Upload 3 files

README.md

@@ -1,42 +1,60 @@
 ---
-language: "en"
-thumbnail:
+language:
+- de
+thumbnail: null
 pipeline_tag: automatic-speech-recognition
 tags:
 - CTC
 - pytorch
 - speechbrain
 - Transformer
-license: "apache-2.0"
+license: apache-2.0
 datasets:
 - commonvoice
 metrics:
 - wer
 - cer
+model-index:
+- name: asr-wav2vec2-commonvoice-de
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: CommonVoice Corpus 10.0/ (German)
+      type: mozilla-foundation/common_voice_10_1
+      config: de
+      split: test
+      args:
+        language: de
+    metrics:
+    - name: Test WER
+      type: wer
+      value: '9.54'
 ---
 
 <iframe src="https://ghbtns.com/github-btn.html?user=speechbrain&repo=speechbrain&type=star&count=true&size=large&v=2" frameborder="0" scrolling="0" width="170" height="30" title="GitHub"></iframe>
 <br/><br/>
 
-# wav2vec 2.0 with CTC trained on CommonVoice English (No LM)
+# wav2vec 2.0 with CTC trained on CommonVoice German (No LM)
 
 This repository provides all the necessary tools to perform automatic speech
-recognition from an end-to-end system pretrained on CommonVoice (English Language) within
+recognition from an end-to-end system pretrained on CommonVoice (German Language) within
 SpeechBrain. For a better experience, we encourage you to learn more about
-[SpeechBrain](https://speechbrain.github.io). 
+[SpeechBrain](https://speechbrain.github.io).
 
 The performance of the model is the following:
 
 | Release | Test CER | Test WER | GPUs |
-|:--------------:|:--------------:|:--------------:| :--------:|
-| 15-08-23 | 7.92 | 16.68 | 1xV100 32GB |
+|:-------------:|:--------------:|:--------------:| :--------:|
+| 16-08-22 | 2.40 | 9.54 | 1xRTXA6000 48GB |
 
 ## Pipeline description
 
 This ASR system is composed of 2 different but linked blocks:
-- Tokenizer (unigram) that transforms words into subword units and trained with
-the train transcriptions (train.tsv) of CommonVoice (EN).
-- Acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model ([wav2vec2-lv60-large](https://huggingface.co/facebook/wav2vec2-large-lv60)) is combined with two DNN layers and finetuned on CommonVoice En. 
+- Tokenizer (char) that transforms words into chars and trained with
+the train transcriptions (train.tsv) of CommonVoice (DE).  
+- Acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model ([wav2vec2-large-xlsr-53-german](https://huggingface.co/facebook/wav2vec2-large-xlsr-53-german)) is combined with two DNN layers and finetuned on CommonVoice DE. 
 The obtained final acoustic representation is given to the CTC decoder.
 
 The system is trained with recordings sampled at 16kHz (single channel).
@@ -53,13 +71,13 @@ pip install speechbrain transformers
 Please notice that we encourage you to read our tutorials and learn more about
 [SpeechBrain](https://speechbrain.github.io).
 
-### Transcribing your own audio files (in English)
+### Transcribing your own audio files (in German)
 
 ```python
-from speechbrain.pretrained import EncoderDecoderASR
+from speechbrain.pretrained import EncoderASR
 
-asr_model = EncoderDecoderASR.from_hparams(source="speechbrain/asr-wav2vec2-commonvoice-14-en", savedir="pretrained_models/asr-wav2vec2-commonvoice-14-en")
-asr_model.transcribe_file("speechbrain/asr-wav2vec2-commonvoice-14-en/example.wav")
+asr_model = EncoderASR.from_hparams(source="speechbrain/asr-wav2vec2-commonvoice-de", savedir="pretrained_models/asr-wav2vec2-commonvoice-de")
+asr_model.transcribe_file("speechbrain/asr-wav2vec2-commonvoice-de/example-de.wav")
 
 ```
 ### Inference on GPU
@@ -85,10 +103,10 @@ pip install -e .
 3. Run Training:
 ```bash
 cd recipes/CommonVoice/ASR/seq2seq
-python train.py hparams/train_en_with_wav2vec.yaml --data_folder=your_data_folder
+python train.py hparams/train_de_with_wav2vec.yaml --data_folder=your_data_folder
 ```
 
-You can find our training results (models, logs, etc) [here](https://www.dropbox.com/sh/ch10cnbhf1faz3w/AACdHFG65LC6582H0Tet_glTa?dl=0).
+You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/19G2Zm8896QSVDqVfs7PS_W86-K0-5xeC?usp=sharing).
 
 ### Limitations
 The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.
@@ -113,4 +131,4 @@ Please, cite SpeechBrain if you use it for your research or business.
   primaryClass={eess.AS},
   note={arXiv:2106.04624}
 }
-```
+```

config.json

@@ -1,69 +1,69 @@
 {
-  "speechbrain_interface": "EncoderDecoderASR",
-  "activation_dropout": 0.1,
-  "apply_spec_augment": true,
-  "architectures": [
-    "Wav2Vec2Model"
-  ],
-  "attention_dropout": 0.1,
-  "bos_token_id": 1,
-  "conv_bias": true,
-  "conv_dim": [
-    512,
-    512,
-    512,
-    512,
-    512,
-    512,
-    512
-  ],
-  "conv_kernel": [
-    10,
-    3,
-    3,
-    3,
-    3,
-    2,
-    2
-  ],
-  "conv_stride": [
-    5,
-    2,
-    2,
-    2,
-    2,
-    2,
-    2
-  ],
-  "ctc_loss_reduction": "sum",
-  "ctc_zero_infinity": false,
-  "do_stable_layer_norm": true,
-  "eos_token_id": 2,
-  "feat_extract_activation": "gelu",
-  "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.1,
-  "final_dropout": 0.1,
-  "gradient_checkpointing": false,
-  "hidden_act": "gelu",
-  "hidden_dropout": 0.1,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 1024,
-  "initializer_range": 0.02,
-  "intermediate_size": 4096,
-  "layer_norm_eps": 1e-05,
-  "layerdrop": 0.1,
-  "mask_feature_length": 10,
-  "mask_feature_prob": 0.0,
-  "mask_time_length": 10,
-  "mask_time_prob": 0.05,
-  "model_type": "wav2vec2",
-  "num_attention_heads": 16,
-  "num_conv_pos_embedding_groups": 16,
-  "num_conv_pos_embeddings": 128,
-  "num_feat_extract_layers": 7,
-  "num_hidden_layers": 24,
-  "pad_token_id": 0,
-  "transformers_version": "4.4.0.dev0",
-  "vocab_size": 32
+    "speechbrain_interface": "EncoderASR",
+    "activation_dropout": 0.1,
+    "apply_spec_augment": true,
+    "architectures": [
+        "Wav2Vec2Model"
+    ],
+    "attention_dropout": 0.1,
+    "bos_token_id": 1,
+    "conv_bias": true,
+    "conv_dim": [
+        512,
+        512,
+        512,
+        512,
+        512,
+        512,
+        512
+    ],
+    "conv_kernel": [
+        10,
+        3,
+        3,
+        3,
+        3,
+        2,
+        2
+    ],
+    "conv_stride": [
+        5,
+        2,
+        2,
+        2,
+        2,
+        2,
+        2
+    ],
+    "ctc_loss_reduction": "sum",
+    "ctc_zero_infinity": false,
+    "do_stable_layer_norm": true,
+    "eos_token_id": 2,
+    "feat_extract_activation": "gelu",
+    "feat_extract_dropout": 0.0,
+    "feat_extract_norm": "layer",
+    "feat_proj_dropout": 0.1,
+    "final_dropout": 0.1,
+    "gradient_checkpointing": false,
+    "hidden_act": "gelu",
+    "hidden_dropout": 0.1,
+    "hidden_dropout_prob": 0.1,
+    "hidden_size": 1024,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "layer_norm_eps": 1e-05,
+    "layerdrop": 0.1,
+    "mask_feature_length": 10,
+    "mask_feature_prob": 0.0,
+    "mask_time_length": 10,
+    "mask_time_prob": 0.05,
+    "model_type": "wav2vec2",
+    "num_attention_heads": 16,
+    "num_conv_pos_embedding_groups": 16,
+    "num_conv_pos_embeddings": 128,
+    "num_feat_extract_layers": 7,
+    "num_hidden_layers": 24,
+    "pad_token_id": 0,
+    "transformers_version": "4.21.1",
+    "vocab_size": 32
 }

hyperparams.yaml

@@ -1,138 +1,91 @@
+# Generated 2022-08-12 from:
+# /netscratch/sagar/thesis/speechbrain/recipes/CommonVoice_de/ASR/CTC/hparams/train_with_wav2vec.yaml
+# yamllint disable
 # ################################
-# Model: wav2vec2 + DNN + CTC/Attention
+# Model: wav2vec2 + DNN + CTC
 # Augmentation: SpecAugment
-# Authors: Titouan Parcollet 2021
+# Authors: Sung-Lin Yeh 2021
 # ################################
 
-sample_rate: 16000
-wav2vec2_hub: facebook/wav2vec2-large-lv60
-
 # BPE parameters
-token_type: unigram  # ["unigram", "bpe", "char"]
+token_type: char  # ["unigram", "bpe", "char"]
 character_coverage: 1.0
 
 # Model parameters
-activation: !name:torch.nn.LeakyReLU
-dnn_layers: 2
+# activation: !name:torch.nn.LeakyReLU
 dnn_neurons: 1024
-emb_size: 128
-dec_neurons: 1024
+wav2vec_output_dim: 1024
+dropout: 0.15
+
+sample_rate: 16000
+
+wav2vec2_hub: facebook/wav2vec2-large-xlsr-53-german
 
 # Outputs
-output_neurons: 1000  # BPE size, index(blank/eos/bos) = 0
+output_neurons: 32  # BPE size, index(blank/eos/bos) = 0
 
 # Decoding parameters
 # Be sure that the bos and eos index match with the BPEs ones
 blank_index: 0
 bos_index: 1
 eos_index: 2
-min_decode_ratio: 0.0
-max_decode_ratio: 1.0
-beam_size: 10
-eos_threshold: 1.5
-using_max_attn_shift: True
-max_attn_shift: 140
-ctc_weight_decode: 0.0
-temperature: 1.50
-
-# enc: !new:speechbrain.lobes.models.VanillaNN.VanillaNN
-#     input_shape: [null, null, 1024]
-#     activation: !ref <activation>
-#     dnn_blocks: !ref <dnn_layers>
-#     dnn_neurons: !ref <dnn_neurons>
 
 enc: !new:speechbrain.nnet.containers.Sequential
-    input_shape: [null, null, 1024]
-    linear1: !name:speechbrain.nnet.linear.Linear
-        n_neurons: !ref <dnn_neurons>
-        bias: True
-    bn1: !name:speechbrain.nnet.normalization.BatchNorm1d
-    activation: !new:torch.nn.LeakyReLU
-
-    linear2: !name:speechbrain.nnet.linear.Linear
-        n_neurons: !ref <dnn_neurons>
-        bias: True
-    bn2: !name:speechbrain.nnet.normalization.BatchNorm1d
-    activation2: !new:torch.nn.LeakyReLU
-
-    linear3: !name:speechbrain.nnet.linear.Linear
-        n_neurons: !ref <dnn_neurons>
-        bias: True
-    bn3: !name:speechbrain.nnet.normalization.BatchNorm1d
-    activation3: !new:torch.nn.LeakyReLU
+  input_shape: [null, null, !ref <wav2vec_output_dim>]
+  linear1: !name:speechbrain.nnet.linear.Linear
+    n_neurons: !ref <dnn_neurons>
+    bias: True
+  bn1: !name:speechbrain.nnet.normalization.BatchNorm1d
+  activation: !new:torch.nn.LeakyReLU
+  drop: !new:torch.nn.Dropout
+    p: !ref <dropout>
+  linear2: !name:speechbrain.nnet.linear.Linear
+    n_neurons: !ref <dnn_neurons>
+    bias: True
+  bn2: !name:speechbrain.nnet.normalization.BatchNorm1d
+  activation2: !new:torch.nn.LeakyReLU
+  drop2: !new:torch.nn.Dropout
+    p: !ref <dropout>
+  linear3: !name:speechbrain.nnet.linear.Linear
+    n_neurons: !ref <dnn_neurons>
+    bias: True
+  bn3: !name:speechbrain.nnet.normalization.BatchNorm1d
+  activation3: !new:torch.nn.LeakyReLU
 
 wav2vec2: !new:speechbrain.lobes.models.huggingface_wav2vec.HuggingFaceWav2Vec2
-    source: !ref <wav2vec2_hub>
-    output_norm: True
-    freeze: True
-    save_path: model_checkpoints
-
-emb: !new:speechbrain.nnet.embedding.Embedding
-    num_embeddings: !ref <output_neurons>
-    embedding_dim: !ref <emb_size>
-
-dec: !new:speechbrain.nnet.RNN.AttentionalRNNDecoder
-    enc_dim: !ref <dnn_neurons>
-    input_size: !ref <emb_size>
-    rnn_type: gru
-    attn_type: location
-    hidden_size: 1024
-    attn_dim: 1024
-    num_layers: 1
-    scaling: 1.0
-    channels: 10
-    kernel_size: 100
-    re_init: True
-    dropout: 0.0
+  source: !ref <wav2vec2_hub>
+  output_norm: True
+  freeze: True
+  save_path: wav2vec2_checkpoint
 
 ctc_lin: !new:speechbrain.nnet.linear.Linear
-    input_size: !ref <dnn_neurons>
-    n_neurons: !ref <output_neurons>
-
-seq_lin: !new:speechbrain.nnet.linear.Linear
-    input_size: !ref <dec_neurons>
-    n_neurons: !ref <output_neurons>
+  input_size: !ref <dnn_neurons>
+  n_neurons: !ref <output_neurons>
 
 log_softmax: !new:speechbrain.nnet.activations.Softmax
-    apply_log: True
+  apply_log: True
 
 ctc_cost: !name:speechbrain.nnet.losses.ctc_loss
-    blank_index: !ref <blank_index>
-
-seq_cost: !name:speechbrain.nnet.losses.nll_loss
-    label_smoothing: 0.1
+  blank_index: !ref <blank_index>
 
 asr_model: !new:torch.nn.ModuleList
-    - [!ref <enc>, !ref <emb>, !ref <dec>, !ref <ctc_lin>, !ref <seq_lin>]
+    - [!ref <enc>, !ref <ctc_lin>]
 
 tokenizer: !new:sentencepiece.SentencePieceProcessor
 
 encoder: !new:speechbrain.nnet.containers.LengthsCapableSequential
     wav2vec2: !ref <wav2vec2>
     enc: !ref <enc>
-
-decoder: !new:speechbrain.decoders.S2SRNNBeamSearcher
-    embedding: !ref <emb>
-    decoder: !ref <dec>
-    linear: !ref <seq_lin>
-    ctc_linear: !ref <ctc_lin>
-    bos_index: !ref <bos_index>
-    eos_index: !ref <eos_index>
-    blank_index: !ref <blank_index>
-    min_decode_ratio: !ref <min_decode_ratio>
-    max_decode_ratio: !ref <max_decode_ratio>
-    beam_size: !ref <beam_size>
-    eos_threshold: !ref <eos_threshold>
-    using_max_attn_shift: !ref <using_max_attn_shift>
-    max_attn_shift: !ref <max_attn_shift>
-    temperature: !ref <temperature>
+    ctc_lin: !ref <ctc_lin>
 
 modules:
-    encoder: !ref <encoder>
-    decoder: !ref <decoder>
+  encoder: !ref <encoder>
+
+decoding_function: !name:speechbrain.decoders.ctc_greedy_decode
+    blank_id: !ref <blank_index>
 
 pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
-    loadables:
-        wav2vec2: !ref <wav2vec2>
-        asr: !ref <asr_model>
-        tokenizer: !ref <tokenizer>
+   loadables:
+      wav2vec2: !ref <wav2vec2>
+      asr: !ref <asr_model>
+      tokenizer: !ref <tokenizer>