Titouan commited on
Commit
42dba6c
1 Parent(s): f9c7e11

update model

Browse files
Files changed (17) hide show
  1. README.md +8 -9
  2. asr.ckpt +2 -2
  3. hyperparams.yaml +26 -48
  4. model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9 +75 -0
  5. model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.json +1 -0
  6. model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.lock +0 -0
  7. model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9 +75 -0
  8. model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.json +1 -0
  9. model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.lock +0 -0
  10. model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23 +8 -0
  11. model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.json +1 -0
  12. model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.lock +0 -0
  13. model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23 +8 -0
  14. model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.json +1 -0
  15. model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.lock +0 -0
  16. tokenizer.ckpt +0 -0
  17. wav2vec2.ckpt +1 -1
README.md CHANGED
@@ -4,7 +4,6 @@ thumbnail:
4
  tags:
5
  - automatic-speech-recognition
6
  - CTC
7
- - Attention
8
  - pytorch
9
  - speechbrain
10
  - Transformer
@@ -24,21 +23,21 @@ metrics:
24
  This repository provides all the necessary tools to perform automatic speech
25
  recognition from an end-to-end system pretrained on CommonVoice (French Language) within
26
  SpeechBrain. For a better experience, we encourage you to learn more about
27
- [SpeechBrain](https://speechbrain.github.io).
28
 
29
  The performance of the model is the following:
30
 
31
  | Release | Test CER | Test WER | GPUs |
32
  |:-------------:|:--------------:|:--------------:| :--------:|
33
- | 29-04-21 | 9.78 | 13.34 | 2xV100 32GB |
34
 
35
  ## Pipeline description
36
 
37
  This ASR system is composed of 2 different but linked blocks:
38
  - Tokenizer (unigram) that transforms words into subword units and trained with
39
  the train transcriptions (train.tsv) of CommonVoice (FR).
40
- - Acoustic model (wav2vec2.0 + CTC/Attention). A pretrained wav2vec 2.0 model ([LeBenchmark/wav2vec2-FR-M-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-M-large)) is combined with two DNN layers and finetuned on CommonVoice FR.
41
- The obtained final acoustic representation is given to the CTC and attention decoders.
42
 
43
 
44
  ## Install SpeechBrain
@@ -55,9 +54,9 @@ Please notice that we encourage you to read our tutorials and learn more about
55
  ### Transcribing your own audio files (in French)
56
 
57
  ```python
58
- from speechbrain.pretrained import EncoderDecoderASR
59
 
60
- asr_model = EncoderDecoderASR.from_hparams(source="speechbrain/asr-wav2vec2-commonvoice-fr", savedir="pretrained_models/asr-crdnn-commonvoice-fr")
61
  asr_model.transcribe_file("example-fr.wav")
62
 
63
  ```
@@ -80,11 +79,11 @@ pip install -e .
80
 
81
  3. Run Training:
82
  ```bash
83
- cd recipes/CommonVoice/ASR/seq2seq
84
  python train_with_wav2vec.py hparams/train_fr_with_wav2vec.yaml --data_folder=your_data_folder
85
  ```
86
 
87
- You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1tjz6IZmVRkuRE97E7h1cXFoGTer7pT73?usp=sharing).
88
 
89
  ### Limitations
90
  The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.
 
4
  tags:
5
  - automatic-speech-recognition
6
  - CTC
 
7
  - pytorch
8
  - speechbrain
9
  - Transformer
 
23
  This repository provides all the necessary tools to perform automatic speech
24
  recognition from an end-to-end system pretrained on CommonVoice (French Language) within
25
  SpeechBrain. For a better experience, we encourage you to learn more about
26
+ [SpeechBrain](https://speechbrain.github.io).
27
 
28
  The performance of the model is the following:
29
 
30
  | Release | Test CER | Test WER | GPUs |
31
  |:-------------:|:--------------:|:--------------:| :--------:|
32
+ | 24-08-21 | 3.19 | 9.96 | 2xV100 32GB |
33
 
34
  ## Pipeline description
35
 
36
  This ASR system is composed of 2 different but linked blocks:
37
  - Tokenizer (unigram) that transforms words into subword units and trained with
38
  the train transcriptions (train.tsv) of CommonVoice (FR).
39
+ - Acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model ([LeBenchmark/wav2vec2-FR-7K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large)) is combined with two DNN layers and finetuned on CommonVoice FR.
40
+ The obtained final acoustic representation is given to the CTC greedy decoder.
41
 
42
 
43
  ## Install SpeechBrain
 
54
  ### Transcribing your own audio files (in French)
55
 
56
  ```python
57
+ from speechbrain.pretrained import EncoderASR
58
 
59
+ asr_model = EncoderASR.from_hparams(source="speechbrain/asr-wav2vec2-commonvoice-fr", savedir="pretrained_models/asr-wav2vec2-commonvoice-fr")
60
  asr_model.transcribe_file("example-fr.wav")
61
 
62
  ```
 
79
 
80
  3. Run Training:
81
  ```bash
82
+ cd recipes/CommonVoice/ASR/CTC/
83
  python train_with_wav2vec.py hparams/train_fr_with_wav2vec.yaml --data_folder=your_data_folder
84
  ```
85
 
86
+ You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1T9DfdZwcNI9CURxhLCi8GA5JVz8adiY8?usp=sharing).
87
 
88
  ### Limitations
89
  The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.
asr.ckpt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8f9b32cfe8a7d10fa852874b5507661d95cdb8c9c8dd9add45976e786e08c52e
3
- size 60570064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64ba475ed7be735d4ac054c2d537f22251b80f6ecb65cb04217eb0d1ed50a143
3
+ size 12963902
hyperparams.yaml CHANGED
@@ -5,7 +5,7 @@
5
  # ################################
6
 
7
  sample_rate: 16000
8
- wav2vec2_hub: LeBenchmark/wav2vec2-FR-M-large
9
 
10
  # BPE parameters
11
  token_type: unigram # ["unigram", "bpe", "char"]
@@ -19,7 +19,7 @@ emb_size: 128
19
  dec_neurons: 1024
20
 
21
  # Outputs
22
- output_neurons: 500 # BPE size, index(blank/eos/bos) = 0
23
 
24
  # Decoding parameters
25
  # Be sure that the bos and eos index match with the BPEs ones
@@ -35,11 +35,27 @@ max_attn_shift: 140
35
  ctc_weight_decode: 0.0
36
  temperature: 1.50
37
 
38
- enc: !new:speechbrain.lobes.models.VanillaNN.VanillaNN
39
  input_shape: [null, null, 1024]
40
- activation: !ref <activation>
41
- dnn_blocks: !ref <dnn_layers>
42
- dnn_neurons: !ref <dnn_neurons>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
43
 
44
  wav2vec2: !new:speechbrain.lobes.models.huggingface_wav2vec.HuggingFaceWav2Vec2
45
  source: !ref <wav2vec2_hub>
@@ -48,69 +64,31 @@ wav2vec2: !new:speechbrain.lobes.models.huggingface_wav2vec.HuggingFaceWav2Vec2
48
  pretrain: False
49
  save_path: model_checkpoints
50
 
51
- emb: !new:speechbrain.nnet.embedding.Embedding
52
- num_embeddings: !ref <output_neurons>
53
- embedding_dim: !ref <emb_size>
54
-
55
- dec: !new:speechbrain.nnet.RNN.AttentionalRNNDecoder
56
- enc_dim: !ref <dnn_neurons>
57
- input_size: !ref <emb_size>
58
- rnn_type: gru
59
- attn_type: location
60
- hidden_size: 1024
61
- attn_dim: 1024
62
- num_layers: 1
63
- scaling: 1.0
64
- channels: 10
65
- kernel_size: 100
66
- re_init: True
67
- dropout: 0.15
68
-
69
  ctc_lin: !new:speechbrain.nnet.linear.Linear
70
  input_size: !ref <dnn_neurons>
71
  n_neurons: !ref <output_neurons>
72
 
73
- seq_lin: !new:speechbrain.nnet.linear.Linear
74
- input_size: !ref <dec_neurons>
75
- n_neurons: !ref <output_neurons>
76
-
77
  log_softmax: !new:speechbrain.nnet.activations.Softmax
78
  apply_log: True
79
 
80
  ctc_cost: !name:speechbrain.nnet.losses.ctc_loss
81
  blank_index: !ref <blank_index>
82
 
83
- seq_cost: !name:speechbrain.nnet.losses.nll_loss
84
- label_smoothing: 0.1
85
-
86
  asr_model: !new:torch.nn.ModuleList
87
- - [!ref <enc>, !ref <emb>, !ref <dec>, !ref <ctc_lin>, !ref <seq_lin>]
88
 
89
  tokenizer: !new:sentencepiece.SentencePieceProcessor
90
 
91
  encoder: !new:speechbrain.nnet.containers.LengthsCapableSequential
92
  wav2vec2: !ref <wav2vec2>
93
  enc: !ref <enc>
 
94
 
95
- decoder: !new:speechbrain.decoders.S2SRNNBeamSearcher
96
- embedding: !ref <emb>
97
- decoder: !ref <dec>
98
- linear: !ref <seq_lin>
99
- ctc_linear: !ref <ctc_lin>
100
- bos_index: !ref <bos_index>
101
- eos_index: !ref <eos_index>
102
- blank_index: !ref <blank_index>
103
- min_decode_ratio: !ref <min_decode_ratio>
104
- max_decode_ratio: !ref <max_decode_ratio>
105
- beam_size: !ref <beam_size>
106
- eos_threshold: !ref <eos_threshold>
107
- using_max_attn_shift: !ref <using_max_attn_shift>
108
- max_attn_shift: !ref <max_attn_shift>
109
- temperature: !ref <temperature>
110
 
111
  modules:
112
  encoder: !ref <encoder>
113
- decoder: !ref <decoder>
114
 
115
  pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
116
  loadables:
 
5
  # ################################
6
 
7
  sample_rate: 16000
8
+ wav2vec2_hub: LeBenchmark/wav2vec2-FR-7K-large
9
 
10
  # BPE parameters
11
  token_type: unigram # ["unigram", "bpe", "char"]
 
19
  dec_neurons: 1024
20
 
21
  # Outputs
22
+ output_neurons: 76 # BPE size, index(blank/eos/bos) = 0
23
 
24
  # Decoding parameters
25
  # Be sure that the bos and eos index match with the BPEs ones
 
35
  ctc_weight_decode: 0.0
36
  temperature: 1.50
37
 
38
+ enc: !new:speechbrain.nnet.containers.Sequential
39
  input_shape: [null, null, 1024]
40
+ linear1: !name:speechbrain.nnet.linear.Linear
41
+ n_neurons: 1024
42
+ bias: True
43
+ bn1: !name:speechbrain.nnet.normalization.BatchNorm1d
44
+ activation: !new:torch.nn.LeakyReLU
45
+ drop: !new:torch.nn.Dropout
46
+ p: 0.15
47
+ linear2: !name:speechbrain.nnet.linear.Linear
48
+ n_neurons: 1024
49
+ bias: True
50
+ bn2: !name:speechbrain.nnet.normalization.BatchNorm1d
51
+ activation2: !new:torch.nn.LeakyReLU
52
+ drop2: !new:torch.nn.Dropout
53
+ p: 0.15
54
+ linear3: !name:speechbrain.nnet.linear.Linear
55
+ n_neurons: 1024
56
+ bias: True
57
+ bn3: !name:speechbrain.nnet.normalization.BatchNorm1d
58
+ activation3: !new:torch.nn.LeakyReLU
59
 
60
  wav2vec2: !new:speechbrain.lobes.models.huggingface_wav2vec.HuggingFaceWav2Vec2
61
  source: !ref <wav2vec2_hub>
 
64
  pretrain: False
65
  save_path: model_checkpoints
66
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
67
  ctc_lin: !new:speechbrain.nnet.linear.Linear
68
  input_size: !ref <dnn_neurons>
69
  n_neurons: !ref <output_neurons>
70
 
 
 
 
 
71
  log_softmax: !new:speechbrain.nnet.activations.Softmax
72
  apply_log: True
73
 
74
  ctc_cost: !name:speechbrain.nnet.losses.ctc_loss
75
  blank_index: !ref <blank_index>
76
 
 
 
 
77
  asr_model: !new:torch.nn.ModuleList
78
+ - [!ref <enc>, !ref <ctc_lin>]
79
 
80
  tokenizer: !new:sentencepiece.SentencePieceProcessor
81
 
82
  encoder: !new:speechbrain.nnet.containers.LengthsCapableSequential
83
  wav2vec2: !ref <wav2vec2>
84
  enc: !ref <enc>
85
+ ctc_lin: !ref <ctc_lin>
86
 
87
+ decoding_function: !name:speechbrain.decoders.ctc_greedy_decode
88
+ blank_id: !ref <blank_index>
 
 
 
 
 
 
 
 
 
 
 
 
 
89
 
90
  modules:
91
  encoder: !ref <encoder>
 
92
 
93
  pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
94
  loadables:
model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9 ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "apply_spec_augment": true,
4
+ "architectures": [
5
+ "Wav2Vec2Model"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "bos_token_id": 1,
9
+ "conv_bias": true,
10
+ "conv_dim": [
11
+ 512,
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512
18
+ ],
19
+ "conv_kernel": [
20
+ 10,
21
+ 3,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 2,
26
+ 2
27
+ ],
28
+ "conv_stride": [
29
+ 5,
30
+ 2,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2
36
+ ],
37
+ "ctc_loss_reduction": "sum",
38
+ "ctc_zero_infinity": false,
39
+ "do_stable_layer_norm": true,
40
+ "eos_token_id": 2,
41
+ "feat_extract_activation": "gelu",
42
+ "feat_extract_dropout": 0.0,
43
+ "feat_extract_norm": "layer",
44
+ "feat_proj_dropout": 0.1,
45
+ "final_dropout": 0.0,
46
+ "gradient_checkpointing": false,
47
+ "hidden_act": "gelu",
48
+ "hidden_dropout": 0.1,
49
+ "hidden_size": 1024,
50
+ "initializer_range": 0.02,
51
+ "intermediate_size": 4096,
52
+ "layer_norm_eps": 1e-05,
53
+ "layerdrop": 0.1,
54
+ "mask_channel_length": 10,
55
+ "mask_channel_min_space": 1,
56
+ "mask_channel_other": 0.0,
57
+ "mask_channel_prob": 0.0,
58
+ "mask_channel_selection": "static",
59
+ "mask_feature_length": 10,
60
+ "mask_feature_prob": 0.0,
61
+ "mask_time_length": 10,
62
+ "mask_time_min_space": 1,
63
+ "mask_time_other": 0.0,
64
+ "mask_time_prob": 0.075,
65
+ "mask_time_selection": "static",
66
+ "model_type": "wav2vec2",
67
+ "num_attention_heads": 16,
68
+ "num_conv_pos_embedding_groups": 16,
69
+ "num_conv_pos_embeddings": 128,
70
+ "num_feat_extract_layers": 7,
71
+ "num_hidden_layers": 24,
72
+ "pad_token_id": 0,
73
+ "transformers_version": "4.5.1",
74
+ "vocab_size": 32
75
+ }
model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"url": "https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-large/resolve/main/config.json", "etag": "\"5565ad893213f0e049dcfd8a397c20224e7e26b9\""}
model_checkpoints/1ad9ee8b47c176b02563689d28740ad00aa9941ddb9d28ab323af9e5ffc9b5dc.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.lock ADDED
File without changes
model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9 ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "apply_spec_augment": true,
4
+ "architectures": [
5
+ "Wav2Vec2Model"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "bos_token_id": 1,
9
+ "conv_bias": true,
10
+ "conv_dim": [
11
+ 512,
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512
18
+ ],
19
+ "conv_kernel": [
20
+ 10,
21
+ 3,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 2,
26
+ 2
27
+ ],
28
+ "conv_stride": [
29
+ 5,
30
+ 2,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2
36
+ ],
37
+ "ctc_loss_reduction": "sum",
38
+ "ctc_zero_infinity": false,
39
+ "do_stable_layer_norm": true,
40
+ "eos_token_id": 2,
41
+ "feat_extract_activation": "gelu",
42
+ "feat_extract_dropout": 0.0,
43
+ "feat_extract_norm": "layer",
44
+ "feat_proj_dropout": 0.1,
45
+ "final_dropout": 0.0,
46
+ "gradient_checkpointing": false,
47
+ "hidden_act": "gelu",
48
+ "hidden_dropout": 0.1,
49
+ "hidden_size": 1024,
50
+ "initializer_range": 0.02,
51
+ "intermediate_size": 4096,
52
+ "layer_norm_eps": 1e-05,
53
+ "layerdrop": 0.1,
54
+ "mask_channel_length": 10,
55
+ "mask_channel_min_space": 1,
56
+ "mask_channel_other": 0.0,
57
+ "mask_channel_prob": 0.0,
58
+ "mask_channel_selection": "static",
59
+ "mask_feature_length": 10,
60
+ "mask_feature_prob": 0.0,
61
+ "mask_time_length": 10,
62
+ "mask_time_min_space": 1,
63
+ "mask_time_other": 0.0,
64
+ "mask_time_prob": 0.075,
65
+ "mask_time_selection": "static",
66
+ "model_type": "wav2vec2",
67
+ "num_attention_heads": 16,
68
+ "num_conv_pos_embedding_groups": 16,
69
+ "num_conv_pos_embeddings": 128,
70
+ "num_feat_extract_layers": 7,
71
+ "num_hidden_layers": 24,
72
+ "pad_token_id": 0,
73
+ "transformers_version": "4.5.1",
74
+ "vocab_size": 32
75
+ }
model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"url": "https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large/resolve/main/config.json", "etag": "\"5565ad893213f0e049dcfd8a397c20224e7e26b9\""}
model_checkpoints/47c38b5cad9a39b412be044270cd24897dcb7586ee61a0d6c0ce6ca9f4a3eff6.3aa7c2002067dfc71f74e269e463f76f247952b1abebe4841d03c98c534483b9.lock ADDED
File without changes
model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23 ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_size": 1,
4
+ "padding_side": "right",
5
+ "padding_value": 0.0,
6
+ "return_attention_mask": true,
7
+ "sampling_rate": 16000
8
+ }
model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"url": "https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large/resolve/main/preprocessor_config.json", "etag": "\"0886a48276922a77013d8aa4681192138ae90d90\""}
model_checkpoints/86ed03fdf2dc6dfd5e306b11948471c225fe9080a51c2b5f2f58a708e59f65fa.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.lock ADDED
File without changes
model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23 ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_size": 1,
4
+ "padding_side": "right",
5
+ "padding_value": 0.0,
6
+ "return_attention_mask": true,
7
+ "sampling_rate": 16000
8
+ }
model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"url": "https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-large/resolve/main/preprocessor_config.json", "etag": "\"0886a48276922a77013d8aa4681192138ae90d90\""}
model_checkpoints/907639155bef046ba66b16c7a377f8cd45a6a81323bb2bb8feb817962e525368.fcd266b775b7f33ba9b607a0fee7cc615aeb2eb281586f046280492ea380ae23.lock ADDED
File without changes
tokenizer.ckpt CHANGED
Binary files a/tokenizer.ckpt and b/tokenizer.ckpt differ
 
wav2vec2.ckpt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ae1869d41bd746312a183ce45f3119696d6a275680b0a01a7e5d2ebeba7e8a42
3
  size 1261930757
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b2d9f900fd7a57a30bdc6220606f1ccf37f582f07aab7a5b75213ac46c38204
3
  size 1261930757