path=output_dir / f"feats_{partition}", + manifest_path=src_dir / f"cuts_{partition}.jsonl.gz", + batch_duration=500, + num_workers=4, + storage_type=LilcomChunkyWriter, + ) + + +if __name__ == "__main__": + formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + + logging.basicConfig(format=formatter, level=logging.INFO) + + args = get_args() + compute_fbank_spgispeech(args) diff --git a/egs/spgispeech/ASR/local/prepare_lang.py b/egs/spgispeech/ASR/local/prepare_lang.py new file mode 120000 index 000000000..747f2ab39 --- /dev/null +++ b/egs/spgispeech/ASR/local/prepare_lang.py @@ -0,0 +1 @@ +../../../librispeech/ASR/local/prepare_lang.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/local/prepare_lang_bpe.py b/egs/spgispeech/ASR/local/prepare_lang_bpe.py new file mode 120000 index 000000000..36b40e7fc --- /dev/null +++ b/egs/spgispeech/ASR/local/prepare_lang_bpe.py @@ -0,0 +1 @@ +../../../librispeech/ASR/local/prepare_lang_bpe.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/local/prepare_splits.py b/egs/spgispeech/ASR/local/prepare_splits.py new file mode 100755 index 000000000..2d1818649 --- /dev/null +++ b/egs/spgispeech/ASR/local/prepare_splits.py @@ -0,0 +1,79 @@ +#!/usr/bin/env python3 +# Copyright 2022 Johns Hopkins University (authors: Desh Raj) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +This file splits the training set into train and dev sets. +""" +import logging +from pathlib import Path + +import torch +from lhotse import CutSet + +from lhotse.recipes.utils import read_manifests_if_cached + +# Torch's multithreaded behavior needs to be disabled or +# it wastes a lot of CPU and slow things down. +# Do this outside of main() in case it needs to take effect +# even when we are not invoking the main (e.g. when spawning subprocesses). +torch.set_num_threads(1) +torch.set_num_interop_threads(1) + + +def split_spgispeech_train(): + src_dir = Path("data/manifests") + + manifests = read_manifests_if_cached( + dataset_parts=["train", "val"], + output_dir=src_dir, + prefix="spgispeech", + suffix="jsonl.gz", + lazy=True, + ) + assert manifests is not None + + train_dev_cuts = CutSet.from_manifests( + recordings=manifests["train"]["recordings"], + supervisions=manifests["train"]["supervisions"], + ) + dev_cuts = train_dev_cuts.subset(first=4000) + train_cuts = train_dev_cuts.filter(lambda c: c not in dev_cuts) + + # Add speed perturbation + train_cuts = ( + train_cuts + train_cuts.perturb_speed(0.9) + train_cuts.perturb_speed(1.1) + ) + + # Write the manifests to disk. + train_cuts.to_file(src_dir / "cuts_train_raw.jsonl.gz") + dev_cuts.to_file(src_dir / "cuts_dev_raw.jsonl.gz") + + # Also write the val set to disk. + val_cuts = CutSet.from_manifests( + recordings=manifests["val"]["recordings"], + supervisions=manifests["val"]["supervisions"], + ) + val_cuts.to_file(src_dir / "cuts_val_raw.jsonl.gz") + + +if __name__ == "__main__": + formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + + logging.basicConfig(format=formatter, level=logging.INFO) + + split_spgispeech_train() diff --git a/egs/spgispeech/ASR/local/train_bpe_model.py b/egs/spgispeech/ASR/local/train_bpe_model.py new file mode 120000 index 000000000..6fad36421 --- /dev/null +++ b/egs/spgispeech/ASR/local/train_bpe_model.py @@ -0,0 +1 @@ +../../../librispeech/ASR/local/train_bpe_model.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/prepare.sh b/egs/spgispeech/ASR/prepare.sh new file mode 100755 index 000000000..231ebd742 --- /dev/null +++ b/egs/spgispeech/ASR/prepare.sh @@ -0,0 +1,196 @@ +#!/usr/bin/env bash + +set -eou pipefail + +nj=20 +stage=-1 +stop_stage=100 + +# We assume dl_dir (download dir) contains the following +# directories and files. If not, they will be downloaded +# by this script automatically. +# +# - $dl_dir/spgispeech +# You can find train.csv, val.csv, train, and val in this directory, which belong +# to the SPGISpeech dataset. +# +# - $dl_dir/musan +# This directory contains the following directories downloaded from +# http://www.openslr.org/17/ +# +# - music +# - noise +# - speech +dl_dir=$PWD/download + +. shared/parse_options.sh || exit 1 + +# vocab size for sentence piece models. +# It will generate data/lang_bpe_xxx, +# data/lang_bpe_yyy if the array contains xxx, yyy +vocab_sizes=( + 500 +) + +# All files generated by this script are saved in "data". +# You can safely remove "data" and rerun this script to regenerate it. +mkdir -p data + +log() { + # This function is from espnet + local fname=${BASH_SOURCE[1]##*/} + echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*" +} + +log "dl_dir: $dl_dir" + +if [ $stage -le 0 ] && [ $stop_stage -ge 0 ]; then + log "Stage 0: Download data" + + # If you have pre-downloaded it to /path/to/spgispeech, + # you can create a symlink + # + # ln -sfv /path/to/spgispeech $dl_dir/spgispeech + # + if [ ! -d $dl_dir/spgispeech/train.csv ]; then + lhotse download spgispeech $dl_dir + fi + + # If you have pre-downloaded it to /path/to/musan, + # you can create a symlink + # + # ln -sfv /path/to/musan $dl_dir/ + # + if [ ! -d $dl_dir/musan ]; then + lhotse download musan $dl_dir + fi +fi + +if [ $stage -le 1 ] && [ $stop_stage -ge 1 ]; then + log "Stage 1: Prepare SPGISpeech manifest (may take ~1h)" + # We assume that you have downloaded the SPGISpeech corpus + # to $dl_dir/spgispeech. We perform text normalization for the transcripts. + mkdir -p data/manifests + lhotse prepare spgispeech -j $nj --normalize-text $dl_dir/spgispeech data/manifests +fi + +if [ $stage -le 2 ] && [ $stop_stage -ge 2 ]; then + log "Stage 2: Prepare musan manifest" + # We assume that you have downloaded the musan corpus + # to data/musan + mkdir -p data/manifests + lhotse prepare musan $dl_dir/musan data/manifests + lhotse combine data/manifests/recordings_{music,speech,noise}.json data/manifests/recordings_musan.jsonl.gz + lhotse cut simple -r data/manifests/recordings_musan.jsonl.gz data/manifests/cuts_musan_raw.jsonl.gz +fi + +if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then + log "Stage 3: Split train into train and dev and create cut sets." + python local/prepare_splits.py +fi + +if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then + log "Stage 4: Compute fbank features for spgispeech dev and val" + mkdir -p data/fbank + python local/compute_fbank_spgispeech.py --test +fi + +if [ $stage -le 5 ] && [ $stop_stage -ge 5 ]; then + log "Stage 5: Compute fbank features for train" + mkdir -p data/fbank + python local/compute_fbank_spgispeech.py --train --num-splits 20 + + log "Combine features from train splits (may take ~1h)" + if [ ! -f data/manifests/cuts_train.jsonl.gz ]; then + pieces=$(find data/manifests -name "cuts_train_[0-9]*.jsonl.gz") + lhotse combine $pieces data/manifests/cuts_train.jsonl.gz + fi + gunzip -c data/manifests/train_cuts.jsonl.gz | shuf | gzip -c > data/manifests/train_cuts_shuf.jsonl.gz +fi + +if [ $stage -le 6 ] && [ $stop_stage -ge 6 ]; then + log "Stage 6: Compute fbank features for musan" + mkdir -p data/fbank + python local/compute_fbank_musan.py +fi + +if [ $stage -le 7 ] && [ $stop_stage -ge 7 ]; then + log "Stage 7: Dump transcripts for LM training" + mkdir -p data/lm + gunzip -c data/manifests/cuts_train_raw.jsonl.gz \ + | jq '.supervisions[0].text' \ + | sed 's:"::g' \ + > data/lm/transcript_words.txt +fi + +if [ $stage -le 8 ] && [ $stop_stage -ge 8 ]; then + log "Stage 8: Prepare BPE based lang" + + for vocab_size in ${vocab_sizes[@]}; do + lang_dir=data/lang_bpe_${vocab_size} + mkdir -p $lang_dir + + # Add special words to words.txt + echo " 0" > $lang_dir/words.txt + echo "!SIL 1" >> $lang_dir/words.txt + echo "[UNK] 2" >> $lang_dir/words.txt + + # Add regular words to words.txt + gunzip -c data/manifests/cuts_train_raw.jsonl.gz \ + | jq '.supervisions[0].text' \ + | sed 's:"::g' \ + | sed 's: :\n:g' \ + | sort \ + | uniq \ + | sed '/^$/d' \ + | awk '{print $0,NR+2}' \ + >> $lang_dir/words.txt + + # Add remaining special word symbols expected by LM scripts. + num_words=$(cat $lang_dir/words.txt | wc -l) + echo " ${num_words}" >> $lang_dir/words.txt + num_words=$(cat $lang_dir/words.txt | wc -l) + echo " ${num_words}" >> $lang_dir/words.txt + num_words=$(cat $lang_dir/words.txt | wc -l) + echo "#0 ${num_words}" >> $lang_dir/words.txt + + ./local/train_bpe_model.py \ + --lang-dir $lang_dir \ + --vocab-size $vocab_size \ + --transcript data/lm/transcript_words.txt + + if [ ! -f $lang_dir/L_disambig.pt ]; then + ./local/prepare_lang_bpe.py --lang-dir $lang_dir + fi + done +fi + +if [ $stage -le 9 ] && [ $stop_stage -ge 9 ]; then + log "Stage 9: Train LM" + lm_dir=data/lm + + if [ ! -f $lm_dir/G.arpa ]; then + ./shared/make_kn_lm.py \ + -ngram-order 3 \ + -text $lm_dir/transcript_words.txt \ + -lm $lm_dir/G.arpa + fi + + if [ ! -f $lm_dir/G_3_gram.fst.txt ]; then + python3 -m kaldilm \ + --read-symbol-table="data/lang_phone/words.txt" \ + --disambig-symbol='#0' \ + --max-order=3 \ + $lm_dir/G.arpa > $lm_dir/G_3_gram.fst.txt + fi +fi + +if [ $stage -le 10 ] && [ $stop_stage -ge 10 ]; then + log "Stage 10: Compile HLG" + ./local/compile_hlg.py --lang-dir data/lang_phone + + for vocab_size in ${vocab_sizes[@]}; do + lang_dir=data/lang_bpe_${vocab_size} + ./local/compile_hlg.py --lang-dir $lang_dir + done +fi diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/__init__.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/__init__.py new file mode 100644 index 000000000..e69de29bb diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/asr_datamodule.py new file mode 100644 index 000000000..f165f6e60 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/asr_datamodule.py @@ -0,0 +1,366 @@ +# Copyright 2021 Piotr Żelasko +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +import argparse +import logging +from functools import lru_cache +from pathlib import Path +from typing import Any, Dict, Optional + +import torch +from lhotse import CutSet, Fbank, FbankConfig, load_manifest, load_manifest_lazy +from lhotse.dataset import ( + CutConcatenate, + CutMix, + DynamicBucketingSampler, + K2SpeechRecognitionDataset, + PrecomputedFeatures, + SpecAugment, +) +from lhotse.dataset.input_strategies import OnTheFlyFeatures +from lhotse.utils import fix_random_seed +from torch.utils.data import DataLoader +from tqdm import tqdm + +from icefall.utils import str2bool + + +class _SeedWorkers: + def __init__(self, seed: int): + self.seed = seed + + def __call__(self, worker_id: int): + fix_random_seed(self.seed + worker_id) + + +class SPGISpeechAsrDataModule: + """ + DataModule for k2 ASR experiments. + It assumes there is always one train and valid dataloader, + but there can be multiple test dataloaders (e.g. LibriSpeech test-clean + and test-other). + It contains all the common data pipeline modules used in ASR + experiments, e.g.: + - dynamic batch size, + - bucketing samplers, + - cut concatenation, + - augmentation, + - on-the-fly feature extraction + This class should be derived for specific corpora used in ASR tasks. + """ + + def __init__(self, args: argparse.Namespace): + self.args = args + + @classmethod + def add_arguments(cls, parser: argparse.ArgumentParser): + group = parser.add_argument_group( + title="ASR data related options", + description="These options are used for the preparation of " + "PyTorch DataLoaders from Lhotse CutSet's -- they control the " + "effective batch sizes, sampling strategies, applied data " + "augmentations, etc.", + ) + group.add_argument( + "--manifest-dir", + type=Path, + default=Path("data/manifests"), + help="Path to directory with train/valid/test cuts.", + ) + group.add_argument( + "--enable-musan", + type=str2bool, + default=True, + help="When enabled, select noise from MUSAN and mix it " + "with training dataset. ", + ) + group.add_argument( + "--concatenate-cuts", + type=str2bool, + default=False, + help="When enabled, utterances (cuts) will be concatenated " + "to minimize the amount of padding.", + ) + group.add_argument( + "--duration-factor", + type=float, + default=1.0, + help="Determines the maximum duration of a concatenated cut " + "relative to the duration of the longest cut in a batch.", + ) + group.add_argument( + "--gap", + type=float, + default=1.0, + help="The amount of padding (in seconds) inserted between " + "concatenated cuts. This padding is filled with noise when " + "noise augmentation is used.", + ) + group.add_argument( + "--max-duration", + type=int, + default=100.0, + help="Maximum pooled recordings duration (seconds) in a " + "single batch. You can reduce it if it causes CUDA OOM.", + ) + group.add_argument( + "--num-buckets", + type=int, + default=30, + help="The number of buckets for the BucketingSampler" + "(you might want to increase it for larger datasets).", + ) + group.add_argument( + "--on-the-fly-feats", + type=str2bool, + default=False, + help="When enabled, use on-the-fly cut mixing and feature " + "extraction. Will drop existing precomputed feature manifests " + "if available.", + ) + group.add_argument( + "--shuffle", + type=str2bool, + default=True, + help="When enabled (=default), the examples will be " + "shuffled for each epoch.", + ) + + group.add_argument( + "--num-workers", + type=int, + default=8, + help="The number of training dataloader workers that " + "collect the batches.", + ) + group.add_argument( + "--enable-spec-aug", + type=str2bool, + default=True, + help="When enabled, use SpecAugment for training dataset.", + ) + group.add_argument( + "--spec-aug-time-warp-factor", + type=int, + default=80, + help="Used only when --enable-spec-aug is True. " + "It specifies the factor for time warping in SpecAugment. " + "Larger values mean more warping. " + "A value less than 1 means to disable time warp.", + ) + + def train_dataloaders( + self, + cuts_train: CutSet, + sampler_state_dict: Optional[Dict[str, Any]] = None, + ) -> DataLoader: + """ + Args: + cuts_train: + CutSet for training. + sampler_state_dict: + The state dict for the training sampler. + """ + logging.info("About to get Musan cuts") + cuts_musan = load_manifest( + self.args.manifest_dir / "cuts_musan.jsonl.gz" + ) + + transforms = [] + if self.args.enable_musan: + logging.info("Enable MUSAN") + transforms.append( + CutMix( + cuts=cuts_musan, prob=0.5, snr=(10, 20), preserve_id=True + ) + ) + else: + logging.info("Disable MUSAN") + + if self.args.concatenate_cuts: + logging.info( + f"Using cut concatenation with duration factor " + f"{self.args.duration_factor} and gap {self.args.gap}." + ) + # Cut concatenation should be the first transform in the list, + # so that if we e.g. mix noise in, it will fill the gaps between + # different utterances. + transforms = [ + CutConcatenate( + duration_factor=self.args.duration_factor, gap=self.args.gap + ) + ] + transforms + + input_transforms = [] + if self.args.enable_spec_aug: + logging.info("Enable SpecAugment") + logging.info( + f"Time warp factor: {self.args.spec_aug_time_warp_factor}" + ) + input_transforms.append( + SpecAugment( + time_warp_factor=self.args.spec_aug_time_warp_factor, + num_frame_masks=2, + features_mask_size=27, + num_feature_masks=2, + frames_mask_size=100, + ) + ) + else: + logging.info("Disable SpecAugment") + + logging.info("About to create train dataset") + if self.args.on_the_fly_feats: + train = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_strategy=OnTheFlyFeatures( + Fbank(FbankConfig(num_mel_bins=80)) + ), + input_transforms=input_transforms, + ) + else: + train = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_transforms=input_transforms, + ) + + logging.info("Using DynamicBucketingSampler.") + train_sampler = DynamicBucketingSampler( + cuts_train, + max_duration=self.args.max_duration, + shuffle=False, + num_buckets=self.args.num_buckets, + drop_last=True, + ) + logging.info("About to create train dataloader") + + if sampler_state_dict is not None: + logging.info("Loading sampler state dict") + train_sampler.load_state_dict(sampler_state_dict) + + # 'seed' is derived from the current random state, which will have + # previously been set in the main process. + seed = torch.randint(0, 100000, ()).item() + worker_init_fn = _SeedWorkers(seed) + + train_dl = DataLoader( + train, + sampler=train_sampler, + batch_size=None, + num_workers=self.args.num_workers, + persistent_workers=False, + worker_init_fn=worker_init_fn, + ) + + return train_dl + + def valid_dataloaders(self, cuts_valid: CutSet) -> DataLoader: + + transforms = [] + if self.args.concatenate_cuts: + transforms = [ + CutConcatenate( + duration_factor=self.args.duration_factor, gap=self.args.gap + ) + ] + transforms + + logging.info("About to create dev dataset") + if self.args.on_the_fly_feats: + validate = K2SpeechRecognitionDataset( + cut_transforms=transforms, + input_strategy=OnTheFlyFeatures( + Fbank(FbankConfig(num_mel_bins=80)) + ), + ) + else: + validate = K2SpeechRecognitionDataset( + cut_transforms=transforms, + ) + valid_sampler = DynamicBucketingSampler( + cuts_valid, + max_duration=self.args.max_duration, + shuffle=False, + ) + logging.info("About to create dev dataloader") + valid_dl = DataLoader( + validate, + sampler=valid_sampler, + batch_size=None, + num_workers=2, + persistent_workers=False, + ) + + return valid_dl + + def test_dataloaders(self, cuts: CutSet) -> DataLoader: + logging.debug("About to create test dataset") + test = K2SpeechRecognitionDataset( + input_strategy=OnTheFlyFeatures(Fbank(FbankConfig(num_mel_bins=80))) + if self.args.on_the_fly_feats + else PrecomputedFeatures(), + ) + sampler = DynamicBucketingSampler( + cuts, max_duration=self.args.max_duration, shuffle=False + ) + logging.debug("About to create test dataloader") + test_dl = DataLoader( + test, + batch_size=None, + sampler=sampler, + num_workers=self.args.num_workers, + ) + return test_dl + + @lru_cache() + def train_cuts(self) -> CutSet: + logging.info("About to get SPGISpeech train cuts") + return load_manifest_lazy( + self.args.manifest_dir / "cuts_train_shuf.jsonl.gz" + ) + + @lru_cache() + def dev_cuts(self) -> CutSet: + logging.info("About to get SPGISpeech dev cuts") + return load_manifest_lazy(self.args.manifest_dir / "cuts_dev.jsonl.gz") + + @lru_cache() + def val_cuts(self) -> CutSet: + logging.info("About to get SPGISpeech val cuts") + return load_manifest_lazy(self.args.manifest_dir / "cuts_val.jsonl.gz") + + +def test(): + parser = argparse.ArgumentParser() + SPGISpeechAsrDataModule.add_arguments(parser) + args = parser.parse_args() + adm = SPGISpeechAsrDataModule(args) + + cuts = adm.train_cuts() + dl = adm.train_dataloaders(cuts) + for i, batch in tqdm(enumerate(dl)): + if i == 100: + break + + cuts = adm.dev_cuts() + dl = adm.valid_dataloaders(cuts) + for i, batch in tqdm(enumerate(dl)): + if i == 100: + break + + +if __name__ == "__main__": + test() diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/beam_search.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/beam_search.py new file mode 120000 index 000000000..e24eca39f --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/beam_search.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/beam_search.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/conformer.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/conformer.py new file mode 120000 index 000000000..a65957180 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/conformer.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/conformer.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/decode.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/decode.py new file mode 100755 index 000000000..ae49d166b --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/decode.py @@ -0,0 +1,594 @@ +#!/usr/bin/env python3 +# +# Copyright 2021 Xiaomi Corporation (Author: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +""" +Usage: +(1) greedy search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 \ + --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method greedy_search + +(2) beam search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 \ + --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method beam_search \ + --beam-size 4 + +(3) modified beam search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 \ + --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 100 \ + --decoding-method modified_beam_search \ + --beam-size 4 + +(4) fast beam search +./pruned_transducer_stateless2/decode.py \ + --iter 696000 \ + --avg 10 \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --max-duration 1500 \ + --decoding-method fast_beam_search \ + --beam 4 \ + --max-contexts 4 \ + --max-states 8 +""" + + +import argparse +import logging +from collections import defaultdict +from pathlib import Path +from typing import Dict, List, Optional, Tuple + +import k2 +import sentencepiece as spm +import torch +import torch.nn as nn +from asr_datamodule import SPGISpeechAsrDataModule +from beam_search import ( + beam_search, + fast_beam_search_one_best, + greedy_search, + greedy_search_batch, + modified_beam_search, +) +from train import get_params, get_transducer_model + +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) +from icefall.utils import ( + AttributeDict, + setup_logger, + store_transcripts, + write_error_stats, +) + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--epoch", + type=int, + default=20, + help="""It specifies the checkpoint to use for decoding. + Note: Epoch counts from 0. + You can specify --avg to use more checkpoints for model averaging.""", + ) + + parser.add_argument( + "--iter", + type=int, + default=0, + help="""If positive, --epoch is ignored and it + will use the checkpoint exp_dir/checkpoint-iter.pt. + You can specify --avg to use more checkpoints for model averaging. + """, + ) + + parser.add_argument( + "--avg", + type=int, + default=10, + help="Number of checkpoints to average. Automatically select " + "consecutive checkpoints before the checkpoint specified by " + "'--epoch' and '--iter'", + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="The experiment dir", + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--decoding-method", + type=str, + default="greedy_search", + help="""Possible values are: + - greedy_search + - beam_search + - modified_beam_search + - fast_beam_search + """, + ) + + parser.add_argument( + "--beam-size", + type=int, + default=4, + help="""An interger indicating how many candidates we will keep for each + frame. Used only when --decoding-method is beam_search or + modified_beam_search.""", + ) + + parser.add_argument( + "--beam", + type=float, + default=4, + help="""A floating point value to calculate the cutoff score during beam + search (i.e., `cutoff = max-score - beam`), which is the same as the + `beam` in Kaldi. + Used only when --decoding-method is fast_beam_search""", + ) + + parser.add_argument( + "--max-contexts", + type=int, + default=4, + help="""Used only when --decoding-method is + fast_beam_search""", + ) + + parser.add_argument( + "--max-states", + type=int, + default=8, + help="""Used only when --decoding-method is + fast_beam_search""", + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + parser.add_argument( + "--max-sym-per-frame", + type=int, + default=1, + help="""Maximum number of symbols per frame. + Used only when --decoding_method is greedy_search""", + ) + + return parser + + +def decode_one_batch( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + batch: dict, + decoding_graph: Optional[k2.Fsa] = None, +) -> Dict[str, List[List[str]]]: + """Decode one batch and return the result in a dict. The dict has the + following format: + + - key: It indicates the setting used for decoding. For example, + if greedy_search is used, it would be "greedy_search" + If beam search with a beam size of 7 is used, it would be + "beam_7" + - value: It contains the decoding result. `len(value)` equals to + batch size. `value[i]` is the decoding result for the i-th + utterance in the given batch. + Args: + params: + It's the return value of :func:`get_params`. + model: + The neural model. + sp: + The BPE model. + batch: + It is the return value from iterating + `lhotse.dataset.K2SpeechRecognitionDataset`. See its documentation + for the format of the `batch`. + decoding_graph: + The decoding graph. Can be either a `k2.trivial_graph` or HLG, Used + only when --decoding_method is fast_beam_search. + Returns: + Return the decoding result. See above description for the format of + the returned dict. + """ + device = model.device + feature = batch["inputs"] + assert feature.ndim == 3 + + feature = feature.to(device) + # at entry, feature is (N, T, C) + + supervisions = batch["supervisions"] + feature_lens = supervisions["num_frames"].to(device) + + encoder_out, encoder_out_lens = model.encoder( + x=feature, x_lens=feature_lens + ) + hyps = [] + + if params.decoding_method == "fast_beam_search": + hyp_tokens = fast_beam_search_one_best( + model=model, + decoding_graph=decoding_graph, + encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, + beam=params.beam, + max_contexts=params.max_contexts, + max_states=params.max_states, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + elif ( + params.decoding_method == "greedy_search" + and params.max_sym_per_frame == 1 + ): + hyp_tokens = greedy_search_batch( + model=model, + encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + elif params.decoding_method == "modified_beam_search": + hyp_tokens = modified_beam_search( + model=model, + encoder_out=encoder_out, + encoder_out_lens=encoder_out_lens, + beam=params.beam_size, + ) + for hyp in sp.decode(hyp_tokens): + hyps.append(hyp.split()) + else: + batch_size = encoder_out.size(0) + + for i in range(batch_size): + # fmt: off + encoder_out_i = encoder_out[i:i+1, :encoder_out_lens[i]] + # fmt: on + if params.decoding_method == "greedy_search": + hyp = greedy_search( + model=model, + encoder_out=encoder_out_i, + max_sym_per_frame=params.max_sym_per_frame, + ) + elif params.decoding_method == "beam_search": + hyp = beam_search( + model=model, + encoder_out=encoder_out_i, + beam=params.beam_size, + ) + else: + raise ValueError( + f"Unsupported decoding method: {params.decoding_method}" + ) + hyps.append(sp.decode(hyp).split()) + + if params.decoding_method == "greedy_search": + return {"greedy_search": hyps} + elif params.decoding_method == "fast_beam_search": + return { + ( + f"beam_{params.beam}_" + f"max_contexts_{params.max_contexts}_" + f"max_states_{params.max_states}" + ): hyps + } + else: + return {f"beam_size_{params.beam_size}": hyps} + + +def decode_dataset( + dl: torch.utils.data.DataLoader, + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + decoding_graph: Optional[k2.Fsa] = None, +) -> Dict[str, List[Tuple[List[str], List[str]]]]: + """Decode dataset. + + Args: + dl: + PyTorch's dataloader containing the dataset to decode. + params: + It is returned by :func:`get_params`. + model: + The neural model. + sp: + The BPE model. + decoding_graph: + The decoding graph. Can be either a `k2.trivial_graph` or HLG, Used + only when --decoding_method is fast_beam_search. + Returns: + Return a dict, whose key may be "greedy_search" if greedy search + is used, or it may be "beam_7" if beam size of 7 is used. + Its value is a list of tuples. Each tuple contains two elements: + The first is the reference transcript, and the second is the + predicted result. + """ + num_cuts = 0 + + try: + num_batches = len(dl) + except TypeError: + num_batches = "?" + + if params.decoding_method == "greedy_search": + log_interval = 100 + else: + log_interval = 2 + + results = defaultdict(list) + for batch_idx, batch in enumerate(dl): + texts = batch["supervisions"]["text"] + + hyps_dict = decode_one_batch( + params=params, + model=model, + sp=sp, + decoding_graph=decoding_graph, + batch=batch, + ) + + for name, hyps in hyps_dict.items(): + this_batch = [] + assert len(hyps) == len(texts) + for hyp_words, ref_text in zip(hyps, texts): + ref_words = ref_text.split() + this_batch.append((ref_words, hyp_words)) + + results[name].extend(this_batch) + + num_cuts += len(texts) + + if batch_idx % log_interval == 0: + batch_str = f"{batch_idx}/{num_batches}" + + logging.info( + f"batch {batch_str}, cuts processed until now is {num_cuts}" + ) + return results + + +def save_results( + params: AttributeDict, + test_set_name: str, + results_dict: Dict[str, List[Tuple[List[int], List[int]]]], +): + test_set_wers = dict() + test_set_cers = dict() + for key, results in results_dict.items(): + recog_path = ( + params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt" + ) + store_transcripts(filename=recog_path, texts=results) + logging.info(f"The transcripts are stored in {recog_path}") + + # The following prints out WERs, per-word error statistics and aligned + # ref/hyp pairs. + wers_filename = ( + params.res_dir / f"wers-{test_set_name}-{key}-{params.suffix}.txt" + ) + with open(wers_filename, "w") as f: + wer = write_error_stats( + f, f"{test_set_name}-{key}", results, enable_log=True + ) + test_set_wers[key] = wer + + # we also compute CER for spgispeech dataset. + results_char = [] + for res in results: + results_char.append((list("".join(res[0])), list("".join(res[1])))) + cers_filename = ( + params.res_dir / f"cers-{test_set_name}-{key}-{params.suffix}.txt" + ) + with open(cers_filename, "w") as f: + cer = write_error_stats( + f, f"{test_set_name}-{key}", results_char, enable_log=True + ) + test_set_cers[key] = cer + + logging.info("Wrote detailed error stats to {}".format(wers_filename)) + + test_set_wers = { + k: v for k, v in sorted(test_set_wers.items(), key=lambda x: x[1]) + } + test_set_cers = { + k: v for k, v in sorted(test_set_cers.items(), key=lambda x: x[1]) + } + errs_info = ( + params.res_dir + / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt" + ) + with open(errs_info, "w") as f: + print("settings\tWER\tCER", file=f) + for key in test_set_wers: + print( + "{}\t{}\t{}".format( + key, test_set_wers[key], test_set_cers[key] + ), + file=f, + ) + + s = "\nFor {}, WER/CER of different settings are:\n".format(test_set_name) + note = "\tbest for {}".format(test_set_name) + for key in test_set_wers: + s += "{}\t{}\t{}{}\n".format( + key, test_set_wers[key], test_set_cers[key], note + ) + note = "" + logging.info(s) + + +@torch.no_grad() +def main(): + parser = get_parser() + SPGISpeechAsrDataModule.add_arguments(parser) + args = parser.parse_args() + args.exp_dir = Path(args.exp_dir) + + params = get_params() + params.update(vars(args)) + + assert params.decoding_method in ( + "greedy_search", + "beam_search", + "fast_beam_search", + "modified_beam_search", + ) + params.res_dir = params.exp_dir / params.decoding_method + + if params.iter > 0: + params.suffix = f"iter-{params.iter}-avg-{params.avg}" + else: + params.suffix = f"epoch-{params.epoch}-avg-{params.avg}" + + if "fast_beam_search" in params.decoding_method: + params.suffix += f"-beam-{params.beam}" + params.suffix += f"-max-contexts-{params.max_contexts}" + params.suffix += f"-max-states-{params.max_states}" + elif "beam_search" in params.decoding_method: + params.suffix += ( + f"-{params.decoding_method}-beam-size-{params.beam_size}" + ) + else: + params.suffix += f"-context-{params.context_size}" + params.suffix += f"-max-sym-per-frame-{params.max_sym_per_frame}" + + setup_logger(f"{params.res_dir}/log-decode-{params.suffix}") + logging.info("Decoding started") + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", 0) + + logging.info(f"Device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + if params.iter > 0: + filenames = find_checkpoints(params.exp_dir, iteration=-params.iter)[ + : params.avg + ] + if len(filenames) == 0: + raise ValueError( + f"No checkpoints found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + elif len(filenames) < params.avg: + raise ValueError( + f"Not enough checkpoints ({len(filenames)}) found for" + f" --iter {params.iter}, --avg {params.avg}" + ) + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: + load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) + else: + start = params.epoch - params.avg + 1 + filenames = [] + for i in range(start, params.epoch + 1): + if start >= 0: + filenames.append(f"{params.exp_dir}/epoch-{i}.pt") + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + + model.to(device) + model.eval() + model.device = device + + if params.decoding_method == "fast_beam_search": + decoding_graph = k2.trivial_graph(params.vocab_size - 1, device=device) + else: + decoding_graph = None + + num_param = sum([p.numel() for p in model.parameters()]) + logging.info(f"Number of model parameters: {num_param}") + + spgispeech = SPGISpeechAsrDataModule(args) + + dev_cuts = spgispeech.dev_cuts() + val_cuts = spgispeech.val_cuts() + + dev_dl = spgispeech.test_dataloaders(dev_cuts) + val_dl = spgispeech.test_dataloaders(val_cuts) + + test_sets = ["dev", "val"] + test_dl = [dev_dl, val_dl] + + for test_set, test_dl in zip(test_sets, test_dl): + results_dict = decode_dataset( + dl=test_dl, + params=params, + model=model, + sp=sp, + decoding_graph=decoding_graph, + ) + + save_results( + params=params, + test_set_name=test_set, + results_dict=results_dict, + ) + + logging.info("Done!") + + +if __name__ == "__main__": + main() diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/decoder.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/decoder.py new file mode 120000 index 000000000..722e1c894 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/decoder.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/decoder.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/encoder_interface.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/encoder_interface.py new file mode 120000 index 000000000..f58253127 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/encoder_interface.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/encoder_interface.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/export.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/export.py new file mode 100755 index 000000000..6119ecf2c --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/export.py @@ -0,0 +1,201 @@ +#!/usr/bin/env python3 +# +# Copyright 2021 Xiaomi Corporation (Author: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +# This script converts several saved checkpoints +# to a single one using model averaging. +""" +Usage: +./pruned_transducer_stateless2/export.py \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --avg-last-n 10 + +It will generate a file exp_dir/pretrained.pt + +To use the generated file with `pruned_transducer_stateless2/decode.py`, +you can do: + + cd /path/to/exp_dir + ln -s pretrained.pt epoch-9999.pt + + cd /path/to/egs/spgispeech/ASR + ./pruned_transducer_stateless2/decode.py \ + --exp-dir ./pruned_transducer_stateless2/exp \ + --epoch 9999 \ + --avg 1 \ + --max-duration 100 \ + --bpe-model data/lang_bpe_500/bpe.model +""" + +import argparse +import logging +from pathlib import Path + +import sentencepiece as spm +import torch +from train import get_params, get_transducer_model + +from icefall.checkpoint import ( + average_checkpoints, + find_checkpoints, + load_checkpoint, +) +from icefall.utils import str2bool + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--epoch", + type=int, + default=28, + help="It specifies the checkpoint to use for decoding." + "Note: Epoch counts from 0.", + ) + + parser.add_argument( + "--avg", + type=int, + default=15, + help="Number of checkpoints to average. Automatically select " + "consecutive checkpoints before the checkpoint specified by " + "'--epoch'. ", + ) + + parser.add_argument( + "--avg-last-n", + type=int, + default=0, + help="""If positive, --epoch and --avg are ignored and it + will use the last n checkpoints exp_dir/checkpoint-xxx.pt + where xxx is the number of processed batches while + saving that checkpoint. + """, + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="""It specifies the directory where all training related + files, e.g., checkpoints, log, etc, are saved + """, + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--jit", + type=str2bool, + default=False, + help="""True to save a model after applying torch.jit.script. + """, + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + + return parser + + +def main(): + args = get_parser().parse_args() + args.exp_dir = Path(args.exp_dir) + + assert args.jit is False, "Support torchscript will be added later" + + params = get_params() + params.update(vars(args)) + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", 0) + + logging.info(f"device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + model.to(device) + + if params.avg_last_n > 0: + filenames = find_checkpoints(params.exp_dir)[: params.avg_last_n] + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + elif params.avg == 1: + load_checkpoint(f"{params.exp_dir}/epoch-{params.epoch}.pt", model) + else: + start = params.epoch - params.avg + 1 + filenames = [] + for i in range(start, params.epoch + 1): + if start >= 0: + filenames.append(f"{params.exp_dir}/epoch-{i}.pt") + logging.info(f"averaging {filenames}") + model.to(device) + model.load_state_dict(average_checkpoints(filenames, device=device)) + + model.eval() + + model.to("cpu") + model.eval() + + if params.jit: + logging.info("Using torch.jit.script") + model = torch.jit.script(model) + filename = params.exp_dir / "cpu_jit.pt" + model.save(str(filename)) + logging.info(f"Saved to {filename}") + else: + logging.info("Not using torch.jit.script") + # Save it using a format so that it can be loaded + # by :func:`load_checkpoint` + filename = params.exp_dir / "pretrained.pt" + torch.save({"model": model.state_dict()}, str(filename)) + logging.info(f"Saved to {filename}") + + +if __name__ == "__main__": + formatter = ( + "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + ) + + logging.basicConfig(format=formatter, level=logging.INFO) + main() diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/joiner.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/joiner.py new file mode 120000 index 000000000..9052f3cbb --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/joiner.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/joiner.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/model.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/model.py new file mode 120000 index 000000000..a99e74334 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/model.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/model.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/optim.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/optim.py new file mode 120000 index 000000000..0a2f285aa --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/optim.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/optim.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/scaling.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/scaling.py new file mode 120000 index 000000000..c10cdfe12 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/scaling.py @@ -0,0 +1 @@ +../../../librispeech/ASR/pruned_transducer_stateless2/scaling.py \ No newline at end of file diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py new file mode 100755 index 000000000..6c66bfb62 --- /dev/null +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py @@ -0,0 +1,1031 @@ +#!/usr/bin/env python3 +# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang, +# Wei Kang +# Mingshuang Luo) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +""" +Usage: + +export CUDA_VISIBLE_DEVICES="0,1,2,3" + +./pruned_transducer_stateless2/train.py \ + --world-size 4 \ + --num-epochs 30 \ + --start-epoch 0 \ + --exp-dir pruned_transducer_stateless2/exp \ + --full-libri 1 \ + --max-duration 300 + +# For mix precision training: + +./pruned_transducer_stateless2/train.py \ + --world-size 4 \ + --num-epochs 30 \ + --start-epoch 0 \ + --use_fp16 1 \ + --exp-dir pruned_transducer_stateless2/exp \ + --full-libri 1 \ + --max-duration 550 + +""" + + +import argparse +import logging +import warnings +from pathlib import Path +from shutil import copyfile +from typing import Any, Dict, Optional, Tuple, Union + +import k2 +import optim +import sentencepiece as spm +import torch +import torch.multiprocessing as mp +import torch.nn as nn +from asr_datamodule import SPGISpeechAsrDataModule +from conformer import Conformer +from decoder import Decoder +from joiner import Joiner +from lhotse.dataset.sampling.base import CutSampler +from lhotse.utils import fix_random_seed +from model import Transducer +from optim import Eden, Eve +from torch import Tensor +from torch.cuda.amp import GradScaler +from torch.nn.parallel import DistributedDataParallel as DDP +from torch.utils.tensorboard import SummaryWriter + +from icefall import diagnostics +from icefall.checkpoint import load_checkpoint, remove_checkpoints +from icefall.checkpoint import save_checkpoint as save_checkpoint_impl +from icefall.checkpoint import save_checkpoint_with_global_batch_idx +from icefall.dist import cleanup_dist, setup_dist +from icefall.env import get_env_info +from icefall.utils import AttributeDict, MetricsTracker, setup_logger, str2bool + +LRSchedulerType = Union[ + torch.optim.lr_scheduler._LRScheduler, optim.LRScheduler +] + + +def get_parser(): + parser = argparse.ArgumentParser( + formatter_class=argparse.ArgumentDefaultsHelpFormatter + ) + + parser.add_argument( + "--world-size", + type=int, + default=1, + help="Number of GPUs for DDP training.", + ) + + parser.add_argument( + "--master-port", + type=int, + default=12354, + help="Master port to use for DDP training.", + ) + + parser.add_argument( + "--tensorboard", + type=str2bool, + default=True, + help="Should various information be logged in tensorboard.", + ) + + parser.add_argument( + "--num-epochs", + type=int, + default=20, + help="Number of epochs to train.", + ) + + parser.add_argument( + "--start-epoch", + type=int, + default=0, + help="""Resume training from from this epoch. + If it is positive, it will load checkpoint from + transducer_stateless2/exp/epoch-{start_epoch-1}.pt + """, + ) + + parser.add_argument( + "--start-batch", + type=int, + default=0, + help="""If positive, --start-epoch is ignored and + it loads the checkpoint from exp-dir/checkpoint-{start_batch}.pt + """, + ) + + parser.add_argument( + "--exp-dir", + type=str, + default="pruned_transducer_stateless2/exp", + help="""The experiment dir. + It specifies the directory where all training related + files, e.g., checkpoints, log, etc, are saved + """, + ) + + parser.add_argument( + "--bpe-model", + type=str, + default="data/lang_bpe_500/bpe.model", + help="Path to the BPE model", + ) + + parser.add_argument( + "--initial-lr", + type=float, + default=0.003, + help="The initial learning rate. This value should not need to be " + "changed.", + ) + + parser.add_argument( + "--lr-batches", + type=float, + default=5000, + help="""Number of steps that affects how rapidly the learning rate decreases. + We suggest not to change this.""", + ) + + parser.add_argument( + "--lr-epochs", + type=float, + default=4, + help="""Number of epochs that affects how rapidly the learning rate decreases. + """, + ) + + parser.add_argument( + "--context-size", + type=int, + default=2, + help="The context size in the decoder. 1 means bigram; " + "2 means tri-gram", + ) + + parser.add_argument( + "--prune-range", + type=int, + default=5, + help="The prune range for rnnt loss, it means how many symbols(context)" + "we are using to compute the loss", + ) + + parser.add_argument( + "--lm-scale", + type=float, + default=0.25, + help="The scale to smooth the loss with lm " + "(output of prediction network) part.", + ) + + parser.add_argument( + "--am-scale", + type=float, + default=0.0, + help="The scale to smooth the loss with am (output of encoder network)" + "part.", + ) + + parser.add_argument( + "--simple-loss-scale", + type=float, + default=0.5, + help="To get pruning ranges, we will calculate a simple version" + "loss(joiner is just addition), this simple loss also uses for" + "training (as a regularization item). We will scale the simple loss" + "with this parameter before adding to the final loss.", + ) + + parser.add_argument( + "--seed", + type=int, + default=42, + help="The seed for random generators intended for reproducibility", + ) + + parser.add_argument( + "--print-diagnostics", + type=str2bool, + default=False, + help="Accumulate stats on activations, print them and exit.", + ) + + parser.add_argument( + "--save-every-n", + type=int, + default=8000, + help="""Save checkpoint after processing this number of batches" + periodically. We save checkpoint to exp-dir/ whenever + params.batch_idx_train % save_every_n == 0. The checkpoint filename + has the form: f'exp-dir/checkpoint-{params.batch_idx_train}.pt' + Note: It also saves checkpoint to `exp-dir/epoch-xxx.pt` at the + end of each epoch where `xxx` is the epoch number counting from 0. + """, + ) + + parser.add_argument( + "--keep-last-k", + type=int, + default=10, + help="""Only keep this number of checkpoints on disk. + For instance, if it is 3, there are only 3 checkpoints + in the exp-dir with filenames `checkpoint-xxx.pt`. + It does not affect checkpoints with name `epoch-xxx.pt`. + """, + ) + + parser.add_argument( + "--use-fp16", + type=str2bool, + default=False, + help="Whether to use half precision training.", + ) + + return parser + + +def get_params() -> AttributeDict: + """Return a dict containing training parameters. + + All training related parameters that are not passed from the commandline + are saved in the variable `params`. + + Commandline options are merged into `params` after they are parsed, so + you can also access them via `params`. + + Explanation of options saved in `params`: + + - best_train_loss: Best training loss so far. It is used to select + the model that has the lowest training loss. It is + updated during the training. + + - best_valid_loss: Best validation loss so far. It is used to select + the model that has the lowest validation loss. It is + updated during the training. + + - best_train_epoch: It is the epoch that has the best training loss. + + - best_valid_epoch: It is the epoch that has the best validation loss. + + - batch_idx_train: Used to writing statistics to tensorboard. It + contains number of batches trained so far across + epochs. + + - log_interval: Print training loss if batch_idx % log_interval` is 0 + + - reset_interval: Reset statistics if batch_idx % reset_interval is 0 + + - valid_interval: Run validation if batch_idx % valid_interval is 0 + + - feature_dim: The model input dim. It has to match the one used + in computing features. + + - subsampling_factor: The subsampling factor for the model. + + - encoder_dim: Hidden dim for multi-head attention model. + + - num_decoder_layers: Number of decoder layer of transformer decoder. + + - warm_step: The warm_step for Noam optimizer. + """ + params = AttributeDict( + { + "best_train_loss": float("inf"), + "best_valid_loss": float("inf"), + "best_train_epoch": -1, + "best_valid_epoch": -1, + "batch_idx_train": 0, + "log_interval": 50, + "reset_interval": 200, + "valid_interval": 3000, # For the 100h subset, use 800 + # parameters for conformer + "feature_dim": 80, + "subsampling_factor": 4, + "encoder_dim": 512, + "nhead": 8, + "dim_feedforward": 2048, + "num_encoder_layers": 12, + # parameters for decoder + "decoder_dim": 512, + # parameters for joiner + "joiner_dim": 512, + # parameters for Noam + "model_warm_step": 3000, # arg given to model, not for lrate + "env_info": get_env_info(), + } + ) + + return params + + +def get_encoder_model(params: AttributeDict) -> nn.Module: + # TODO: We can add an option to switch between Conformer and Transformer + encoder = Conformer( + num_features=params.feature_dim, + subsampling_factor=params.subsampling_factor, + d_model=params.encoder_dim, + nhead=params.nhead, + dim_feedforward=params.dim_feedforward, + num_encoder_layers=params.num_encoder_layers, + ) + return encoder + + +def get_decoder_model(params: AttributeDict) -> nn.Module: + decoder = Decoder( + vocab_size=params.vocab_size, + decoder_dim=params.decoder_dim, + blank_id=params.blank_id, + context_size=params.context_size, + ) + return decoder + + +def get_joiner_model(params: AttributeDict) -> nn.Module: + joiner = Joiner( + encoder_dim=params.encoder_dim, + decoder_dim=params.decoder_dim, + joiner_dim=params.joiner_dim, + vocab_size=params.vocab_size, + ) + return joiner + + +def get_transducer_model(params: AttributeDict) -> nn.Module: + encoder = get_encoder_model(params) + decoder = get_decoder_model(params) + joiner = get_joiner_model(params) + + model = Transducer( + encoder=encoder, + decoder=decoder, + joiner=joiner, + encoder_dim=params.encoder_dim, + decoder_dim=params.decoder_dim, + joiner_dim=params.joiner_dim, + vocab_size=params.vocab_size, + ) + return model + + +def load_checkpoint_if_available( + params: AttributeDict, + model: nn.Module, + optimizer: Optional[torch.optim.Optimizer] = None, + scheduler: Optional[LRSchedulerType] = None, +) -> Optional[Dict[str, Any]]: + """Load checkpoint from file. + + If params.start_batch is positive, it will load the checkpoint from + `params.exp_dir/checkpoint-{params.start_batch}.pt`. Otherwise, if + params.start_epoch is positive, it will load the checkpoint from + `params.start_epoch - 1`. + + Apart from loading state dict for `model` and `optimizer` it also updates + `best_train_epoch`, `best_train_loss`, `best_valid_epoch`, + and `best_valid_loss` in `params`. + + Args: + params: + The return value of :func:`get_params`. + model: + The training model. + optimizer: + The optimizer that we are using. + scheduler: + The scheduler that we are using. + Returns: + Return a dict containing previously saved training info. + """ + if params.start_batch > 0: + filename = params.exp_dir / f"checkpoint-{params.start_batch}.pt" + elif params.start_epoch > 0: + filename = params.exp_dir / f"epoch-{params.start_epoch-1}.pt" + else: + return None + + assert filename.is_file(), f"{filename} does not exist!" + + saved_params = load_checkpoint( + filename, + model=model, + optimizer=optimizer, + scheduler=scheduler, + ) + + keys = [ + "best_train_epoch", + "best_valid_epoch", + "batch_idx_train", + "best_train_loss", + "best_valid_loss", + ] + for k in keys: + params[k] = saved_params[k] + + if params.start_batch > 0: + if "cur_epoch" in saved_params: + params["start_epoch"] = saved_params["cur_epoch"] + + if "cur_batch_idx" in saved_params: + params["cur_batch_idx"] = saved_params["cur_batch_idx"] + + return saved_params + + +def save_checkpoint( + params: AttributeDict, + model: nn.Module, + optimizer: Optional[torch.optim.Optimizer] = None, + scheduler: Optional[LRSchedulerType] = None, + sampler: Optional[CutSampler] = None, + scaler: Optional[GradScaler] = None, + rank: int = 0, +) -> None: + """Save model, optimizer, scheduler and training stats to file. + + Args: + params: + It is returned by :func:`get_params`. + model: + The training model. + optimizer: + The optimizer used in the training. + sampler: + The sampler for the training dataset. + scaler: + The scaler used for mix precision training. + """ + if rank != 0: + return + filename = params.exp_dir / f"epoch-{params.cur_epoch}.pt" + save_checkpoint_impl( + filename=filename, + model=model, + params=params, + optimizer=optimizer, + scheduler=scheduler, + sampler=sampler, + scaler=scaler, + rank=rank, + ) + + if params.best_train_epoch == params.cur_epoch: + best_train_filename = params.exp_dir / "best-train-loss.pt" + copyfile(src=filename, dst=best_train_filename) + + if params.best_valid_epoch == params.cur_epoch: + best_valid_filename = params.exp_dir / "best-valid-loss.pt" + copyfile(src=filename, dst=best_valid_filename) + + +def compute_loss( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + batch: dict, + is_training: bool, + warmup: float = 1.0, +) -> Tuple[Tensor, MetricsTracker]: + """ + Compute CTC loss given the model and its inputs. + + Args: + params: + Parameters for training. See :func:`get_params`. + model: + The model for training. It is an instance of Conformer in our case. + batch: + A batch of data. See `lhotse.dataset.K2SpeechRecognitionDataset()` + for the content in it. + is_training: + True for training. False for validation. When it is True, this + function enables autograd during computation; when it is False, it + disables autograd. + warmup: a floating point value which increases throughout training; + values >= 1.0 are fully warmed up and have all modules present. + """ + device = model.device + feature = batch["inputs"] + # at entry, feature is (N, T, C) + assert feature.ndim == 3 + feature = feature.to(device) + + supervisions = batch["supervisions"] + feature_lens = supervisions["num_frames"].to(device) + + texts = batch["supervisions"]["text"] + y = sp.encode(texts, out_type=int) + y = k2.RaggedTensor(y).to(device) + + with torch.set_grad_enabled(is_training): + simple_loss, pruned_loss = model( + x=feature, + x_lens=feature_lens, + y=y, + prune_range=params.prune_range, + am_scale=params.am_scale, + lm_scale=params.lm_scale, + warmup=warmup, + ) + # after the main warmup step, we keep pruned_loss_scale small + # for the same amount of time (model_warm_step), to avoid + # overwhelming the simple_loss and causing it to diverge, + # in case it had not fully learned the alignment yet. + pruned_loss_scale = ( + 0.0 + if warmup < 1.0 + else (0.1 if warmup > 1.0 and warmup < 2.0 else 1.0) + ) + loss = ( + params.simple_loss_scale * simple_loss + + pruned_loss_scale * pruned_loss + ) + + assert loss.requires_grad == is_training + + info = MetricsTracker() + with warnings.catch_warnings(): + warnings.simplefilter("ignore") + info["frames"] = ( + (feature_lens // params.subsampling_factor).sum().item() + ) + + # Note: We use reduction=sum while computing the loss. + info["loss"] = loss.detach().cpu().item() + info["simple_loss"] = simple_loss.detach().cpu().item() + info["pruned_loss"] = pruned_loss.detach().cpu().item() + + return loss, info + + +def compute_validation_loss( + params: AttributeDict, + model: nn.Module, + sp: spm.SentencePieceProcessor, + valid_dl: torch.utils.data.DataLoader, + world_size: int = 1, +) -> MetricsTracker: + """Run the validation process.""" + model.eval() + + tot_loss = MetricsTracker() + + for batch_idx, batch in enumerate(valid_dl): + loss, loss_info = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=False, + ) + assert loss.requires_grad is False + tot_loss = tot_loss + loss_info + + if world_size > 1: + tot_loss.reduce(loss.device) + + loss_value = tot_loss["loss"] / tot_loss["frames"] + if loss_value < params.best_valid_loss: + params.best_valid_epoch = params.cur_epoch + params.best_valid_loss = loss_value + + return tot_loss + + +def train_one_epoch( + params: AttributeDict, + model: nn.Module, + optimizer: torch.optim.Optimizer, + scheduler: LRSchedulerType, + sp: spm.SentencePieceProcessor, + train_dl: torch.utils.data.DataLoader, + valid_dl: torch.utils.data.DataLoader, + scaler: GradScaler, + tb_writer: Optional[SummaryWriter] = None, + world_size: int = 1, + rank: int = 0, +) -> None: + """Train the model for one epoch. + + The training loss from the mean of all frames is saved in + `params.train_loss`. It runs the validation process every + `params.valid_interval` batches. + + Args: + params: + It is returned by :func:`get_params`. + model: + The model for training. + optimizer: + The optimizer we are using. + scheduler: + The learning rate scheduler, we call step() every step. + train_dl: + Dataloader for the training dataset. + valid_dl: + Dataloader for the validation dataset. + scaler: + The scaler used for mix precision training. + tb_writer: + Writer to write log messages to tensorboard. + world_size: + Number of nodes in DDP training. If it is 1, DDP is disabled. + rank: + The rank of the node in DDP training. If no DDP is used, it should + be set to 0. + """ + model.train() + + tot_loss = MetricsTracker() + + cur_batch_idx = params.get("cur_batch_idx", 0) + + for batch_idx, batch in enumerate(train_dl): + if batch_idx < cur_batch_idx: + continue + cur_batch_idx = batch_idx + + params.batch_idx_train += 1 + batch_size = len(batch["supervisions"]["text"]) + + with torch.cuda.amp.autocast(enabled=params.use_fp16): + loss, loss_info = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=True, + warmup=(params.batch_idx_train / params.model_warm_step), + ) + # summary stats + tot_loss = (tot_loss * (1 - 1 / params.reset_interval)) + loss_info + + # NOTE: We use reduction==sum and loss is computed over utterances + # in the batch and there is no normalization to it so far. + scaler.scale(loss).backward() + scheduler.step_batch(params.batch_idx_train) + scaler.step(optimizer) + scaler.update() + optimizer.zero_grad() + + if params.print_diagnostics and batch_idx == 5: + return + + if ( + params.batch_idx_train > 0 + and params.batch_idx_train % params.save_every_n == 0 + ): + params.cur_batch_idx = batch_idx + save_checkpoint_with_global_batch_idx( + out_dir=params.exp_dir, + global_batch_idx=params.batch_idx_train, + model=model, + params=params, + optimizer=optimizer, + scheduler=scheduler, + sampler=train_dl.sampler, + scaler=scaler, + rank=rank, + ) + del params.cur_batch_idx + remove_checkpoints( + out_dir=params.exp_dir, + topk=params.keep_last_k, + rank=rank, + ) + + if batch_idx % params.log_interval == 0: + cur_lr = scheduler.get_last_lr()[0] + logging.info( + f"Epoch {params.cur_epoch}, " + f"batch {batch_idx}, loss[{loss_info}], " + f"tot_loss[{tot_loss}], batch size: {batch_size}, " + f"lr: {cur_lr:.2e}" + ) + + if tb_writer is not None: + tb_writer.add_scalar( + "train/learning_rate", cur_lr, params.batch_idx_train + ) + + loss_info.write_summary( + tb_writer, "train/current_", params.batch_idx_train + ) + tot_loss.write_summary( + tb_writer, "train/tot_", params.batch_idx_train + ) + + if batch_idx > 0 and batch_idx % params.valid_interval == 0: + logging.info("Computing validation loss") + valid_info = compute_validation_loss( + params=params, + model=model, + sp=sp, + valid_dl=valid_dl, + world_size=world_size, + ) + model.train() + logging.info(f"Epoch {params.cur_epoch}, validation: {valid_info}") + if tb_writer is not None: + valid_info.write_summary( + tb_writer, "train/valid_", params.batch_idx_train + ) + + loss_value = tot_loss["loss"] / tot_loss["frames"] + params.train_loss = loss_value + if params.train_loss < params.best_train_loss: + params.best_train_epoch = params.cur_epoch + params.best_train_loss = params.train_loss + + +def run(rank, world_size, args): + """ + Args: + rank: + It is a value between 0 and `world_size-1`, which is + passed automatically by `mp.spawn()` in :func:`main`. + The node with rank 0 is responsible for saving checkpoint. + world_size: + Number of GPUs for DDP training. + args: + The return value of get_parser().parse_args() + """ + params = get_params() + params.update(vars(args)) + + fix_random_seed(params.seed) + if world_size > 1: + setup_dist(rank, world_size, params.master_port) + + setup_logger(f"{params.exp_dir}/log/log-train") + logging.info("Training started") + + if args.tensorboard and rank == 0: + tb_writer = SummaryWriter(log_dir=f"{params.exp_dir}/tensorboard") + else: + tb_writer = None + + device = torch.device("cpu") + if torch.cuda.is_available(): + device = torch.device("cuda", rank) + logging.info(f"Device: {device}") + + sp = spm.SentencePieceProcessor() + sp.load(params.bpe_model) + + # is defined in local/train_bpe_model.py + params.blank_id = sp.piece_to_id("") + params.vocab_size = sp.get_piece_size() + + logging.info(params) + + logging.info("About to create model") + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + logging.info(f"Number of model parameters: {num_param}") + + checkpoints = load_checkpoint_if_available(params=params, model=model) + + model.to(device) + if world_size > 1: + logging.info("Using DDP") + model = DDP(model, device_ids=[rank]) + model.device = device + + optimizer = Eve(model.parameters(), lr=params.initial_lr) + + scheduler = Eden(optimizer, params.lr_batches, params.lr_epochs) + + if checkpoints and "optimizer" in checkpoints: + logging.info("Loading optimizer state dict") + optimizer.load_state_dict(checkpoints["optimizer"]) + + if ( + checkpoints + and "scheduler" in checkpoints + and checkpoints["scheduler"] is not None + ): + logging.info("Loading scheduler state dict") + scheduler.load_state_dict(checkpoints["scheduler"]) + + if params.print_diagnostics: + opts = diagnostics.TensorDiagnosticOptions( + 2 ** 22 + ) # allow 4 megabytes per sub-module + diagnostic = diagnostics.attach_diagnostics(model, opts) + + spgispeech = SPGISpeechAsrDataModule(args) + + train_cuts = spgispeech.train_cuts() + + # Ideally we should filter utterances that are too long or too short, + # but SPGISpeech contains regular length utterances so we don't need to + # do that. Here are the statistics of the training data (obtained by + # `train_cuts.describe()`): + + # Cuts count: 5886320 + # Total duration (hours): 15070.1 + # Speech duration (hours): 15070.1 (100.0%) + # *** + # Duration statistics (seconds): + # mean 9.2 + # std 2.8 + # min 4.6 + # 25% 6.9 + # 50% 8.9 + # 75% 11.2 + # 99% 16.0 + # 99.5% 16.3 + # 99.9% 16.6 + # max 16.7 + + if params.start_batch > 0 and checkpoints and "sampler" in checkpoints: + # We only load the sampler's state dict when it loads a checkpoint + # saved in the middle of an epoch + sampler_state_dict = checkpoints["sampler"] + else: + sampler_state_dict = None + + train_dl = spgispeech.train_dataloaders( + train_cuts, sampler_state_dict=sampler_state_dict + ) + + valid_cuts = spgispeech.dev_cuts() + valid_dl = spgispeech.valid_dataloaders(valid_cuts) + + if not params.print_diagnostics: + scan_pessimistic_batches_for_oom( + model=model, + train_dl=train_dl, + optimizer=optimizer, + sp=sp, + params=params, + ) + + scaler = GradScaler(enabled=params.use_fp16) + if checkpoints and "grad_scaler" in checkpoints: + logging.info("Loading grad scaler state dict") + scaler.load_state_dict(checkpoints["grad_scaler"]) + + for epoch in range(params.start_epoch, params.num_epochs): + scheduler.step_epoch(epoch) + fix_random_seed(params.seed + epoch) + train_dl.sampler.set_epoch(epoch) + + if tb_writer is not None: + tb_writer.add_scalar("train/epoch", epoch, params.batch_idx_train) + + params.cur_epoch = epoch + + train_one_epoch( + params=params, + model=model, + optimizer=optimizer, + scheduler=scheduler, + sp=sp, + train_dl=train_dl, + valid_dl=valid_dl, + scaler=scaler, + tb_writer=tb_writer, + world_size=world_size, + rank=rank, + ) + + if params.print_diagnostics: + diagnostic.print_diagnostics() + break + + save_checkpoint( + params=params, + model=model, + optimizer=optimizer, + scheduler=scheduler, + sampler=train_dl.sampler, + scaler=scaler, + rank=rank, + ) + + logging.info("Done!") + + if world_size > 1: + torch.distributed.barrier() + cleanup_dist() + + +def display_and_save_batch( + batch: dict, + params: AttributeDict, + sp: spm.SentencePieceProcessor, +) -> None: + """Display the batch statistics and save the batch into disk. + Args: + batch: + A batch of data. See `lhotse.dataset.K2SpeechRecognitionDataset()` + for the content in it. + params: + Parameters for training. See :func:`get_params`. + sp: + The BPE model. + """ + from lhotse.utils import uuid4 + + filename = f"{params.exp_dir}/batch-{uuid4()}.pt" + logging.info(f"Saving batch to {filename}") + torch.save(batch, filename) + + supervisions = batch["supervisions"] + features = batch["inputs"] + + logging.info(f"features shape: {features.shape}") + + y = sp.encode(supervisions["text"], out_type=int) + num_tokens = sum(len(i) for i in y) + logging.info(f"num tokens: {num_tokens}") + + +def scan_pessimistic_batches_for_oom( + model: nn.Module, + train_dl: torch.utils.data.DataLoader, + optimizer: torch.optim.Optimizer, + sp: spm.SentencePieceProcessor, + params: AttributeDict, +): + from lhotse.dataset import find_pessimistic_batches + + logging.info( + "Sanity check -- see if any of the batches in epoch 0 would cause OOM." + ) + batches, crit_values = find_pessimistic_batches(train_dl.sampler) + for criterion, cuts in batches.items(): + batch = train_dl.dataset[cuts] + try: + # warmup = 0.0 is so that the derivs for the pruned loss stay zero + # (i.e. are not remembered by the decaying-average in adam), because + # we want to avoid these params being subject to shrinkage in adam. + with torch.cuda.amp.autocast(enabled=params.use_fp16): + loss, _ = compute_loss( + params=params, + model=model, + sp=sp, + batch=batch, + is_training=True, + warmup=0.0, + ) + loss.backward() + optimizer.step() + optimizer.zero_grad() + except RuntimeError as e: + if "CUDA out of memory" in str(e): + logging.error( + "Your GPU ran out of memory with the current " + "max_duration setting. We recommend decreasing " + "max_duration and trying again.\n" + f"Failing criterion: {criterion} " + f"(={crit_values[criterion]}) ..." + ) + display_and_save_batch(batch, params=params, sp=sp) + raise + + +def main(): + parser = get_parser() + SPGISpeechAsrDataModule.add_arguments(parser) + args = parser.parse_args() + args.exp_dir = Path(args.exp_dir) + + world_size = args.world_size + assert world_size >= 1 + if world_size > 1: + mp.spawn(run, args=(world_size, args), nprocs=world_size, join=True) + else: + run(rank=0, world_size=1, args=args) + + +torch.set_num_threads(1) +torch.set_num_interop_threads(1) + +if __name__ == "__main__": + main() diff --git a/egs/spgispeech/ASR/shared b/egs/spgispeech/ASR/shared new file mode 120000 index 000000000..4c5e91438 --- /dev/null +++ b/egs/spgispeech/ASR/shared @@ -0,0 +1 @@ +../../../icefall/shared/ \ No newline at end of file From f6ce135608968bfe736040345743dfbc2847cd8b Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Mon, 16 May 2022 21:46:59 +0800 Subject: [PATCH 10/16] Various fixes to support torch script. (#371) * Various fixes to support torch script. * Add tests to ensure that the model is torch scriptable. * Update tests. --- .github/workflows/test.yml | 34 ++++++++- .../ASR/pruned_transducer_stateless/export.py | 7 +- .../test_decoder.py | 2 + .../pruned_transducer_stateless/test_model.py | 50 ++++++++++++++ .../pruned_transducer_stateless2/conformer.py | 11 +-- .../pruned_transducer_stateless2/export.py | 7 +- .../pruned_transducer_stateless2/scaling.py | 27 ++++++-- .../test_model.py | 50 ++++++++++++++ .../pruned_transducer_stateless3/export.py | 7 +- .../test_model.py | 50 ++++++++++++++ .../test_scaling.py | 69 +++++++++++++++++++ .../test_model.py | 50 ++++++++++++++ .../ASR/transducer_lstm/encoder.py | 4 +- egs/librispeech/ASR/transducer_lstm/model.py | 3 +- .../ASR/transducer_stateless/conformer.py | 10 +-- .../ASR/transducer_stateless/export.py | 7 +- .../ASR/transducer_stateless/joiner.py | 6 +- .../ASR/transducer_stateless/test_model.py | 49 +++++++++++++ .../ASR/transducer_stateless2/export.py | 7 +- .../ASR/transducer_stateless2/joiner.py | 9 ++- .../ASR/transducer_stateless2/test_model.py | 49 +++++++++++++ .../export.py | 7 +- 22 files changed, 480 insertions(+), 35 deletions(-) create mode 100755 egs/librispeech/ASR/pruned_transducer_stateless/test_model.py create mode 100755 egs/librispeech/ASR/pruned_transducer_stateless2/test_model.py create mode 100755 egs/librispeech/ASR/pruned_transducer_stateless3/test_model.py create mode 100755 egs/librispeech/ASR/pruned_transducer_stateless3/test_scaling.py create mode 100755 egs/librispeech/ASR/pruned_transducer_stateless4/test_model.py create mode 100755 egs/librispeech/ASR/transducer_stateless/test_model.py create mode 100755 egs/librispeech/ASR/transducer_stateless2/test_model.py diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml index f2c63a3b8..fce14c460 100644 --- a/.github/workflows/test.yml +++ b/.github/workflows/test.yml @@ -103,11 +103,26 @@ jobs: cd egs/librispeech/ASR/conformer_ctc pytest -v -s + cd ../pruned_transducer_stateless + pytest -v -s + + cd ../pruned_transducer_stateless2 + pytest -v -s + + cd ../pruned_transducer_stateless3 + pytest -v -s + + cd ../pruned_transducer_stateless4 + pytest -v -s + + cd ../transducer_stateless + pytest -v -s + if [[ ${{ matrix.torchaudio }} == "0.10.0" ]]; then cd ../transducer pytest -v -s - cd ../transducer_stateless + cd ../transducer_stateless2 pytest -v -s cd ../transducer_lstm @@ -128,11 +143,26 @@ jobs: cd egs/librispeech/ASR/conformer_ctc pytest -v -s + cd ../pruned_transducer_stateless + pytest -v -s + + cd ../pruned_transducer_stateless2 + pytest -v -s + + cd ../pruned_transducer_stateless3 + pytest -v -s + + cd ../pruned_transducer_stateless4 + pytest -v -s + + cd ../transducer_stateless + pytest -v -s + if [[ ${{ matrix.torchaudio }} == "0.10.0" ]]; then cd ../transducer pytest -v -s - cd ../transducer_stateless + cd ../transducer_stateless2 pytest -v -s cd ../transducer_lstm diff --git a/egs/librispeech/ASR/pruned_transducer_stateless/export.py b/egs/librispeech/ASR/pruned_transducer_stateless/export.py index 7d2a07817..a4210831c 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless/export.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless/export.py @@ -116,8 +116,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -159,6 +157,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" diff --git a/egs/librispeech/ASR/pruned_transducer_stateless/test_decoder.py b/egs/librispeech/ASR/pruned_transducer_stateless/test_decoder.py index 937d55c2a..36712018d 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless/test_decoder.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless/test_decoder.py @@ -29,6 +29,7 @@ from decoder import Decoder def test_decoder(): vocab_size = 3 blank_id = 0 + unk_id = 2 embedding_dim = 128 context_size = 4 @@ -36,6 +37,7 @@ def test_decoder(): vocab_size=vocab_size, embedding_dim=embedding_dim, blank_id=blank_id, + unk_id=unk_id, context_size=context_size, ) N = 100 diff --git a/egs/librispeech/ASR/pruned_transducer_stateless/test_model.py b/egs/librispeech/ASR/pruned_transducer_stateless/test_model.py new file mode 100755 index 000000000..5c49025bd --- /dev/null +++ b/egs/librispeech/ASR/pruned_transducer_stateless/test_model.py @@ -0,0 +1,50 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./pruned_transducer_stateless/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + params.unk_id = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/conformer.py b/egs/librispeech/ASR/pruned_transducer_stateless2/conformer.py index 257936b59..840d847cb 100644 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/conformer.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/conformer.py @@ -112,10 +112,13 @@ class Conformer(EncoderInterface): x, pos_emb = self.encoder_pos(x) x = x.permute(1, 0, 2) # (N, T, C) -> (T, N, C) - with warnings.catch_warnings(): - warnings.simplefilter("ignore") - # Caution: We assume the subsampling factor is 4! - lengths = ((x_lens - 1) // 2 - 1) // 2 + # Caution: We assume the subsampling factor is 4! + + # lengths = ((x_lens - 1) // 2 - 1) // 2 # issue an warning + # + # Note: rounding_mode in torch.div() is available only in torch >= 1.8.0 + lengths = (((x_lens - 1) >> 1) - 1) >> 1 + assert x.size(0) == lengths.max().item() mask = make_pad_mask(lengths) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/export.py b/egs/librispeech/ASR/pruned_transducer_stateless2/export.py index 6b3a7a9ff..cff9c7377 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/export.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/export.py @@ -131,8 +131,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -191,6 +189,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/scaling.py b/egs/librispeech/ASR/pruned_transducer_stateless2/scaling.py index f89d2963e..5ee4bab98 100644 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/scaling.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/scaling.py @@ -212,7 +212,10 @@ class ScaledLinear(nn.Linear): return self.weight * self.weight_scale.exp() def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() + if self.bias is None or self.bias_scale is None: + return None + + return self.bias * self.bias_scale.exp() def forward(self, input: Tensor) -> Tensor: return torch.nn.functional.linear( @@ -255,7 +258,11 @@ class ScaledConv1d(nn.Conv1d): return self.weight * self.weight_scale.exp() def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() + bias = self.bias + bias_scale = self.bias_scale + if bias is None or bias_scale is None: + return None + return bias * bias_scale.exp() def forward(self, input: Tensor) -> Tensor: F = torch.nn.functional @@ -269,7 +276,7 @@ class ScaledConv1d(nn.Conv1d): self.get_weight(), self.get_bias(), self.stride, - _single(0), + (0,), self.dilation, self.groups, ) @@ -319,7 +326,12 @@ class ScaledConv2d(nn.Conv2d): return self.weight * self.weight_scale.exp() def get_bias(self): - return None if self.bias is None else self.bias * self.bias_scale.exp() + # see https://github.com/pytorch/pytorch/issues/24135 + bias = self.bias + bias_scale = self.bias_scale + if bias is None or bias_scale is None: + return None + return bias * bias_scale.exp() def _conv_forward(self, input, weight): F = torch.nn.functional @@ -333,7 +345,7 @@ class ScaledConv2d(nn.Conv2d): weight, self.get_bias(), self.stride, - _pair(0), + (0, 0), self.dilation, self.groups, ) @@ -398,6 +410,9 @@ class ActivationBalancer(torch.nn.Module): self.max_abs = max_abs def forward(self, x: Tensor) -> Tensor: + if torch.jit.is_scripting(): + return x + return ActivationBalancerFunction.apply( x, self.channel_dim, @@ -444,6 +459,8 @@ class DoubleSwish(torch.nn.Module): """Return double-swish activation function which is an approximation to Swish(Swish(x)), that we approximate closely with x * sigmoid(x-1). """ + if torch.jit.is_scripting(): + return x * torch.sigmoid(x - 1.0) return DoubleSwishFunction.apply(x) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/test_model.py b/egs/librispeech/ASR/pruned_transducer_stateless2/test_model.py new file mode 100755 index 000000000..9d5c6376d --- /dev/null +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/test_model.py @@ -0,0 +1,50 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./pruned_transducer_stateless2/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + params.unk_id = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/export.py b/egs/librispeech/ASR/pruned_transducer_stateless3/export.py index 0cdb0b957..e674fb360 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/export.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/export.py @@ -132,8 +132,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -192,6 +190,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/test_model.py b/egs/librispeech/ASR/pruned_transducer_stateless3/test_model.py new file mode 100755 index 000000000..9a060c5fb --- /dev/null +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/test_model.py @@ -0,0 +1,50 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./pruned_transducer_stateless3/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + params.unk_id = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/test_scaling.py b/egs/librispeech/ASR/pruned_transducer_stateless3/test_scaling.py new file mode 100755 index 000000000..e9dfe6d5e --- /dev/null +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/test_scaling.py @@ -0,0 +1,69 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./pruned_transducer_stateless3/test_scaling.py +""" + +import torch +from scaling import ActivationBalancer, ScaledConv1d, ScaledConv2d + + +def test_scaled_conv1d(): + for bias in [True, False]: + conv1d = ScaledConv1d( + 3, + 6, + kernel_size=1, + stride=1, + padding=0, + bias=bias, + ) + torch.jit.script(conv1d) + + +def test_scaled_conv2d(): + for bias in [True, False]: + conv2d = ScaledConv2d( + in_channels=1, + out_channels=3, + kernel_size=3, + padding=1, + bias=bias, + ) + torch.jit.script(conv2d) + + +def test_activation_balancer(): + act = ActivationBalancer( + channel_dim=1, max_abs=10.0, min_positive=0.05, max_positive=1.0 + ) + torch.jit.script(act) + + +def main(): + test_scaled_conv1d() + test_scaled_conv2d() + test_activation_balancer() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/pruned_transducer_stateless4/test_model.py b/egs/librispeech/ASR/pruned_transducer_stateless4/test_model.py new file mode 100755 index 000000000..b1832d0ec --- /dev/null +++ b/egs/librispeech/ASR/pruned_transducer_stateless4/test_model.py @@ -0,0 +1,50 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./pruned_transducer_stateless4/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + params.unk_id = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/transducer_lstm/encoder.py b/egs/librispeech/ASR/transducer_lstm/encoder.py index 860a84bb1..3dc992dd2 100644 --- a/egs/librispeech/ASR/transducer_lstm/encoder.py +++ b/egs/librispeech/ASR/transducer_lstm/encoder.py @@ -94,7 +94,7 @@ class LstmEncoder(EncoderInterface): ) if False: - # It is commented out as DPP complains that not all parameters are + # It is commented out as DDP complains that not all parameters are # used. Need more checks later for the reason. # # Caution: We assume the dataloader returns utterances with @@ -107,7 +107,7 @@ class LstmEncoder(EncoderInterface): ) packed_rnn_out, _ = self.rnn(packed_x) - rnn_out, _ = pad_packed_sequence(packed_x, batch_first=True) + rnn_out, _ = pad_packed_sequence(packed_rnn_out, batch_first=True) else: rnn_out, _ = self.rnn(x) diff --git a/egs/librispeech/ASR/transducer_lstm/model.py b/egs/librispeech/ASR/transducer_lstm/model.py index 31843b60e..e37558a98 100644 --- a/egs/librispeech/ASR/transducer_lstm/model.py +++ b/egs/librispeech/ASR/transducer_lstm/model.py @@ -97,8 +97,7 @@ class Transducer(nn.Module): y_lens = row_splits[1:] - row_splits[:-1] blank_id = self.decoder.blank_id - sos_id = self.decoder.sos_id - sos_y = add_sos(y, sos_id=sos_id) + sos_y = add_sos(y, sos_id=blank_id) sos_y_padded = sos_y.pad(mode="constant", padding_value=blank_id) sos_y_padded = sos_y_padded.to(torch.int64) diff --git a/egs/librispeech/ASR/transducer_stateless/conformer.py b/egs/librispeech/ASR/transducer_stateless/conformer.py index 488c82386..51f13b73f 100644 --- a/egs/librispeech/ASR/transducer_stateless/conformer.py +++ b/egs/librispeech/ASR/transducer_stateless/conformer.py @@ -109,10 +109,12 @@ class Conformer(Transformer): x, pos_emb = self.encoder_pos(x) x = x.permute(1, 0, 2) # (N, T, C) -> (T, N, C) - with warnings.catch_warnings(): - warnings.simplefilter("ignore") - # Caution: We assume the subsampling factor is 4! - lengths = ((x_lens - 1) // 2 - 1) // 2 + # Caution: We assume the subsampling factor is 4! + + # lengths = ((x_lens - 1) // 2 - 1) // 2 # issue an warning + # + # Note: rounding_mode in torch.div() is available only in torch >= 1.8.0 + lengths = (((x_lens - 1) >> 1) - 1) >> 1 assert x.size(0) == lengths.max().item() mask = make_pad_mask(lengths) diff --git a/egs/librispeech/ASR/transducer_stateless/export.py b/egs/librispeech/ASR/transducer_stateless/export.py index 5687260df..8bd0bdea1 100755 --- a/egs/librispeech/ASR/transducer_stateless/export.py +++ b/egs/librispeech/ASR/transducer_stateless/export.py @@ -183,8 +183,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -226,6 +224,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" diff --git a/egs/librispeech/ASR/transducer_stateless/joiner.py b/egs/librispeech/ASR/transducer_stateless/joiner.py index b0ba7fd83..93cccbd8c 100644 --- a/egs/librispeech/ASR/transducer_stateless/joiner.py +++ b/egs/librispeech/ASR/transducer_stateless/joiner.py @@ -14,6 +14,8 @@ # See the License for the specific language governing permissions and # limitations under the License. +from typing import List + import torch import torch.nn as nn @@ -55,8 +57,8 @@ class Joiner(nn.Module): N = encoder_out.size(0) - encoder_out_len = encoder_out_len.tolist() - decoder_out_len = decoder_out_len.tolist() + encoder_out_len: List[int] = encoder_out_len.tolist() + decoder_out_len: List[int] = decoder_out_len.tolist() encoder_out_list = [ encoder_out[i, : encoder_out_len[i], :] for i in range(N) diff --git a/egs/librispeech/ASR/transducer_stateless/test_model.py b/egs/librispeech/ASR/transducer_stateless/test_model.py new file mode 100755 index 000000000..9e500f477 --- /dev/null +++ b/egs/librispeech/ASR/transducer_stateless/test_model.py @@ -0,0 +1,49 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./transducer_stateless/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/transducer_stateless2/export.py b/egs/librispeech/ASR/transducer_stateless2/export.py index 7a68f69ff..57c1a6094 100755 --- a/egs/librispeech/ASR/transducer_stateless2/export.py +++ b/egs/librispeech/ASR/transducer_stateless2/export.py @@ -115,8 +115,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -158,6 +156,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" diff --git a/egs/librispeech/ASR/transducer_stateless2/joiner.py b/egs/librispeech/ASR/transducer_stateless2/joiner.py index 765f0be8b..4882f9971 100644 --- a/egs/librispeech/ASR/transducer_stateless2/joiner.py +++ b/egs/librispeech/ASR/transducer_stateless2/joiner.py @@ -14,6 +14,8 @@ # See the License for the specific language governing permissions and # limitations under the License. +from typing import Optional + import torch import torch.nn as nn @@ -30,7 +32,8 @@ class Joiner(nn.Module): self, encoder_out: torch.Tensor, decoder_out: torch.Tensor, - *unused, + unused_encoder_out_len: Optional[torch.Tensor] = None, + unused_decoder_out_len: Optional[torch.Tensor] = None, ) -> torch.Tensor: """ Args: @@ -38,10 +41,12 @@ class Joiner(nn.Module): Output from the encoder. Its shape is (N, T, self.input_dim). decoder_out: Output from the decoder. Its shape is (N, U, self.input_dim). - unused: + unused_encoder_out_len: This is a placeholder so that we can reuse transducer_stateless/beam_search.py in this folder as that script assumes the joiner networks accepts 4 inputs. + unused_decoder_out_len: + Just a placeholder. Returns: Return a tensor of shape (N, T, U, self.output_dim). """ diff --git a/egs/librispeech/ASR/transducer_stateless2/test_model.py b/egs/librispeech/ASR/transducer_stateless2/test_model.py new file mode 100755 index 000000000..bd2230a45 --- /dev/null +++ b/egs/librispeech/ASR/transducer_stateless2/test_model.py @@ -0,0 +1,49 @@ +#!/usr/bin/env python3 +# Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang) +# +# See ../../../../LICENSE for clarification regarding multiple authors +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +""" +To run this file, do: + + cd icefall/egs/librispeech/ASR + python ./transducer_stateless2/test_model.py +""" + +import torch +from train import get_params, get_transducer_model + + +def test_model(): + params = get_params() + params.vocab_size = 500 + params.blank_id = 0 + params.context_size = 2 + + model = get_transducer_model(params) + + num_param = sum([p.numel() for p in model.parameters()]) + print(f"Number of model parameters: {num_param}") + model.__class__.forward = torch.jit.ignore(model.__class__.forward) + torch.jit.script(model) + + +def main(): + test_model() + + +if __name__ == "__main__": + main() diff --git a/egs/librispeech/ASR/transducer_stateless_multi_datasets/export.py b/egs/librispeech/ASR/transducer_stateless_multi_datasets/export.py index 7d14d011d..b6b69d932 100755 --- a/egs/librispeech/ASR/transducer_stateless_multi_datasets/export.py +++ b/egs/librispeech/ASR/transducer_stateless_multi_datasets/export.py @@ -184,8 +184,6 @@ def main(): args = get_parser().parse_args() args.exp_dir = Path(args.exp_dir) - assert args.jit is False, "Support torchscript will be added later" - params = get_params() params.update(vars(args)) @@ -229,6 +227,11 @@ def main(): model.eval() if params.jit: + # We won't use the forward() method of the model in C++, so just ignore + # it here. + # Otherwise, one of its arguments is a ragged tensor and is not + # torch scriptabe. + model.__class__.forward = torch.jit.ignore(model.__class__.forward) logging.info("Using torch.jit.script") model = torch.jit.script(model) filename = params.exp_dir / "cpu_jit.pt" From c2c46ea02311ac30bd366efafcc70c1d1f7e0efc Mon Sep 17 00:00:00 2001 From: Daniel Povey Date: Wed, 18 May 2022 21:42:57 +0800 Subject: [PATCH 11/16] Update diagnostics, hopefully print more stats. # Conflicts: # egs/librispeech/ASR/pruned_transducer_stateless4b/train.py --- icefall/diagnostics.py | 290 +++++++++++++++++------------------------ 1 file changed, 120 insertions(+), 170 deletions(-) diff --git a/icefall/diagnostics.py b/icefall/diagnostics.py index bc8fe3069..d0a0461d0 100644 --- a/icefall/diagnostics.py +++ b/icefall/diagnostics.py @@ -19,7 +19,7 @@ import random from typing import List, Optional, Tuple - +from dataclasses import dataclass import torch from torch import Tensor, nn @@ -94,138 +94,12 @@ def get_tensor_stats( return x, count -def get_diagnostics_for_dim( - dim: int, - tensors: List[Tensor], - options: TensorDiagnosticOptions, - sizes_same: bool, - stats_type: str, -) -> str: - """ - This function gets diagnostics for a dimension of a module. - - Args: - dim: - the dimension to analyze, with 0 <= dim < tensors[0].ndim - options: - options object - sizes_same: - True if all the tensor sizes are the same on this dimension - stats_type: either "abs" or "positive" or "eigs" or "value", - imdictates the type of stats we accumulate, abs is mean absolute - value, "positive" is proportion of positive to nonnegative values, - "eigs" is eigenvalues after doing outer product on this dim, sum - over all other dimes. - Returns: - Diagnostic as a string, either percentiles or the actual values, - see the code. Will return the empty string if the diagnostics did - not make sense to print out for this dimension, e.g. dimension - mismatch and stats_type == "eigs". - """ - - # stats_and_counts is a list of pair (Tensor, int) - stats_and_counts = [get_tensor_stats(x, dim, stats_type) for x in tensors] - stats = [x[0] for x in stats_and_counts] - counts = [x[1] for x in stats_and_counts] - - if stats_type == "eigs": - try: - stats = torch.stack(stats).sum(dim=0) - except: # noqa - return "" - count = sum(counts) - stats = stats / count - try: - eigs, _ = torch.symeig(stats) - stats = eigs.abs().sqrt() - except: # noqa - print("Error getting eigenvalues, trying another method.") - eigs = torch.linalg.eigvals(stats) - stats = eigs.abs().sqrt() - # sqrt so it reflects data magnitude, like stddev- not variance - elif sizes_same: - stats = torch.stack(stats).sum(dim=0) - count = sum(counts) - stats = stats / count - else: - stats = [x[0] / x[1] for x in stats_and_counts] - stats = torch.cat(stats, dim=0) - if stats_type == "rms": - stats = stats.sqrt() - - # if `summarize` we print percentiles of the stats; else, - # we print out individual elements. - summarize = (not sizes_same) or options.dim_is_summarized(stats.numel()) - if summarize: - # print out percentiles. - stats = stats.sort()[0] - num_percentiles = 10 - size = stats.numel() - percentiles = [] - for i in range(num_percentiles + 1): - index = (i * (size - 1)) // num_percentiles - percentiles.append(stats[index].item()) - percentiles = ["%.2g" % x for x in percentiles] - percentiles = " ".join(percentiles) - ans = f"percentiles: [{percentiles}]" - else: - ans = stats.tolist() - ans = ["%.2g" % x for x in ans] - ans = "[" + " ".join(ans) + "]" - if stats_type == "value": - # This norm is useful because it is strictly less than the largest - # sqrt(eigenvalue) of the variance, which we print out, and shows, - # speaking in an approximate way, how much of that largest eigenvalue - # can be attributed to the mean of the distribution. - norm = (stats ** 2).sum().sqrt().item() - mean = stats.mean().item() - rms = (stats ** 2).mean().sqrt().item() - ans += f", norm={norm:.2g}, mean={mean:.2g}, rms={rms:.2g}" - else: - mean = stats.mean().item() - rms = (stats ** 2).mean().sqrt().item() - ans += f", mean={mean:.2g}, rms={rms:.2g}" - return ans -def print_diagnostics_for_dim( - name: str, dim: int, tensors: List[Tensor], options: TensorDiagnosticOptions -): - """This function prints diagnostics for a dimension of a tensor. - - Args: - name: - The tensor name. - dim: - The dimension to analyze, with 0 <= dim < tensors[0].ndim. - tensors: - List of cached tensors to get the stats. - options: - Options object. - """ - - ndim = tensors[0].ndim - if ndim > 1: - stats_types = ["abs", "positive", "value", "rms"] - if tensors[0].shape[dim] <= options.max_eig_dim: - stats_types.append("eigs") - else: - stats_types = ["value", "abs"] - - for stats_type in stats_types: - sizes = [x.shape[dim] for x in tensors] - sizes_same = all([x == sizes[0] for x in sizes]) - s = get_diagnostics_for_dim( - dim, tensors, options, sizes_same, stats_type - ) - if s == "": - continue - - min_size = min(sizes) - max_size = max(sizes) - size_str = f"{min_size}" if sizes_same else f"{min_size}..{max_size}" - # stats_type will be "abs" or "positive". - print(f"module={name}, dim={dim}, size={size_str}, {stats_type} {s}") +@dataclass +class TensorAndCount: + tensor: Tensor + count: int class TensorDiagnostic(object): @@ -238,12 +112,23 @@ class TensorDiagnostic(object): name: The tensor name. """ - def __init__(self, opts: TensorDiagnosticOptions, name: str): self.name = name self.opts = opts - # A list to cache the tensors. - self.saved_tensors = [] + + + self.stats = None # we'll later assign a list to this data member. It's a list of dict. + + # the keys into self.stats[dim] are strings, whose values can be + # "abs", "value", "positive", "rms", "value". + # The values e.g. self.stats[dim]["rms"] are lists of dataclass TensorAndCount, + # containing a tensor and its associated count (which is the sum of the other dims + # that we aggregated over, e.g. the number of frames and/or batch elements and/or + # channels. + # ... we actually accumulate the Tensors / counts any time we have the same-dim tensor, + # only adding a new element to the list if there was a different dim. + # if the string in the key is "eigs", if we detect a length mismatch we put None as the value. + def accumulate(self, x): """Accumulate tensors.""" @@ -251,50 +136,115 @@ class TensorDiagnostic(object): x = x[0] if not isinstance(x, Tensor): return - if x.device == torch.device("cpu"): - x = x.detach().clone() - else: - x = x.detach().to("cpu", non_blocking=True) - self.saved_tensors.append(x) - num = len(self.saved_tensors) - if num & (num - 1) == 0: # power of 2.. - self._limit_memory() + x = x.detach().clone() + if x.ndim == 0: + x = x.unsqueeze(0) + ndim = x.ndim + if self.stats is None: + self.stats = [ dict() for _ in range(ndim) ] - def _limit_memory(self): - """Only keep the newly cached tensors to limit memory.""" - if len(self.saved_tensors) > 1024: - self.saved_tensors = self.saved_tensors[-1024:] - return + for dim in range(ndim): + this_dim_stats = self.stats[dim] + if ndim > 1: + stats_types = ["abs", "positive", "value", "rms"] + if x.shape[dim] <= self.opts.max_eig_dim: + stats_types.append("eigs") + else: + stats_types = ["value", "abs"] + this_dict = self.stats[dim] + for stats_type in stats_types: + stats, count = get_tensor_stats(x, dim, stats_type) + if not stats_type in this_dim_stats: + this_dim_stats[stats_type] = [] # list of TensorAndCount + + done = False + if this_dim_stats[stats_type] is None: + # we can reach here if we detected for stats_type "eigs" that + # where was more than one different size for this dim. Then we + # disable accumulating this stats type, as it uses too much memory. + continue + for s in this_dim_stats[stats_type]: + if s.tensor.shape == stats.shape: + s.tensor += stats + s.count += count + done = True + break + if not done: + if this_dim_stats[stats_type] != [] and stats_type == "eigs": + # >1 size encountered on this dim, e.g. it's a batch or time dimension, + # don't accumulat "eigs" stats type, it uses too much memory + this_dim_stats[stats_type] = None + else: + this_dim_stats[stats_type].append(TensorAndCount(stats, count)) - tot_mem = 0.0 - for i in reversed(range(len(self.saved_tensors))): - tot_mem += ( - self.saved_tensors[i].numel() - * self.saved_tensors[i].element_size() - ) - if tot_mem > self.opts.memory_limit: - self.saved_tensors = self.saved_tensors[i:] - return def print_diagnostics(self): """Print diagnostics for each dimension of the tensor.""" - if len(self.saved_tensors) == 0: - print("{name}: no stats".format(name=self.name)) - return + for dim, this_dim_stats in enumerate(self.stats): + for stats_type, stats_list in this_dim_stats.items(): + # stats_type could be "rms", "value", "abs", "eigs", "positive". + # "value" could be a list of TensorAndCount, or None + if stats_list is None: + assert stats_type == "eigs" + continue - if self.saved_tensors[0].ndim == 0: - # Ensure there is at least one dim. - self.saved_tensors = [x.unsqueeze(0) for x in self.saved_tensors] + if stats_type == "eigs": + assert len(stats_list) == 1 + stats = stats_list[0].tensor / stats_list[0].count + try: + eigs, _ = torch.symeig(stats) + stats = eigs.abs().sqrt() + except: # noqa + print("Error getting eigenvalues, trying another method.") + eigs = torch.linalg.eigvals(stats) + stats = eigs.abs().sqrt() + # sqrt so it reflects data magnitude, like stddev- not variance + elif len(stats_list) == 1: + stats = stats_list[0].tensor / stats_list[0].count + else: + stats = torch.cat([x.tensor / x.count for x in stats_list], dim=0) - try: - device = torch.device("cuda") - except: # noqa - device = torch.device("cpu") + if stats_type == "rms": + # we stored the square; after aggregation we need to take sqrt. + stats = stats.sqrt() + + # if `summarize` we print percentiles of the stats; else, + # we print out individual elements. + summarize = (len(stats_list) > 1) or self.opts.dim_is_summarized(stats.numel()) + if summarize: # usually `summarize` will be true + # print out percentiles. + stats = stats.sort()[0] + num_percentiles = 10 + size = stats.numel() + percentiles = [] + for i in range(num_percentiles + 1): + index = (i * (size - 1)) // num_percentiles + percentiles.append(stats[index].item()) + percentiles = ["%.2g" % x for x in percentiles] + percentiles = " ".join(percentiles) + ans = f"percentiles: [{percentiles}]" + else: + ans = stats.tolist() + ans = ["%.2g" % x for x in ans] + ans = "[" + " ".join(ans) + "]" + if stats_type == "value": + # This norm is useful because it is strictly less than the largest + # sqrt(eigenvalue) of the variance, which we print out, and shows, + # speaking in an approximate way, how much of that largest eigenvalue + # can be attributed to the mean of the distribution. + norm = (stats ** 2).sum().sqrt().item() + ans += f", norm={norm:.2g}" + mean = stats.mean().item() + rms = (stats ** 2).mean().sqrt().item() + ans += f", mean={mean:.2g}, rms={rms:.2g}" + + # OK, "ans" contains the actual stats, e.g. + # ans = "percentiles: [0.43 0.46 0.48 0.49 0.49 0.5 0.51 0.52 0.53 0.54 0.59], mean=0.5, rms=0.5" + + sizes = [x.tensor.shape[0] for x in stats_list] + size_str = f"{sizes[0]}" if len(sizes) == 1 else f"{min(sizes)}..{max(sizes)}" + print(f"module={self.name}, dim={dim}, size={size_str}, {stats_type} {ans}") - ndim = self.saved_tensors[0].ndim - tensors = [x.to(device) for x in self.saved_tensors] - for dim in range(ndim): - print_diagnostics_for_dim(self.name, dim, tensors, self.opts) class ModelDiagnostic(object): From c0fdfabaf32a3a5c7bda1d04a7263b7aa7a268dd Mon Sep 17 00:00:00 2001 From: Daniel Povey Date: Thu, 19 May 2022 11:30:56 +0800 Subject: [PATCH 12/16] Remove memory-limit options arg --- icefall/diagnostics.py | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/icefall/diagnostics.py b/icefall/diagnostics.py index d0a0461d0..1cd685d37 100644 --- a/icefall/diagnostics.py +++ b/icefall/diagnostics.py @@ -28,16 +28,12 @@ class TensorDiagnosticOptions(object): """Options object for tensor diagnostics: Args: - memory_limit: - The maximum number of bytes per tensor - (limits how many copies of the tensor we cache). max_eig_dim: The maximum dimension for which we print out eigenvalues (limited for speed reasons). """ - def __init__(self, memory_limit: int = (2 ** 20), max_eig_dim: int = 512): - self.memory_limit = memory_limit + def __init__(self, max_eig_dim: int = 512): self.max_eig_dim = max_eig_dim def dim_is_summarized(self, size: int): From c736b39c7d20c37faabe7b511566a0d9b584de03 Mon Sep 17 00:00:00 2001 From: Daniel Povey Date: Thu, 19 May 2022 11:35:54 +0800 Subject: [PATCH 13/16] Remove unnecessary option for diagnostics code, collect on more batches --- egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py | 7 ++----- egs/librispeech/ASR/pruned_transducer_stateless2/train.py | 7 ++----- egs/librispeech/ASR/pruned_transducer_stateless3/train.py | 7 ++----- egs/librispeech/ASR/pruned_transducer_stateless4/train.py | 7 ++----- egs/librispeech/ASR/transducer_stateless/train.py | 7 ++----- egs/librispeech/ASR/transducer_stateless2/train.py | 7 ++----- egs/spgispeech/ASR/pruned_transducer_stateless2/train.py | 7 ++----- 7 files changed, 14 insertions(+), 35 deletions(-) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py index 4421ce2aa..83ae25561 100755 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py @@ -689,7 +689,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -831,10 +831,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) gigaspeech = GigaSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/train.py b/egs/librispeech/ASR/pruned_transducer_stateless2/train.py index 51c1a231a..eed2df755 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/train.py @@ -695,7 +695,7 @@ def train_one_epoch( display_and_save_batch(batch, params=params, sp=sp) raise - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -839,10 +839,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py index 037f99bc7..f5a25a226 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py @@ -767,7 +767,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -938,10 +938,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeech(manifest_dir=args.manifest_dir) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless4/train.py b/egs/librispeech/ASR/pruned_transducer_stateless4/train.py index 4ff69d521..ca7207122 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless4/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless4/train.py @@ -724,7 +724,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -888,10 +888,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/transducer_stateless/train.py b/egs/librispeech/ASR/transducer_stateless/train.py index 89f754b20..cb7f08a09 100755 --- a/egs/librispeech/ASR/transducer_stateless/train.py +++ b/egs/librispeech/ASR/transducer_stateless/train.py @@ -523,7 +523,7 @@ def train_one_epoch( loss.backward() clip_grad_norm_(model.parameters(), 5.0, 2.0) optimizer.step() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if batch_idx % params.log_interval == 0: @@ -635,10 +635,7 @@ def run(rank, world_size, args): librispeech = LibriSpeechAsrDataModule(args) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) train_cuts = librispeech.train_clean_100_cuts() if params.full_libri: diff --git a/egs/librispeech/ASR/transducer_stateless2/train.py b/egs/librispeech/ASR/transducer_stateless2/train.py index 8ceffb489..cb13e317c 100755 --- a/egs/librispeech/ASR/transducer_stateless2/train.py +++ b/egs/librispeech/ASR/transducer_stateless2/train.py @@ -511,7 +511,7 @@ def train_one_epoch( loss.backward() clip_grad_norm_(model.parameters(), 5.0, 2.0) optimizer.step() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if batch_idx % params.log_interval == 0: @@ -623,10 +623,7 @@ def run(rank, world_size, args): librispeech = LibriSpeechAsrDataModule(args) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) train_cuts = librispeech.train_clean_100_cuts() if params.full_libri: diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py index 6c66bfb62..dda29b3e5 100755 --- a/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py @@ -690,7 +690,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -832,10 +832,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) spgispeech = SPGISpeechAsrDataModule(args) From 4e23fb2252cafd94ad4c401abc310a5e56fc1f8b Mon Sep 17 00:00:00 2001 From: Daniel Povey Date: Thu, 19 May 2022 11:45:59 +0800 Subject: [PATCH 14/16] Improve diagnostics code memory-wise and accumulate more stats. (#373) * Update diagnostics, hopefully print more stats. # Conflicts: # egs/librispeech/ASR/pruned_transducer_stateless4b/train.py * Remove memory-limit options arg * Remove unnecessary option for diagnostics code, collect on more batches --- .../ASR/pruned_transducer_stateless2/train.py | 7 +- .../ASR/pruned_transducer_stateless2/train.py | 7 +- .../ASR/pruned_transducer_stateless3/train.py | 7 +- .../ASR/pruned_transducer_stateless4/train.py | 7 +- .../ASR/transducer_stateless/train.py | 7 +- .../ASR/transducer_stateless2/train.py | 7 +- .../ASR/pruned_transducer_stateless2/train.py | 7 +- icefall/diagnostics.py | 296 +++++++----------- 8 files changed, 135 insertions(+), 210 deletions(-) diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py index 4421ce2aa..83ae25561 100755 --- a/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/gigaspeech/ASR/pruned_transducer_stateless2/train.py @@ -689,7 +689,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -831,10 +831,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) gigaspeech = GigaSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless2/train.py b/egs/librispeech/ASR/pruned_transducer_stateless2/train.py index 51c1a231a..eed2df755 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless2/train.py @@ -695,7 +695,7 @@ def train_one_epoch( display_and_save_batch(batch, params=params, sp=sp) raise - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -839,10 +839,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py index 037f99bc7..f5a25a226 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/train.py @@ -767,7 +767,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -938,10 +938,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeech(manifest_dir=args.manifest_dir) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless4/train.py b/egs/librispeech/ASR/pruned_transducer_stateless4/train.py index 4ff69d521..ca7207122 100755 --- a/egs/librispeech/ASR/pruned_transducer_stateless4/train.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless4/train.py @@ -724,7 +724,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -888,10 +888,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) librispeech = LibriSpeechAsrDataModule(args) diff --git a/egs/librispeech/ASR/transducer_stateless/train.py b/egs/librispeech/ASR/transducer_stateless/train.py index 89f754b20..cb7f08a09 100755 --- a/egs/librispeech/ASR/transducer_stateless/train.py +++ b/egs/librispeech/ASR/transducer_stateless/train.py @@ -523,7 +523,7 @@ def train_one_epoch( loss.backward() clip_grad_norm_(model.parameters(), 5.0, 2.0) optimizer.step() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if batch_idx % params.log_interval == 0: @@ -635,10 +635,7 @@ def run(rank, world_size, args): librispeech = LibriSpeechAsrDataModule(args) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) train_cuts = librispeech.train_clean_100_cuts() if params.full_libri: diff --git a/egs/librispeech/ASR/transducer_stateless2/train.py b/egs/librispeech/ASR/transducer_stateless2/train.py index 8ceffb489..cb13e317c 100755 --- a/egs/librispeech/ASR/transducer_stateless2/train.py +++ b/egs/librispeech/ASR/transducer_stateless2/train.py @@ -511,7 +511,7 @@ def train_one_epoch( loss.backward() clip_grad_norm_(model.parameters(), 5.0, 2.0) optimizer.step() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if batch_idx % params.log_interval == 0: @@ -623,10 +623,7 @@ def run(rank, world_size, args): librispeech = LibriSpeechAsrDataModule(args) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) train_cuts = librispeech.train_clean_100_cuts() if params.full_libri: diff --git a/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py index 6c66bfb62..dda29b3e5 100755 --- a/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py +++ b/egs/spgispeech/ASR/pruned_transducer_stateless2/train.py @@ -690,7 +690,7 @@ def train_one_epoch( scaler.update() optimizer.zero_grad() - if params.print_diagnostics and batch_idx == 5: + if params.print_diagnostics and batch_idx == 30: return if ( @@ -832,10 +832,7 @@ def run(rank, world_size, args): scheduler.load_state_dict(checkpoints["scheduler"]) if params.print_diagnostics: - opts = diagnostics.TensorDiagnosticOptions( - 2 ** 22 - ) # allow 4 megabytes per sub-module - diagnostic = diagnostics.attach_diagnostics(model, opts) + diagnostic = diagnostics.attach_diagnostics(model) spgispeech = SPGISpeechAsrDataModule(args) diff --git a/icefall/diagnostics.py b/icefall/diagnostics.py index bc8fe3069..1cd685d37 100644 --- a/icefall/diagnostics.py +++ b/icefall/diagnostics.py @@ -19,7 +19,7 @@ import random from typing import List, Optional, Tuple - +from dataclasses import dataclass import torch from torch import Tensor, nn @@ -28,16 +28,12 @@ class TensorDiagnosticOptions(object): """Options object for tensor diagnostics: Args: - memory_limit: - The maximum number of bytes per tensor - (limits how many copies of the tensor we cache). max_eig_dim: The maximum dimension for which we print out eigenvalues (limited for speed reasons). """ - def __init__(self, memory_limit: int = (2 ** 20), max_eig_dim: int = 512): - self.memory_limit = memory_limit + def __init__(self, max_eig_dim: int = 512): self.max_eig_dim = max_eig_dim def dim_is_summarized(self, size: int): @@ -94,138 +90,12 @@ def get_tensor_stats( return x, count -def get_diagnostics_for_dim( - dim: int, - tensors: List[Tensor], - options: TensorDiagnosticOptions, - sizes_same: bool, - stats_type: str, -) -> str: - """ - This function gets diagnostics for a dimension of a module. - - Args: - dim: - the dimension to analyze, with 0 <= dim < tensors[0].ndim - options: - options object - sizes_same: - True if all the tensor sizes are the same on this dimension - stats_type: either "abs" or "positive" or "eigs" or "value", - imdictates the type of stats we accumulate, abs is mean absolute - value, "positive" is proportion of positive to nonnegative values, - "eigs" is eigenvalues after doing outer product on this dim, sum - over all other dimes. - Returns: - Diagnostic as a string, either percentiles or the actual values, - see the code. Will return the empty string if the diagnostics did - not make sense to print out for this dimension, e.g. dimension - mismatch and stats_type == "eigs". - """ - - # stats_and_counts is a list of pair (Tensor, int) - stats_and_counts = [get_tensor_stats(x, dim, stats_type) for x in tensors] - stats = [x[0] for x in stats_and_counts] - counts = [x[1] for x in stats_and_counts] - - if stats_type == "eigs": - try: - stats = torch.stack(stats).sum(dim=0) - except: # noqa - return "" - count = sum(counts) - stats = stats / count - try: - eigs, _ = torch.symeig(stats) - stats = eigs.abs().sqrt() - except: # noqa - print("Error getting eigenvalues, trying another method.") - eigs = torch.linalg.eigvals(stats) - stats = eigs.abs().sqrt() - # sqrt so it reflects data magnitude, like stddev- not variance - elif sizes_same: - stats = torch.stack(stats).sum(dim=0) - count = sum(counts) - stats = stats / count - else: - stats = [x[0] / x[1] for x in stats_and_counts] - stats = torch.cat(stats, dim=0) - if stats_type == "rms": - stats = stats.sqrt() - - # if `summarize` we print percentiles of the stats; else, - # we print out individual elements. - summarize = (not sizes_same) or options.dim_is_summarized(stats.numel()) - if summarize: - # print out percentiles. - stats = stats.sort()[0] - num_percentiles = 10 - size = stats.numel() - percentiles = [] - for i in range(num_percentiles + 1): - index = (i * (size - 1)) // num_percentiles - percentiles.append(stats[index].item()) - percentiles = ["%.2g" % x for x in percentiles] - percentiles = " ".join(percentiles) - ans = f"percentiles: [{percentiles}]" - else: - ans = stats.tolist() - ans = ["%.2g" % x for x in ans] - ans = "[" + " ".join(ans) + "]" - if stats_type == "value": - # This norm is useful because it is strictly less than the largest - # sqrt(eigenvalue) of the variance, which we print out, and shows, - # speaking in an approximate way, how much of that largest eigenvalue - # can be attributed to the mean of the distribution. - norm = (stats ** 2).sum().sqrt().item() - mean = stats.mean().item() - rms = (stats ** 2).mean().sqrt().item() - ans += f", norm={norm:.2g}, mean={mean:.2g}, rms={rms:.2g}" - else: - mean = stats.mean().item() - rms = (stats ** 2).mean().sqrt().item() - ans += f", mean={mean:.2g}, rms={rms:.2g}" - return ans -def print_diagnostics_for_dim( - name: str, dim: int, tensors: List[Tensor], options: TensorDiagnosticOptions -): - """This function prints diagnostics for a dimension of a tensor. - - Args: - name: - The tensor name. - dim: - The dimension to analyze, with 0 <= dim < tensors[0].ndim. - tensors: - List of cached tensors to get the stats. - options: - Options object. - """ - - ndim = tensors[0].ndim - if ndim > 1: - stats_types = ["abs", "positive", "value", "rms"] - if tensors[0].shape[dim] <= options.max_eig_dim: - stats_types.append("eigs") - else: - stats_types = ["value", "abs"] - - for stats_type in stats_types: - sizes = [x.shape[dim] for x in tensors] - sizes_same = all([x == sizes[0] for x in sizes]) - s = get_diagnostics_for_dim( - dim, tensors, options, sizes_same, stats_type - ) - if s == "": - continue - - min_size = min(sizes) - max_size = max(sizes) - size_str = f"{min_size}" if sizes_same else f"{min_size}..{max_size}" - # stats_type will be "abs" or "positive". - print(f"module={name}, dim={dim}, size={size_str}, {stats_type} {s}") +@dataclass +class TensorAndCount: + tensor: Tensor + count: int class TensorDiagnostic(object): @@ -238,12 +108,23 @@ class TensorDiagnostic(object): name: The tensor name. """ - def __init__(self, opts: TensorDiagnosticOptions, name: str): self.name = name self.opts = opts - # A list to cache the tensors. - self.saved_tensors = [] + + + self.stats = None # we'll later assign a list to this data member. It's a list of dict. + + # the keys into self.stats[dim] are strings, whose values can be + # "abs", "value", "positive", "rms", "value". + # The values e.g. self.stats[dim]["rms"] are lists of dataclass TensorAndCount, + # containing a tensor and its associated count (which is the sum of the other dims + # that we aggregated over, e.g. the number of frames and/or batch elements and/or + # channels. + # ... we actually accumulate the Tensors / counts any time we have the same-dim tensor, + # only adding a new element to the list if there was a different dim. + # if the string in the key is "eigs", if we detect a length mismatch we put None as the value. + def accumulate(self, x): """Accumulate tensors.""" @@ -251,50 +132,115 @@ class TensorDiagnostic(object): x = x[0] if not isinstance(x, Tensor): return - if x.device == torch.device("cpu"): - x = x.detach().clone() - else: - x = x.detach().to("cpu", non_blocking=True) - self.saved_tensors.append(x) - num = len(self.saved_tensors) - if num & (num - 1) == 0: # power of 2.. - self._limit_memory() + x = x.detach().clone() + if x.ndim == 0: + x = x.unsqueeze(0) + ndim = x.ndim + if self.stats is None: + self.stats = [ dict() for _ in range(ndim) ] - def _limit_memory(self): - """Only keep the newly cached tensors to limit memory.""" - if len(self.saved_tensors) > 1024: - self.saved_tensors = self.saved_tensors[-1024:] - return + for dim in range(ndim): + this_dim_stats = self.stats[dim] + if ndim > 1: + stats_types = ["abs", "positive", "value", "rms"] + if x.shape[dim] <= self.opts.max_eig_dim: + stats_types.append("eigs") + else: + stats_types = ["value", "abs"] + this_dict = self.stats[dim] + for stats_type in stats_types: + stats, count = get_tensor_stats(x, dim, stats_type) + if not stats_type in this_dim_stats: + this_dim_stats[stats_type] = [] # list of TensorAndCount + + done = False + if this_dim_stats[stats_type] is None: + # we can reach here if we detected for stats_type "eigs" that + # where was more than one different size for this dim. Then we + # disable accumulating this stats type, as it uses too much memory. + continue + for s in this_dim_stats[stats_type]: + if s.tensor.shape == stats.shape: + s.tensor += stats + s.count += count + done = True + break + if not done: + if this_dim_stats[stats_type] != [] and stats_type == "eigs": + # >1 size encountered on this dim, e.g. it's a batch or time dimension, + # don't accumulat "eigs" stats type, it uses too much memory + this_dim_stats[stats_type] = None + else: + this_dim_stats[stats_type].append(TensorAndCount(stats, count)) - tot_mem = 0.0 - for i in reversed(range(len(self.saved_tensors))): - tot_mem += ( - self.saved_tensors[i].numel() - * self.saved_tensors[i].element_size() - ) - if tot_mem > self.opts.memory_limit: - self.saved_tensors = self.saved_tensors[i:] - return def print_diagnostics(self): """Print diagnostics for each dimension of the tensor.""" - if len(self.saved_tensors) == 0: - print("{name}: no stats".format(name=self.name)) - return + for dim, this_dim_stats in enumerate(self.stats): + for stats_type, stats_list in this_dim_stats.items(): + # stats_type could be "rms", "value", "abs", "eigs", "positive". + # "value" could be a list of TensorAndCount, or None + if stats_list is None: + assert stats_type == "eigs" + continue - if self.saved_tensors[0].ndim == 0: - # Ensure there is at least one dim. - self.saved_tensors = [x.unsqueeze(0) for x in self.saved_tensors] + if stats_type == "eigs": + assert len(stats_list) == 1 + stats = stats_list[0].tensor / stats_list[0].count + try: + eigs, _ = torch.symeig(stats) + stats = eigs.abs().sqrt() + except: # noqa + print("Error getting eigenvalues, trying another method.") + eigs = torch.linalg.eigvals(stats) + stats = eigs.abs().sqrt() + # sqrt so it reflects data magnitude, like stddev- not variance + elif len(stats_list) == 1: + stats = stats_list[0].tensor / stats_list[0].count + else: + stats = torch.cat([x.tensor / x.count for x in stats_list], dim=0) - try: - device = torch.device("cuda") - except: # noqa - device = torch.device("cpu") + if stats_type == "rms": + # we stored the square; after aggregation we need to take sqrt. + stats = stats.sqrt() + + # if `summarize` we print percentiles of the stats; else, + # we print out individual elements. + summarize = (len(stats_list) > 1) or self.opts.dim_is_summarized(stats.numel()) + if summarize: # usually `summarize` will be true + # print out percentiles. + stats = stats.sort()[0] + num_percentiles = 10 + size = stats.numel() + percentiles = [] + for i in range(num_percentiles + 1): + index = (i * (size - 1)) // num_percentiles + percentiles.append(stats[index].item()) + percentiles = ["%.2g" % x for x in percentiles] + percentiles = " ".join(percentiles) + ans = f"percentiles: [{percentiles}]" + else: + ans = stats.tolist() + ans = ["%.2g" % x for x in ans] + ans = "[" + " ".join(ans) + "]" + if stats_type == "value": + # This norm is useful because it is strictly less than the largest + # sqrt(eigenvalue) of the variance, which we print out, and shows, + # speaking in an approximate way, how much of that largest eigenvalue + # can be attributed to the mean of the distribution. + norm = (stats ** 2).sum().sqrt().item() + ans += f", norm={norm:.2g}" + mean = stats.mean().item() + rms = (stats ** 2).mean().sqrt().item() + ans += f", mean={mean:.2g}, rms={rms:.2g}" + + # OK, "ans" contains the actual stats, e.g. + # ans = "percentiles: [0.43 0.46 0.48 0.49 0.49 0.5 0.51 0.52 0.53 0.54 0.59], mean=0.5, rms=0.5" + + sizes = [x.tensor.shape[0] for x in stats_list] + size_str = f"{sizes[0]}" if len(sizes) == 1 else f"{min(sizes)}..{max(sizes)}" + print(f"module={self.name}, dim={dim}, size={size_str}, {stats_type} {ans}") - ndim = self.saved_tensors[0].ndim - tensors = [x.to(device) for x in self.saved_tensors] - for dim in range(ndim): - print_diagnostics_for_dim(self.name, dim, tensors, self.opts) class ModelDiagnostic(object): From 5230e73e41d1fb57c10a78c609ccc548bf8b4cf0 Mon Sep 17 00:00:00 2001 From: Daniel Povey Date: Thu, 19 May 2022 12:49:00 +0800 Subject: [PATCH 15/16] Small fixes --- icefall/diagnostics.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/icefall/diagnostics.py b/icefall/diagnostics.py index 1cd685d37..e8bedc64e 100644 --- a/icefall/diagnostics.py +++ b/icefall/diagnostics.py @@ -272,7 +272,7 @@ class ModelDiagnostic(object): def attach_diagnostics( - model: nn.Module, opts: TensorDiagnosticOptions + model: nn.Module, opts: Optional[TensorDiagnosticOptions] = None ) -> ModelDiagnostic: """Attach a ModelDiagnostic object to the model by 1) registering forward hook and backward hook on each module, to accumulate @@ -335,7 +335,7 @@ def attach_diagnostics( def _test_tensor_diagnostic(): - opts = TensorDiagnosticOptions(2 ** 20, 512) + opts = TensorDiagnosticOptions(512) diagnostic = TensorDiagnostic(opts, "foo") From ec5a112831b54471daa82ddb956112b88128577a Mon Sep 17 00:00:00 2001 From: Mingshuang Luo <37799481+luomingshuang@users.noreply.github.com> Date: Fri, 20 May 2022 19:30:38 +0800 Subject: [PATCH 16/16] [Ready to merge] Do some coding style checks for the latest files (#379) * style check * do changes for .flake8 * a change for compute_fbank_yesno.py --- .flake8 | 14 ++---- .../ASR/local/compute_fbank_musan.py | 21 ++++---- .../ASR/local/compute_fbank_spgispeech.py | 28 ++++++----- egs/spgispeech/ASR/local/prepare_splits.py | 10 ++-- egs/yesno/ASR/local/compute_fbank_yesno.py | 4 +- icefall/diagnostics.py | 49 ++++++++++++------- 6 files changed, 74 insertions(+), 52 deletions(-) diff --git a/.flake8 b/.flake8 index 8c497fac3..dbeec0b0c 100644 --- a/.flake8 +++ b/.flake8 @@ -4,15 +4,11 @@ statistics=true max-line-length = 80 per-file-ignores = # line too long - egs/librispeech/ASR/*/conformer.py: E501, - egs/aishell/ASR/*/conformer.py: E501, - egs/tedlium3/ASR/*/conformer.py: E501, - egs/gigaspeech/ASR/*/conformer.py: E501, - egs/librispeech/ASR/pruned_transducer_stateless2/*.py: E501, - egs/gigaspeech/ASR/pruned_transducer_stateless2/*.py: E501, - egs/librispeech/ASR/pruned_transducer_stateless4/*.py: E501, - egs/librispeech/ASR/*/optim.py: E501, - egs/librispeech/ASR/*/scaling.py: E501, + icefall/diagnostics.py: E501 + egs/*/ASR/*/conformer.py: E501, + egs/*/ASR/pruned_transducer_stateless*/*.py: E501, + egs/*/ASR/*/optim.py: E501, + egs/*/ASR/*/scaling.py: E501, # invalid escape sequence (cause by tex formular), W605 icefall/utils.py: E501, W605 diff --git a/egs/spgispeech/ASR/local/compute_fbank_musan.py b/egs/spgispeech/ASR/local/compute_fbank_musan.py index b4f409ba6..57805a756 100755 --- a/egs/spgispeech/ASR/local/compute_fbank_musan.py +++ b/egs/spgispeech/ASR/local/compute_fbank_musan.py @@ -27,17 +27,15 @@ import logging from pathlib import Path import torch -from lhotse import LilcomChunkyWriter, CutSet, combine +from lhotse import CutSet, LilcomChunkyWriter, combine from lhotse.features.kaldifeat import ( KaldifeatFbank, KaldifeatFbankConfig, - KaldifeatMelOptions, KaldifeatFrameOptions, + KaldifeatMelOptions, ) from lhotse.recipes.utils import read_manifests_if_cached -from icefall.utils import get_executor - # Torch's multithreaded behavior needs to be disabled or # it wastes a lot of CPU and slow things down. # Do this outside of main() in case it needs to take effect @@ -82,23 +80,28 @@ def compute_fbank_musan(): # create chunks of Musan with duration 5 - 10 seconds musan_cuts = ( CutSet.from_manifests( - recordings=combine(part["recordings"] for part in manifests.values()) + recordings=combine( + part["recordings"] for part in manifests.values() + ) ) .cut_into_windows(10.0) .filter(lambda c: c.duration > 5) .compute_and_store_features_batch( extractor=extractor, - storage_path=output_dir / f"feats_musan", - manifest_path=src_dir / f"cuts_musan.jsonl.gz", + storage_path=output_dir / "feats_musan", batch_duration=500, num_workers=4, storage_type=LilcomChunkyWriter, ) ) + logging.info(f"Saving to {musan_cuts_path}") + musan_cuts.to_file(musan_cuts_path) + if __name__ == "__main__": - formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" - + formatter = ( + "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + ) logging.basicConfig(format=formatter, level=logging.INFO) compute_fbank_musan() diff --git a/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py b/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py index cc8c8a670..b67754e2a 100755 --- a/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py +++ b/egs/spgispeech/ASR/local/compute_fbank_spgispeech.py @@ -25,17 +25,15 @@ The generated fbank features are saved in data/fbank. import argparse import logging from pathlib import Path -from tqdm import tqdm import torch -from lhotse import load_manifest_lazy, LilcomChunkyWriter +from lhotse import LilcomChunkyWriter, load_manifest_lazy from lhotse.features.kaldifeat import ( KaldifeatFbank, KaldifeatFbankConfig, - KaldifeatMelOptions, KaldifeatFrameOptions, + KaldifeatMelOptions, ) -from lhotse.manipulation import combine # Torch's multithreaded behavior needs to be disabled or # it wastes a lot of CPU and slow things down. @@ -97,27 +95,32 @@ def compute_fbank_spgispeech(args): ) if args.train: - logging.info(f"Processing train") - cut_set = load_manifest_lazy(src_dir / f"cuts_train_raw.jsonl.gz") + logging.info("Processing train") + cut_set = load_manifest_lazy(src_dir / "cuts_train_raw.jsonl.gz") chunk_size = len(cut_set) // args.num_splits cut_sets = cut_set.split_lazy( output_dir=src_dir / f"cuts_train_raw_split{args.num_splits}", chunk_size=chunk_size, ) start = args.start - stop = min(args.stop, args.num_splits) if args.stop > 0 else args.num_splits + stop = ( + min(args.stop, args.num_splits) + if args.stop > 0 + else args.num_splits + ) num_digits = len(str(args.num_splits)) for i in range(start, stop): idx = f"{i + 1}".zfill(num_digits) + cuts_train_idx_path = src_dir / f"cuts_train_{idx}.jsonl.gz" logging.info(f"Processing train split {i}") cs = cut_sets[i].compute_and_store_features_batch( extractor=extractor, storage_path=output_dir / f"feats_train_{idx}", - manifest_path=src_dir / f"cuts_train_{idx}.jsonl.gz", batch_duration=500, num_workers=4, storage_type=LilcomChunkyWriter, ) + cs.to_file(cuts_train_idx_path) if args.test: for partition in ["dev", "val"]: @@ -125,7 +128,9 @@ def compute_fbank_spgispeech(args): logging.info(f"{partition} already exists - skipping.") continue logging.info(f"Processing {partition}") - cut_set = load_manifest_lazy(src_dir / f"cuts_{partition}_raw.jsonl.gz") + cut_set = load_manifest_lazy( + src_dir / f"cuts_{partition}_raw.jsonl.gz" + ) cut_set = cut_set.compute_and_store_features_batch( extractor=extractor, storage_path=output_dir / f"feats_{partition}", @@ -137,8 +142,9 @@ def compute_fbank_spgispeech(args): if __name__ == "__main__": - formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" - + formatter = ( + "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + ) logging.basicConfig(format=formatter, level=logging.INFO) args = get_args() diff --git a/egs/spgispeech/ASR/local/prepare_splits.py b/egs/spgispeech/ASR/local/prepare_splits.py index 2d1818649..8c8f1c133 100755 --- a/egs/spgispeech/ASR/local/prepare_splits.py +++ b/egs/spgispeech/ASR/local/prepare_splits.py @@ -24,7 +24,6 @@ from pathlib import Path import torch from lhotse import CutSet - from lhotse.recipes.utils import read_manifests_if_cached # Torch's multithreaded behavior needs to be disabled or @@ -56,7 +55,9 @@ def split_spgispeech_train(): # Add speed perturbation train_cuts = ( - train_cuts + train_cuts.perturb_speed(0.9) + train_cuts.perturb_speed(1.1) + train_cuts + + train_cuts.perturb_speed(0.9) + + train_cuts.perturb_speed(1.1) ) # Write the manifests to disk. @@ -72,8 +73,9 @@ def split_spgispeech_train(): if __name__ == "__main__": - formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" - + formatter = ( + "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" + ) logging.basicConfig(format=formatter, level=logging.INFO) split_spgispeech_train() diff --git a/egs/yesno/ASR/local/compute_fbank_yesno.py b/egs/yesno/ASR/local/compute_fbank_yesno.py index 6072d4222..6922ffe10 100755 --- a/egs/yesno/ASR/local/compute_fbank_yesno.py +++ b/egs/yesno/ASR/local/compute_fbank_yesno.py @@ -38,7 +38,9 @@ def compute_fbank_yesno(): "test", ) manifests = read_manifests_if_cached( - dataset_parts=dataset_parts, output_dir=src_dir + dataset_parts=dataset_parts, + output_dir=src_dir, + prefix="yesno", ) assert manifests is not None diff --git a/icefall/diagnostics.py b/icefall/diagnostics.py index e8bedc64e..4850308d9 100644 --- a/icefall/diagnostics.py +++ b/icefall/diagnostics.py @@ -18,8 +18,9 @@ import random -from typing import List, Optional, Tuple from dataclasses import dataclass +from typing import Optional, Tuple + import torch from torch import Tensor, nn @@ -90,8 +91,6 @@ def get_tensor_stats( return x, count - - @dataclass class TensorAndCount: tensor: Tensor @@ -108,12 +107,12 @@ class TensorDiagnostic(object): name: The tensor name. """ + def __init__(self, opts: TensorDiagnosticOptions, name: str): self.name = name self.opts = opts - - self.stats = None # we'll later assign a list to this data member. It's a list of dict. + self.stats = None # we'll later assign a list to this data member. It's a list of dict. # the keys into self.stats[dim] are strings, whose values can be # "abs", "value", "positive", "rms", "value". @@ -125,7 +124,6 @@ class TensorDiagnostic(object): # only adding a new element to the list if there was a different dim. # if the string in the key is "eigs", if we detect a length mismatch we put None as the value. - def accumulate(self, x): """Accumulate tensors.""" if isinstance(x, Tuple): @@ -137,7 +135,7 @@ class TensorDiagnostic(object): x = x.unsqueeze(0) ndim = x.ndim if self.stats is None: - self.stats = [ dict() for _ in range(ndim) ] + self.stats = [dict() for _ in range(ndim)] for dim in range(ndim): this_dim_stats = self.stats[dim] @@ -147,10 +145,10 @@ class TensorDiagnostic(object): stats_types.append("eigs") else: stats_types = ["value", "abs"] - this_dict = self.stats[dim] + for stats_type in stats_types: stats, count = get_tensor_stats(x, dim, stats_type) - if not stats_type in this_dim_stats: + if stats_type not in this_dim_stats: this_dim_stats[stats_type] = [] # list of TensorAndCount done = False @@ -166,13 +164,17 @@ class TensorDiagnostic(object): done = True break if not done: - if this_dim_stats[stats_type] != [] and stats_type == "eigs": + if ( + this_dim_stats[stats_type] != [] + and stats_type == "eigs" + ): # >1 size encountered on this dim, e.g. it's a batch or time dimension, # don't accumulat "eigs" stats type, it uses too much memory this_dim_stats[stats_type] = None else: - this_dim_stats[stats_type].append(TensorAndCount(stats, count)) - + this_dim_stats[stats_type].append( + TensorAndCount(stats, count) + ) def print_diagnostics(self): """Print diagnostics for each dimension of the tensor.""" @@ -191,14 +193,18 @@ class TensorDiagnostic(object): eigs, _ = torch.symeig(stats) stats = eigs.abs().sqrt() except: # noqa - print("Error getting eigenvalues, trying another method.") + print( + "Error getting eigenvalues, trying another method." + ) eigs = torch.linalg.eigvals(stats) stats = eigs.abs().sqrt() # sqrt so it reflects data magnitude, like stddev- not variance elif len(stats_list) == 1: stats = stats_list[0].tensor / stats_list[0].count else: - stats = torch.cat([x.tensor / x.count for x in stats_list], dim=0) + stats = torch.cat( + [x.tensor / x.count for x in stats_list], dim=0 + ) if stats_type == "rms": # we stored the square; after aggregation we need to take sqrt. @@ -206,7 +212,9 @@ class TensorDiagnostic(object): # if `summarize` we print percentiles of the stats; else, # we print out individual elements. - summarize = (len(stats_list) > 1) or self.opts.dim_is_summarized(stats.numel()) + summarize = ( + len(stats_list) > 1 + ) or self.opts.dim_is_summarized(stats.numel()) if summarize: # usually `summarize` will be true # print out percentiles. stats = stats.sort()[0] @@ -238,9 +246,14 @@ class TensorDiagnostic(object): # ans = "percentiles: [0.43 0.46 0.48 0.49 0.49 0.5 0.51 0.52 0.53 0.54 0.59], mean=0.5, rms=0.5" sizes = [x.tensor.shape[0] for x in stats_list] - size_str = f"{sizes[0]}" if len(sizes) == 1 else f"{min(sizes)}..{max(sizes)}" - print(f"module={self.name}, dim={dim}, size={size_str}, {stats_type} {ans}") - + size_str = ( + f"{sizes[0]}" + if len(sizes) == 1 + else f"{min(sizes)}..{max(sizes)}" + ) + print( + f"module={self.name}, dim={dim}, size={size_str}, {stats_type} {ans}" + ) class ModelDiagnostic(object):