mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-09-05 15:14:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

revert asr data module

Browse Source
This commit is contained in:
Yuekai Zhang 2024-01-15 19:59:48 +08:00
parent e883bb60d4
commit eea46458c5
3 changed files with 6 additions and 1179 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
egs/aishell/ASR/tdnn_lstm_ctc/asr_datamodule.py
View File

@ -30,7 +30,7 @@ from lhotse.dataset import (
DynamicBucketingSampler, DynamicBucketingSampler,
K2SpeechRecognitionDataset, K2SpeechRecognitionDataset,
PrecomputedFeatures, PrecomputedFeatures,
SingleCutSampler, SimpleCutSampler,
SpecAugment, SpecAugment,
) )
from lhotse.dataset.input_strategies import OnTheFlyFeatures from lhotse.dataset.input_strategies import OnTheFlyFeatures
@ -198,7 +198,7 @@ class AishellAsrDataModule:
if self.args.enable_musan: if self.args.enable_musan:
logging.info("Enable MUSAN") logging.info("Enable MUSAN")
transforms.append( transforms.append(
CutMix(cuts=cuts_musan, prob=0.5, snr=(10, 20), preserve_id=True) CutMix(cuts=cuts_musan, p=0.5, snr=(10, 20), preserve_id=True)
) )
else: else:
logging.info("Disable MUSAN") logging.info("Disable MUSAN")
@ -278,8 +278,8 @@ class AishellAsrDataModule:
drop_last=self.args.drop_last, drop_last=self.args.drop_last,
) )
else: else:
logging.info("Using SingleCutSampler.") logging.info("Using SimpleCutSampler.")
train_sampler = SingleCutSampler( train_sampler = SimpleCutSampler(
cuts_train, cuts_train,
max_duration=self.args.max_duration, max_duration=self.args.max_duration,
shuffle=self.args.shuffle, shuffle=self.args.shuffle,

1173
egs/aishell/ASR/whisper/optim.py
View File

File diff suppressed because it is too large Load Diff

1
egs/aishell/ASR/whisper/optim.py Symbolic link
View File

@ -0,0 +1 @@
../../../librispeech/ASR/zipformer/optim.py

3
egs/aishell/ASR/whisper/requirements.txt Normal file → Executable file
View File

@ -1,6 +1,5 @@
k2 k2
kaldialign kaldialign
#lhotse==1.18
git+https://github.com/lhotse-speech/lhotse git+https://github.com/lhotse-speech/lhotse
sentencepiece sentencepiece
tensorboard tensorboard
@ -8,4 +7,4 @@ librosa
openai-whisper==20231117 openai-whisper==20231117
zhconv zhconv
WeTextProcessing WeTextProcessing
deepspeed deepspeed