Hugging Face's logo

ZikXewen
/
wav2vec2-large-xlsr-53-thai-demo

Automatic Speech Recognition
Transformers
PyTorch
JAX
wav2vec2
Inference Endpoints
Model card Files Community
Supanat commited on
Commit
c2ca524
•
1 Parent(s): 03de77a

first commit

Browse files
Files changed (9) hide show
  1. config.json +84 -0
  2. optimizer.pt +3 -0
  3. preprocessor_config.json +9 -0
  4. pytorch_model.bin +3 -0
  5. rng_state.pth +3 -0
  6. scaler.pt +3 -0
  7. scheduler.pt +3 -0
  8. trainer_state.json +136 -0
  9. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
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
+ "codevector_dim": 768,
11
+ "contrastive_logits_temperature": 0.1,
12
+ "conv_bias": true,
13
+ "conv_dim": [
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512,
20
+ 512
21
+ ],
22
+ "conv_kernel": [
23
+ 10,
24
+ 3,
25
+ 3,
26
+ 3,
27
+ 3,
28
+ 2,
29
+ 2
30
+ ],
31
+ "conv_stride": [
32
+ 5,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2,
37
+ 2,
38
+ 2
39
+ ],
40
+ "ctc_loss_reduction": "mean",
41
+ "ctc_zero_infinity": false,
42
+ "diversity_loss_weight": 0.1,
43
+ "do_stable_layer_norm": true,
44
+ "eos_token_id": 2,
45
+ "feat_extract_activation": "gelu",
46
+ "feat_extract_dropout": 0.0,
47
+ "feat_extract_norm": "layer",
48
+ "feat_proj_dropout": 0.1,
49
+ "feat_quantizer_dropout": 0.0,
50
+ "final_dropout": 0.0,
51
+ "gradient_checkpointing": true,
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": 0.075,
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": 70,
81
+ "proj_codevector_dim": 768,
82
+ "transformers_version": "4.8.1",
83
+ "vocab_size": 73
84
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:640d3248efc4749756152e3931ba16322af55bfa75cb93adde7b9671703a52f5
3
+ size 2490657553
preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": true,
8
+ "sampling_rate": 16000
9
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:815ce4f48b732198a0e83879aa6353cb6f1954d96b8cdd209ee05cf74b50e066
3
+ size 1262222961
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ea5972ee1e8528291bf676f424b25bfaec398b61d17dfd328bc984546980544
3
+ size 14631
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3de5c4545ae0615ca3f287f4fc8f8c5321fbeba6a51f4e4fce1f5d4155e19ed
3
+ size 559
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13bca7a62c95be7bf889fe09343566183dba689324be024d7d0ed49b9d504ff9
3
+ size 623
trainer_state.json ADDED
@@ -0,0 +1,136 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 10.59504132231405,
5
+ "global_step": 3200,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 1.32,
12
+ "learning_rate": 0.000237,
13
+ "loss": 6.8607,
14
+ "step": 400
15
+ },
16
+ {
17
+ "epoch": 1.32,
18
+ "eval_cer": 0.9999129564346956,
19
+ "eval_loss": 3.4867048263549805,
20
+ "eval_runtime": 258.7779,
21
+ "eval_samples_per_second": 8.455,
22
+ "eval_steps_per_second": 1.059,
23
+ "step": 400
24
+ },
25
+ {
26
+ "epoch": 2.65,
27
+ "learning_rate": 0.0002840252707581227,
28
+ "loss": 2.7103,
29
+ "step": 800
30
+ },
31
+ {
32
+ "epoch": 2.65,
33
+ "eval_cer": 0.3108760934847891,
34
+ "eval_loss": 1.061523675918579,
35
+ "eval_runtime": 261.8355,
36
+ "eval_samples_per_second": 8.356,
37
+ "eval_steps_per_second": 1.046,
38
+ "step": 800
39
+ },
40
+ {
41
+ "epoch": 3.97,
42
+ "learning_rate": 0.0002623646209386281,
43
+ "loss": 1.0054,
44
+ "step": 1200
45
+ },
46
+ {
47
+ "epoch": 3.97,
48
+ "eval_cer": 0.22384703544123835,
49
+ "eval_loss": 0.6996724605560303,
50
+ "eval_runtime": 261.6376,
51
+ "eval_samples_per_second": 8.363,
52
+ "eval_steps_per_second": 1.047,
53
+ "step": 1200
54
+ },
55
+ {
56
+ "epoch": 5.3,
57
+ "learning_rate": 0.00024070397111913356,
58
+ "loss": 0.6836,
59
+ "step": 1600
60
+ },
61
+ {
62
+ "epoch": 5.3,
63
+ "eval_cer": 0.19226472849661255,
64
+ "eval_loss": 0.6144178509712219,
65
+ "eval_runtime": 259.8219,
66
+ "eval_samples_per_second": 8.421,
67
+ "eval_steps_per_second": 1.055,
68
+ "step": 1600
69
+ },
70
+ {
71
+ "epoch": 6.62,
72
+ "learning_rate": 0.00021904332129963898,
73
+ "loss": 0.5372,
74
+ "step": 2000
75
+ },
76
+ {
77
+ "epoch": 6.62,
78
+ "eval_cer": 0.1836474155314735,
79
+ "eval_loss": 0.60425865650177,
80
+ "eval_runtime": 259.9127,
81
+ "eval_samples_per_second": 8.418,
82
+ "eval_steps_per_second": 1.054,
83
+ "step": 2000
84
+ },
85
+ {
86
+ "epoch": 7.95,
87
+ "learning_rate": 0.0001973826714801444,
88
+ "loss": 0.4427,
89
+ "step": 2400
90
+ },
91
+ {
92
+ "epoch": 7.95,
93
+ "eval_cer": 0.16931424177800988,
94
+ "eval_loss": 0.5793657898902893,
95
+ "eval_runtime": 260.1302,
96
+ "eval_samples_per_second": 8.411,
97
+ "eval_steps_per_second": 1.053,
98
+ "step": 2400
99
+ },
100
+ {
101
+ "epoch": 9.27,
102
+ "learning_rate": 0.0001757220216606498,
103
+ "loss": 0.3773,
104
+ "step": 2800
105
+ },
106
+ {
107
+ "epoch": 9.27,
108
+ "eval_cer": 0.1633082357720039,
109
+ "eval_loss": 0.6081105470657349,
110
+ "eval_runtime": 260.6021,
111
+ "eval_samples_per_second": 8.396,
112
+ "eval_steps_per_second": 1.051,
113
+ "step": 2800
114
+ },
115
+ {
116
+ "epoch": 10.6,
117
+ "learning_rate": 0.0001540613718411552,
118
+ "loss": 0.3241,
119
+ "step": 3200
120
+ },
121
+ {
122
+ "epoch": 10.6,
123
+ "eval_cer": 0.15936226081153618,
124
+ "eval_loss": 0.6171871423721313,
125
+ "eval_runtime": 260.2528,
126
+ "eval_samples_per_second": 8.407,
127
+ "eval_steps_per_second": 1.053,
128
+ "step": 3200
129
+ }
130
+ ],
131
+ "max_steps": 6040,
132
+ "num_train_epochs": 20,
133
+ "total_flos": 6.610023346999888e+18,
134
+ "trial_name": null,
135
+ "trial_params": null
136
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c029403bc722a36fa6eae21e60b868e397e2bf017148447664da681a437b0a56
3
+ size 2607