replaced all SingleCutSampler with SimpleCutSampler

k2-fsa · Sep 13, 2023 · cba1f6e · cba1f6e
1 parent b96fdeb
commit cba1f6e
Show file tree

Hide file tree

Showing 20 changed files with 58 additions and 60 deletions.
diff --git a/egs/aidatatang_200zh/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/aidatatang_200zh/ASR/pruned_transducer_stateless2/asr_datamodule.py
@@ -37,7 +37,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -291,8 +291,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/aishell/ASR/tdnn_lstm_ctc/asr_datamodule.py b/egs/aishell/ASR/tdnn_lstm_ctc/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -278,8 +278,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/aishell2/ASR/pruned_transducer_stateless5/asr_datamodule.py b/egs/aishell2/ASR/pruned_transducer_stateless5/asr_datamodule.py
@@ -31,7 +31,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 For AudioSamples
@@ -299,8 +299,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/aishell4/ASR/pruned_transducer_stateless5/asr_datamodule.py b/egs/aishell4/ASR/pruned_transducer_stateless5/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 for AudioSamples
@@ -310,8 +310,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/alimeeting/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/alimeeting/ASR/pruned_transducer_stateless2/asr_datamodule.py
@@ -37,7 +37,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -292,8 +292,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/ami/SURT/dprnn_zipformer/asr_datamodule.py b/egs/ami/SURT/dprnn_zipformer/asr_datamodule.py
@@ -257,7 +257,7 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
+            logging.info("Using SimpleCutSampler.")
             train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,

diff --git a/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py b/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 For AudioSamples
@@ -311,8 +311,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/csj/ASR/local/utils/asr_datamodule.py b/egs/csj/ASR/local/utils/asr_datamodule.py
@@ -31,7 +31,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 For AudioSamples
@@ -339,8 +339,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/gigaspeech/ASR/conformer_ctc/asr_datamodule.py b/egs/gigaspeech/ASR/conformer_ctc/asr_datamodule.py
@@ -27,7 +27,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -264,8 +264,8 @@ def train_dataloaders(self, cuts_train: CutSet) -> DataLoader:
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/gigaspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -297,8 +297,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/libricss/SURT/dprnn_zipformer/asr_datamodule.py b/egs/libricss/SURT/dprnn_zipformer/asr_datamodule.py
@@ -259,7 +259,7 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
+            logging.info("Using SimpleCutSampler.")
             train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,

diff --git a/egs/librispeech/ASR/pruned2_knowledge/asr_datamodule.py b/egs/librispeech/ASR/pruned2_knowledge/asr_datamodule.py
@@ -31,7 +31,7 @@
     CutMix,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -290,8 +290,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/librispeech/ASR/pruned_transducer_stateless7/gigaspeech.py b/egs/librispeech/ASR/pruned_transducer_stateless7/gigaspeech.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -297,8 +297,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/librispeech/ASR/tdnn_lstm_ctc/asr_datamodule.py b/egs/librispeech/ASR/tdnn_lstm_ctc/asr_datamodule.py
@@ -31,7 +31,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 For AudioSamples
@@ -314,8 +314,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/mgb2/ASR/conformer_ctc/asr_datamodule.py b/egs/mgb2/ASR/conformer_ctc/asr_datamodule.py
@@ -17,7 +17,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -270,8 +270,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/tal_csasr/ASR/pruned_transducer_stateless5/asr_datamodule.py b/egs/tal_csasr/ASR/pruned_transducer_stateless5/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import (  # noqa F401 for AudioSamples
@@ -311,8 +311,8 @@ def train_dataloaders(
                 drop_last=self.args.drop_last,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,

diff --git a/egs/tedlium3/ASR/transducer_stateless/asr_datamodule.py b/egs/tedlium3/ASR/transducer_stateless/asr_datamodule.py
@@ -28,7 +28,7 @@
     CutMix,
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -259,8 +259,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,
@@ -282,7 +282,6 @@ def train_dataloaders(
         return train_dl
 
     def valid_dataloaders(self, cuts_valid: CutSet) -> DataLoader:
-
         transforms = []
         if self.args.concatenate_cuts:
             transforms = [
@@ -322,7 +321,6 @@ def valid_dataloaders(self, cuts_valid: CutSet) -> DataLoader:
         return valid_dl
 
     def test_dataloaders(self, cuts_test: CutSet) -> DataLoader:
-
         logging.debug("About to create test dataset")
         if self.args.on_the_fly_feats:
             test = K2SpeechRecognitionDataset(

diff --git a/egs/timit/ASR/tdnn_lstm_ctc/asr_datamodule.py b/egs/timit/ASR/tdnn_lstm_ctc/asr_datamodule.py
@@ -30,7 +30,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -225,8 +225,8 @@ def train_dataloaders(self) -> DataLoader:
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,
@@ -267,7 +267,7 @@ def valid_dataloaders(self) -> DataLoader:
                 cut_transforms=transforms,
                 return_cuts=self.args.return_cuts,
             )
-        valid_sampler = SingleCutSampler(
+        valid_sampler = SimpleCutSampler(
             cuts_valid,
             max_duration=self.args.max_duration,
             shuffle=False,
@@ -298,7 +298,7 @@ def test_dataloaders(self) -> Union[DataLoader, List[DataLoader]]:
                 else PrecomputedFeatures(),
                 return_cuts=self.args.return_cuts,
             )
-            sampler = SingleCutSampler(cuts_test, max_duration=self.args.max_duration)
+            sampler = SimpleCutSampler(cuts_test, max_duration=self.args.max_duration)
             logging.debug("About to create test dataloader")
             test_dl = DataLoader(test, batch_size=None, sampler=sampler, num_workers=1)
             test_loaders.append(test_dl)

diff --git a/egs/wenetspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py b/egs/wenetspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py
@@ -37,7 +37,7 @@
     DynamicBucketingSampler,
     K2SpeechRecognitionDataset,
     PrecomputedFeatures,
-    SingleCutSampler,
+    SimpleCutSampler,
     SpecAugment,
 )
 from lhotse.dataset.input_strategies import OnTheFlyFeatures
@@ -296,8 +296,8 @@ def train_dataloaders(
                 drop_last=True,
             )
         else:
-            logging.info("Using SingleCutSampler.")
-            train_sampler = SingleCutSampler(
+            logging.info("Using SimpleCutSampler.")
+            train_sampler = SimpleCutSampler(
                 cuts_train,
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,