Hugging Face's logo

nikhil6041
/
wav2vec2-commonvoice-hindi

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
nikhil6041 commited on
Commit
e44a4df
•
1 Parent(s): cfbb79a

Training in progress, step 100

Browse files
Files changed (7) hide show
  1. config.json +14 -45
  2. pytorch_model.bin +2 -2
  3. runs/Apr02_03-44-58_b05abfbc56da/1648871215.978871/events.out.tfevents.1648871215.b05abfbc56da.72.1 +3 -0
  4. runs/Apr02_03-44-58_b05abfbc56da/events.out.tfevents.1648871215.b05abfbc56da.72.0 +3 -0
  5. runs/Apr02_03-50-30_b05abfbc56da/1648871447.4040134/events.out.tfevents.1648871447.b05abfbc56da.72.3 +3 -0
  6. runs/Apr02_03-50-30_b05abfbc56da/events.out.tfevents.1648871447.b05abfbc56da.72.2 +3 -0
  7. training_args.bin +1 -1
config.json CHANGED
@@ -1,19 +1,16 @@
1
  {
2
- "_name_or_path": "Harveenchadha/vakyansh_hindi_base_pretrained",
3
  "activation_dropout": 0.0,
4
- "adapter_kernel_size": 3,
5
- "adapter_stride": 2,
6
- "add_adapter": false,
7
  "apply_spec_augment": true,
8
  "architectures": [
9
  "Wav2Vec2ForCTC"
10
  ],
11
- "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
  "codevector_dim": 256,
15
  "contrastive_logits_temperature": 0.1,
16
- "conv_bias": false,
17
  "conv_dim": [
18
  512,
19
  512,
@@ -44,74 +41,46 @@
44
  "ctc_loss_reduction": "mean",
45
  "ctc_zero_infinity": false,
46
  "diversity_loss_weight": 0.1,
47
- "do_stable_layer_norm": false,
48
  "eos_token_id": 2,
49
  "feat_extract_activation": "gelu",
50
- "feat_extract_norm": "group",
 
51
  "feat_proj_dropout": 0.0,
52
  "feat_quantizer_dropout": 0.0,
53
  "final_dropout": 0.0,
54
- "freeze_feat_extract_train": true,
55
  "hidden_act": "gelu",
56
- "hidden_dropout": 0.0,
57
- "hidden_size": 768,
58
  "initializer_range": 0.02,
59
- "intermediate_size": 3072,
60
  "layer_norm_eps": 1e-05,
61
- "layerdrop": 0.0,
62
  "mask_channel_length": 10,
63
  "mask_channel_min_space": 1,
64
  "mask_channel_other": 0.0,
65
  "mask_channel_prob": 0.0,
66
  "mask_channel_selection": "static",
67
  "mask_feature_length": 10,
68
- "mask_feature_min_masks": 0,
69
  "mask_feature_prob": 0.0,
70
  "mask_time_length": 10,
71
- "mask_time_min_masks": 2,
72
  "mask_time_min_space": 1,
73
  "mask_time_other": 0.0,
74
  "mask_time_prob": 0.05,
75
  "mask_time_selection": "static",
76
  "model_type": "wav2vec2",
77
- "no_mask_channel_overlap": false,
78
- "no_mask_time_overlap": false,
79
- "num_adapter_layers": 3,
80
- "num_attention_heads": 12,
81
  "num_codevector_groups": 2,
82
  "num_codevectors_per_group": 320,
83
  "num_conv_pos_embedding_groups": 16,
84
  "num_conv_pos_embeddings": 128,
85
  "num_feat_extract_layers": 7,
86
- "num_hidden_layers": 12,
87
  "num_negatives": 100,
88
- "output_hidden_size": 768,
89
- "pad_token_id": 72,
90
  "proj_codevector_dim": 256,
91
- "tdnn_dilation": [
92
- 1,
93
- 2,
94
- 3,
95
- 1,
96
- 1
97
- ],
98
- "tdnn_dim": [
99
- 512,
100
- 512,
101
- 512,
102
- 512,
103
- 1500
104
- ],
105
- "tdnn_kernel": [
106
- 5,
107
- 3,
108
- 3,
109
- 1,
110
- 1
111
- ],
112
  "torch_dtype": "float32",
113
  "transformers_version": "4.11.3",
114
  "use_weighted_layer_sum": false,
115
- "vocab_size": 75,
116
- "xvector_output_dim": 512
117
  }
 
1
  {
2
+ "_name_or_path": "theainerd/Wav2Vec2-large-xlsr-hindi",
3
  "activation_dropout": 0.0,
 
 
 
4
  "apply_spec_augment": true,
5
  "architectures": [
6
  "Wav2Vec2ForCTC"
7
  ],
8
+ "attention_dropout": 0.1,
9
  "bos_token_id": 1,
10
  "classifier_proj_size": 256,
11
  "codevector_dim": 256,
12
  "contrastive_logits_temperature": 0.1,
13
+ "conv_bias": true,
14
  "conv_dim": [
15
  512,
16
  512,
 
41
  "ctc_loss_reduction": "mean",
42
  "ctc_zero_infinity": false,
43
  "diversity_loss_weight": 0.1,
44
+ "do_stable_layer_norm": true,
45
  "eos_token_id": 2,
46
  "feat_extract_activation": "gelu",
47
+ "feat_extract_dropout": 0.0,
48
+ "feat_extract_norm": "layer",
49
  "feat_proj_dropout": 0.0,
50
  "feat_quantizer_dropout": 0.0,
51
  "final_dropout": 0.0,
 
52
  "hidden_act": "gelu",
53
+ "hidden_dropout": 0.1,
54
+ "hidden_size": 1024,
55
  "initializer_range": 0.02,
56
+ "intermediate_size": 4096,
57
  "layer_norm_eps": 1e-05,
58
+ "layerdrop": 0.1,
59
  "mask_channel_length": 10,
60
  "mask_channel_min_space": 1,
61
  "mask_channel_other": 0.0,
62
  "mask_channel_prob": 0.0,
63
  "mask_channel_selection": "static",
64
  "mask_feature_length": 10,
 
65
  "mask_feature_prob": 0.0,
66
  "mask_time_length": 10,
 
67
  "mask_time_min_space": 1,
68
  "mask_time_other": 0.0,
69
  "mask_time_prob": 0.05,
70
  "mask_time_selection": "static",
71
  "model_type": "wav2vec2",
72
+ "num_attention_heads": 16,
 
 
 
73
  "num_codevector_groups": 2,
74
  "num_codevectors_per_group": 320,
75
  "num_conv_pos_embedding_groups": 16,
76
  "num_conv_pos_embeddings": 128,
77
  "num_feat_extract_layers": 7,
78
+ "num_hidden_layers": 24,
79
  "num_negatives": 100,
80
+ "pad_token_id": 63,
 
81
  "proj_codevector_dim": 256,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
82
  "torch_dtype": "float32",
83
  "transformers_version": "4.11.3",
84
  "use_weighted_layer_sum": false,
85
+ "vocab_size": 64
 
86
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1ac63b2fe3ccc50f34359cfcd860b863e6694666d3430ba09f36526cb4b8f0c4
3
- size 377802263
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93646cba7b3e770705000b1e0cbe287fc57835b24463f75033b7d0d216e025ec
3
+ size 1262186097
runs/Apr02_03-44-58_b05abfbc56da/1648871215.978871/events.out.tfevents.1648871215.b05abfbc56da.72.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:612f705ceb3896eea4349c5f23f21ec512652f3d10e738aad632401bda3dbcc3
3
+ size 4543
runs/Apr02_03-44-58_b05abfbc56da/events.out.tfevents.1648871215.b05abfbc56da.72.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48beedc590e7f16d8f712b583952351ae4f3796dad122cba81a8347f0af8c9fc
3
+ size 4379
runs/Apr02_03-50-30_b05abfbc56da/1648871447.4040134/events.out.tfevents.1648871447.b05abfbc56da.72.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8710680ccd5959692c8d49a7aa94e26f18d6a9dc43e1676309d224d9b4615a00
3
+ size 4543
runs/Apr02_03-50-30_b05abfbc56da/events.out.tfevents.1648871447.b05abfbc56da.72.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f2ab66a8c019f8f14010ca9b7f6ebc348c37c29a7880c14916f4d5e7433b0ec
3
+ size 4691
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3d25cf1611496f98c71c2f696a10183963457f0be6fe85b045f8057fd002b122
3
  size 2863
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0957ba1a13aadf0d7a718e155989208fa89733655ced568487d046c1d23abba8
3
  size 2863