update train.py

egs/librispeech/ASR/whisper/train.py

@@ -23,7 +23,8 @@ torchrun --nproc_per_node 8 ./whisper/train.py \
   --max-duration 200 \
   --exp-dir whisper/exp_large_v2 \
   --model-name large-v2 \
-  --manifest-dir data/fbank_whisper \
+  --full-libri True \
+  --manifest-dir data/fbank_whisper_80D \
   --deepspeed \
   --deepspeed_config ./whisper/ds_config_zero1.json
 
@@ -31,7 +32,8 @@ torchrun --nproc_per_node 8 ./whisper/train.py \
 torchrun --nproc_per_node 8 ./whisper/train.py \
   --max-duration 200 \
   --exp-dir whisper/exp_medium \
-  --manifest-dir data/fbank_whisper \
+  --full-libri True \
+  --manifest-dir data/fbank_whisper_80D \
   --base-lr 1e-5 \
   --model-name medium
 """
@@ -53,7 +55,7 @@ import torch
 import torch.multiprocessing as mp
 import torch.nn as nn
 import whisper
-from asr_datamodule import AishellAsrDataModule
+from asr_datamodule import LibriSpeechAsrDataModule
 from deepspeed.utils.zero_to_fp32 import convert_zero_checkpoint_to_fp32_state_dict
 from label_smoothing import LabelSmoothingLoss
 from lhotse import CutSet, load_manifest
@@ -147,7 +149,7 @@ def get_parser():
         "--model-name",
         type=str,
         default="large-v2",
-        choices=["large-v2", "large-v3", "medium", "small", "tiny"],
+        choices=["large-v2", "large-v3", "medium", "medium.en", "small", "small.en", "tiny", "tiny.en"],
         help="""The model name to use.
         """,
     )
@@ -450,8 +452,7 @@ def compute_loss(
     batch_idx_train = params.batch_idx_train
 
     texts = batch["supervisions"]["text"]
-    # remove spaces in texts
-    texts = [text.replace(" ", "") for text in texts]
+    texts = [t[0] + t[1:].lower() for t in texts]
 
     text_tokens_list = [
         list(tokenizer.sot_sequence_including_notimestamps)
@@ -744,7 +745,7 @@ def run(rank, world_size, args):
     tokenizer = whisper.tokenizer.get_tokenizer(
         model.is_multilingual,
         num_languages=model.num_languages,
-        language="zh",
+        language="en",
         task="transcribe",
     )
 
@@ -800,7 +801,19 @@ def run(rank, world_size, args):
     if params.inf_check:
         register_inf_check_hooks(model)
 
-    aishell = AishellAsrDataModule(args)
+    librispeech = LibriSpeechAsrDataModule(args)
+
+    if params.full_libri:
+        train_cuts = librispeech.train_all_shuf_cuts()
+    else:
+        train_cuts = librispeech.train_clean_100_cuts()
+
+    def remove_short_and_long_utt(c: Cut):
+        if c.duration < 1.0 or c.duration > 20.0:
+            return False
+        return True
+
+    train_cuts = train_cuts.filter(remove_short_and_long_utt)
 
     if params.start_batch > 0 and checkpoints and "sampler" in checkpoints:
         # We only load the sampler's state dict when it loads a checkpoint
@@ -809,8 +822,16 @@ def run(rank, world_size, args):
     else:
         sampler_state_dict = None
 
-    train_dl = aishell.train_dataloaders(aishell.train_cuts())
-    valid_dl = aishell.valid_dataloaders(aishell.valid_cuts())
+    train_dl = librispeech.train_dataloaders(
+        train_cuts, sampler_state_dict=sampler_state_dict
+    )
+
+    valid_cuts = librispeech.dev_clean_cuts()
+    valid_cuts += librispeech.dev_other_cuts()
+
+    # do this to prevent Whisper throwing the length mismatch error
+    valid_cuts = valid_cuts.filter(remove_short_and_long_utt)
+    valid_dl = librispeech.valid_dataloaders(valid_cuts)
 
     scaler = GradScaler(enabled=params.use_fp16, init_scale=1.0)
     if checkpoints and "grad_scaler" in checkpoints:
@@ -911,7 +932,7 @@ def display_and_save_batch(
 
 def main():
     parser = get_parser()
-    AishellAsrDataModule.add_arguments(parser)
+    LibriSpeechAsrDataModule.add_arguments(parser)
     args = parser.parse_args()
     args.exp_dir = Path(args.exp_dir)