[ASR] wav2vec2_en, test=asr (#2637)

* wav2vec2_en, test=asr

* wav2vec2_en, test=asr

* wav2vec2_en, test=asr

docs/source/released_model.md

@@ -22,7 +22,7 @@ Acoustic Model | Training Data | Token-based | Size | Descriptions | CER | WER |
 Model | Pre-Train Method | Pre-Train Data | Finetune Data | Size | Descriptions | CER | WER |  Example Link |
 :-------------:| :------------:| :-----: | -----: | :-----: |:-----:| :-----:  | :-----:  | :-----: | 
 [Wav2vec2-large-960h-lv60-self Model](https://paddlespeech.bj.bcebos.com/wav2vec/wav2vec2-large-960h-lv60-self.pdparams) | wav2vec2 | Librispeech and LV-60k Dataset (5.3w h) | - | 1.18 GB |Pre-trained Wav2vec2.0 Model | - | - | - | 
-[Wav2vec2ASR-large-960h-librispeech Model](https://paddlespeech.bj.bcebos.com/s2t/librispeech/asr3/wav2vec2ASR-large-960h-librispeech_ckpt_1.3.0.model.tar.gz) | wav2vec2 | Librispeech and LV-60k Dataset (5.3w h) | Librispeech (960 h) | 1.18 GB |Encoder: Wav2vec2.0, Decoder: CTC, Decoding method: Greedy search | - | 0.0189 | [Wav2vecASR Librispeech ASR3](../../examples/librispeech/asr3) |
+[Wav2vec2ASR-large-960h-librispeech Model](https://paddlespeech.bj.bcebos.com/s2t/librispeech/asr3/wav2vec2ASR-large-960h-librispeech_ckpt_1.3.1.model.tar.gz) | wav2vec2 | Librispeech and LV-60k Dataset (5.3w h) | Librispeech (960 h) | 718 MB |Encoder: Wav2vec2.0, Decoder: CTC, Decoding method: Greedy search | - | 0.0189 | [Wav2vecASR Librispeech ASR3](../../examples/librispeech/asr3) |
 
 ### Language Model based on NGram
 Language Model | Training Data | Token-based | Size | Descriptions

examples/librispeech/asr3/conf/wav2vec2ASR.yaml

@@ -70,7 +70,6 @@ train_manifest: data/manifest.train
 dev_manifest: data/manifest.dev
 test_manifest: data/manifest.test-clean
 
-
 ###########################################
 #              Dataloader                 #
 ###########################################
@@ -95,6 +94,12 @@ dist_sampler: True
 shortest_first: True
 return_lens_rate: True
   
+############################################
+#             Data Augmentation            #
+############################################
+audio_augment:  # for raw audio 
+  sample_rate: 16000
+  speeds: [95, 100, 105]
 
 ###########################################
 #                 Training                #
@@ -115,6 +120,3 @@ log_interval: 1
 checkpoint:
   kbest_n: 50
   latest_n: 5
-augment: True
-
-

paddlespeech/s2t/exps/wav2vec2/model.py

@@ -71,7 +71,8 @@ class Wav2Vec2ASRTrainer(Trainer):
         wavs_lens_rate = wavs_lens / wav.shape[1]
         target_lens_rate = target_lens / target.shape[1]
         wav = wav[:, :, 0]
-        wav = self.speech_augmentation(wav, wavs_lens_rate)
+        if hasattr(train_conf, 'speech_augment'):
+            wav = self.speech_augmentation(wav, wavs_lens_rate)
         loss = self.model(wav, wavs_lens_rate, target, target_lens_rate)
         # loss div by `batch_size * accum_grad`
         loss /= train_conf.accum_grad
@@ -277,7 +278,9 @@ class Wav2Vec2ASRTrainer(Trainer):
         logger.info("Setup model!")
 
         # setup speech augmentation for wav2vec2
-        self.speech_augmentation = TimeDomainSpecAugment()
+        if hasattr(config, 'audio_augment') and self.train:
+            self.speech_augmentation = TimeDomainSpecAugment(
+                **config.audio_augment)
 
         if not self.train:
             return

paddlespeech/s2t/models/wav2vec2/processing/speech_augmentation.py

@@ -641,14 +641,11 @@ class DropChunk(nn.Layer):
 
 class TimeDomainSpecAugment(nn.Layer):
     """A time-domain approximation of the SpecAugment algorithm.
-
     This augmentation module implements three augmentations in
     the time-domain.
-
      1. Drop chunks of the audio (zero amplitude or white noise)
      2. Drop frequency bands (with band-drop filters)
      3. Speed peturbation (via resampling to slightly different rate)
-
     Arguments
     ---------
     perturb_prob : float from 0 to 1
@@ -677,7 +674,6 @@ class TimeDomainSpecAugment(nn.Layer):
     drop_chunk_noise_factor : float
         The noise factor used to scale the white noise inserted, relative to
         the average amplitude of the utterance. Default 0 (no noise inserted).
-
     Example
     -------
     >>> inputs = paddle.randn([10, 16000])
@@ -718,7 +714,6 @@ class TimeDomainSpecAugment(nn.Layer):
 
     def forward(self, waveforms, lengths):
         """Returns the distorted waveforms.
-
         Arguments
         ---------
         waveforms : tensor