Hugging Face's logo

Taewhoo
/
wav2vec2_esp

TensorBoard
Model card Files Metrics Community
Taewhoo commited on
Commit
6c314fe
•
1 Parent(s): 9f79a5c

Upload with huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +12 -0
  2. wav2vec2_esp_15h/checkpoint-22000/config.json +117 -0
  3. wav2vec2_esp_15h/checkpoint-22000/optimizer.pt +3 -0
  4. wav2vec2_esp_15h/checkpoint-22000/preprocessor_config.json +9 -0
  5. wav2vec2_esp_15h/checkpoint-22000/pytorch_model.bin +3 -0
  6. wav2vec2_esp_15h/checkpoint-22000/rng_state.pth +0 -0
  7. wav2vec2_esp_15h/checkpoint-22000/scaler.pt +0 -0
  8. wav2vec2_esp_15h/checkpoint-22000/scheduler.pt +0 -0
  9. wav2vec2_esp_15h/checkpoint-22000/trainer_state.json +676 -0
  10. wav2vec2_esp_15h/checkpoint-22000/training_args.bin +0 -0
  11. wav2vec2_esp_15h/checkpoint-22500/config.json +117 -0
  12. wav2vec2_esp_15h/checkpoint-22500/optimizer.pt +3 -0
  13. wav2vec2_esp_15h/checkpoint-22500/preprocessor_config.json +9 -0
  14. wav2vec2_esp_15h/checkpoint-22500/pytorch_model.bin +3 -0
  15. wav2vec2_esp_15h/checkpoint-22500/rng_state.pth +0 -0
  16. wav2vec2_esp_15h/checkpoint-22500/scaler.pt +0 -0
  17. wav2vec2_esp_15h/checkpoint-22500/scheduler.pt +0 -0
  18. wav2vec2_esp_15h/checkpoint-22500/trainer_state.json +691 -0
  19. wav2vec2_esp_15h/checkpoint-22500/training_args.bin +0 -0
  20. wav2vec2_esp_15h/runs/Mar08_04-04-36_mint/1678216062.1216896/events.out.tfevents.1678216062.mint +0 -0
  21. wav2vec2_esp_15h/runs/Mar08_04-04-36_mint/events.out.tfevents.1678216062.mint +0 -0
  22. wav2vec2_esp_30h/checkpoint-43500/config.json +117 -0
  23. wav2vec2_esp_30h/checkpoint-43500/optimizer.pt +3 -0
  24. wav2vec2_esp_30h/checkpoint-43500/preprocessor_config.json +9 -0
  25. wav2vec2_esp_30h/checkpoint-43500/pytorch_model.bin +3 -0
  26. wav2vec2_esp_30h/checkpoint-43500/rng_state.pth +0 -0
  27. wav2vec2_esp_30h/checkpoint-43500/scaler.pt +0 -0
  28. wav2vec2_esp_30h/checkpoint-43500/scheduler.pt +0 -0
  29. wav2vec2_esp_30h/checkpoint-43500/trainer_state.json +1321 -0
  30. wav2vec2_esp_30h/checkpoint-43500/training_args.bin +0 -0
  31. wav2vec2_esp_30h/checkpoint-44000/config.json +117 -0
  32. wav2vec2_esp_30h/checkpoint-44000/optimizer.pt +3 -0
  33. wav2vec2_esp_30h/checkpoint-44000/preprocessor_config.json +9 -0
  34. wav2vec2_esp_30h/checkpoint-44000/pytorch_model.bin +3 -0
  35. wav2vec2_esp_30h/checkpoint-44000/rng_state.pth +0 -0
  36. wav2vec2_esp_30h/checkpoint-44000/scaler.pt +0 -0
  37. wav2vec2_esp_30h/checkpoint-44000/scheduler.pt +0 -0
  38. wav2vec2_esp_30h/checkpoint-44000/trainer_state.json +1336 -0
  39. wav2vec2_esp_30h/checkpoint-44000/training_args.bin +0 -0
  40. wav2vec2_esp_30h/runs/Mar08_02-48-55_mint/1678211683.9648502/events.out.tfevents.1678211683.mint +0 -0
  41. wav2vec2_esp_30h/runs/Mar08_02-48-55_mint/events.out.tfevents.1678211683.mint +0 -0
  42. wav2vec2_esp_30h/runs/Mar08_03-58-32_mint/1678215863.173738/events.out.tfevents.1678215863.mint +0 -0
  43. wav2vec2_esp_30h/runs/Mar08_03-58-32_mint/events.out.tfevents.1678215863.mint +0 -0
  44. wav2vec2_esp_5h/checkpoint-8000/config.json +117 -0
  45. wav2vec2_esp_5h/checkpoint-8000/optimizer.pt +3 -0
  46. wav2vec2_esp_5h/checkpoint-8000/preprocessor_config.json +9 -0
  47. wav2vec2_esp_5h/checkpoint-8000/pytorch_model.bin +3 -0
  48. wav2vec2_esp_5h/checkpoint-8000/rng_state.pth +0 -0
  49. wav2vec2_esp_5h/checkpoint-8000/scaler.pt +0 -0
  50. wav2vec2_esp_5h/checkpoint-8000/scheduler.pt +0 -0
.gitattributes ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ wav2vec2_esp_15h/checkpoint-22500/optimizer.pt filter=lfs diff=lfs merge=lfs -text
2
+ wav2vec2_esp_15h/checkpoint-22500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
3
+ wav2vec2_esp_15h/checkpoint-22000/optimizer.pt filter=lfs diff=lfs merge=lfs -text
4
+ wav2vec2_esp_15h/checkpoint-22000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
5
+ wav2vec2_esp_5h/checkpoint-8000/optimizer.pt filter=lfs diff=lfs merge=lfs -text
6
+ wav2vec2_esp_5h/checkpoint-8000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
7
+ wav2vec2_esp_5h/checkpoint-8500/optimizer.pt filter=lfs diff=lfs merge=lfs -text
8
+ wav2vec2_esp_5h/checkpoint-8500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
9
+ wav2vec2_esp_30h/checkpoint-44000/optimizer.pt filter=lfs diff=lfs merge=lfs -text
10
+ wav2vec2_esp_30h/checkpoint-44000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
11
+ wav2vec2_esp_30h/checkpoint-43500/optimizer.pt filter=lfs diff=lfs merge=lfs -text
12
+ wav2vec2_esp_30h/checkpoint-43500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
wav2vec2_esp_15h/checkpoint-22000/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
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": 0.1,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
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": 0.1,
52
+ "feat_quantizer_dropout": 0.0,
53
+ "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
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": 37,
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": "4.17.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 38,
116
+ "xvector_output_dim": 512
117
+ }
wav2vec2_esp_15h/checkpoint-22000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60486088c50cd0525ab854a55277c2dc80e2e8d5c5c76645332befaa06db0358
3
+ size 721723525
wav2vec2_esp_15h/checkpoint-22000/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
+ }
wav2vec2_esp_15h/checkpoint-22000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:627c8266a4f9c844c240dce1a26a3b0daf5c09696c12507ea040b8387fc75134
3
+ size 377691873
wav2vec2_esp_15h/checkpoint-22000/rng_state.pth ADDED
Binary file (14.7 kB). View file
 
wav2vec2_esp_15h/checkpoint-22000/scaler.pt ADDED
Binary file (557 Bytes). View file
 
wav2vec2_esp_15h/checkpoint-22000/scheduler.pt ADDED
Binary file (627 Bytes). View file
 
wav2vec2_esp_15h/checkpoint-22000/trainer_state.json ADDED
@@ -0,0 +1,676 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 28.79581151832461,
5
+ "global_step": 22000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.65,
12
+ "learning_rate": 4.94e-05,
13
+ "loss": 6.5844,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.65,
18
+ "eval_loss": 2.985063076019287,
19
+ "eval_runtime": 66.7207,
20
+ "eval_samples_per_second": 15.348,
21
+ "eval_steps_per_second": 1.918,
22
+ "eval_wer": 1.0,
23
+ "step": 500
24
+ },
25
+ {
26
+ "epoch": 1.31,
27
+ "learning_rate": 9.94e-05,
28
+ "loss": 2.1717,
29
+ "step": 1000
30
+ },
31
+ {
32
+ "epoch": 1.31,
33
+ "eval_loss": 1.0769672393798828,
34
+ "eval_runtime": 66.0484,
35
+ "eval_samples_per_second": 15.504,
36
+ "eval_steps_per_second": 1.938,
37
+ "eval_wer": 0.8259404828747895,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 1.96,
42
+ "learning_rate": 9.774635036496352e-05,
43
+ "loss": 1.0058,
44
+ "step": 1500
45
+ },
46
+ {
47
+ "epoch": 1.96,
48
+ "eval_loss": 0.7441346049308777,
49
+ "eval_runtime": 66.5848,
50
+ "eval_samples_per_second": 15.379,
51
+ "eval_steps_per_second": 1.922,
52
+ "eval_wer": 0.6882650196518809,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 2.62,
57
+ "learning_rate": 9.546532846715329e-05,
58
+ "loss": 0.7316,
59
+ "step": 2000
60
+ },
61
+ {
62
+ "epoch": 2.62,
63
+ "eval_loss": 0.6406537294387817,
64
+ "eval_runtime": 65.9141,
65
+ "eval_samples_per_second": 15.535,
66
+ "eval_steps_per_second": 1.942,
67
+ "eval_wer": 0.617630544637844,
68
+ "step": 2000
69
+ },
70
+ {
71
+ "epoch": 3.27,
72
+ "learning_rate": 9.318430656934307e-05,
73
+ "loss": 0.6249,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 3.27,
78
+ "eval_loss": 0.5814340114593506,
79
+ "eval_runtime": 66.4954,
80
+ "eval_samples_per_second": 15.4,
81
+ "eval_steps_per_second": 1.925,
82
+ "eval_wer": 0.5783267827063447,
83
+ "step": 2500
84
+ },
85
+ {
86
+ "epoch": 3.93,
87
+ "learning_rate": 9.090328467153285e-05,
88
+ "loss": 0.5541,
89
+ "step": 3000
90
+ },
91
+ {
92
+ "epoch": 3.93,
93
+ "eval_loss": 0.5387497544288635,
94
+ "eval_runtime": 66.9564,
95
+ "eval_samples_per_second": 15.294,
96
+ "eval_steps_per_second": 1.912,
97
+ "eval_wer": 0.5449747332959012,
98
+ "step": 3000
99
+ },
100
+ {
101
+ "epoch": 4.58,
102
+ "learning_rate": 8.862226277372263e-05,
103
+ "loss": 0.4727,
104
+ "step": 3500
105
+ },
106
+ {
107
+ "epoch": 4.58,
108
+ "eval_loss": 0.5657380223274231,
109
+ "eval_runtime": 66.1646,
110
+ "eval_samples_per_second": 15.477,
111
+ "eval_steps_per_second": 1.935,
112
+ "eval_wer": 0.5357664233576642,
113
+ "step": 3500
114
+ },
115
+ {
116
+ "epoch": 5.24,
117
+ "learning_rate": 8.634124087591242e-05,
118
+ "loss": 0.4444,
119
+ "step": 4000
120
+ },
121
+ {
122
+ "epoch": 5.24,
123
+ "eval_loss": 0.5406317710876465,
124
+ "eval_runtime": 65.9956,
125
+ "eval_samples_per_second": 15.516,
126
+ "eval_steps_per_second": 1.94,
127
+ "eval_wer": 0.5164514317798989,
128
+ "step": 4000
129
+ },
130
+ {
131
+ "epoch": 5.89,
132
+ "learning_rate": 8.406021897810219e-05,
133
+ "loss": 0.4057,
134
+ "step": 4500
135
+ },
136
+ {
137
+ "epoch": 5.89,
138
+ "eval_loss": 0.5271017551422119,
139
+ "eval_runtime": 66.6343,
140
+ "eval_samples_per_second": 15.367,
141
+ "eval_steps_per_second": 1.921,
142
+ "eval_wer": 0.5116226838854576,
143
+ "step": 4500
144
+ },
145
+ {
146
+ "epoch": 6.54,
147
+ "learning_rate": 8.177919708029196e-05,
148
+ "loss": 0.3663,
149
+ "step": 5000
150
+ },
151
+ {
152
+ "epoch": 6.54,
153
+ "eval_loss": 0.5105618238449097,
154
+ "eval_runtime": 66.576,
155
+ "eval_samples_per_second": 15.381,
156
+ "eval_steps_per_second": 1.923,
157
+ "eval_wer": 0.4793935991016283,
158
+ "step": 5000
159
+ },
160
+ {
161
+ "epoch": 7.2,
162
+ "learning_rate": 7.949817518248176e-05,
163
+ "loss": 0.3406,
164
+ "step": 5500
165
+ },
166
+ {
167
+ "epoch": 7.2,
168
+ "eval_loss": 0.5662369132041931,
169
+ "eval_runtime": 67.3476,
170
+ "eval_samples_per_second": 15.205,
171
+ "eval_steps_per_second": 1.901,
172
+ "eval_wer": 0.48568220101066817,
173
+ "step": 5500
174
+ },
175
+ {
176
+ "epoch": 7.85,
177
+ "learning_rate": 7.721715328467154e-05,
178
+ "loss": 0.3142,
179
+ "step": 6000
180
+ },
181
+ {
182
+ "epoch": 7.85,
183
+ "eval_loss": 0.5632680654525757,
184
+ "eval_runtime": 66.2027,
185
+ "eval_samples_per_second": 15.468,
186
+ "eval_steps_per_second": 1.933,
187
+ "eval_wer": 0.4879281302638967,
188
+ "step": 6000
189
+ },
190
+ {
191
+ "epoch": 8.51,
192
+ "learning_rate": 7.493613138686131e-05,
193
+ "loss": 0.3002,
194
+ "step": 6500
195
+ },
196
+ {
197
+ "epoch": 8.51,
198
+ "eval_loss": 0.5218114852905273,
199
+ "eval_runtime": 67.5136,
200
+ "eval_samples_per_second": 15.167,
201
+ "eval_steps_per_second": 1.896,
202
+ "eval_wer": 0.45659741718135877,
203
+ "step": 6500
204
+ },
205
+ {
206
+ "epoch": 9.16,
207
+ "learning_rate": 7.265967153284671e-05,
208
+ "loss": 0.2845,
209
+ "step": 7000
210
+ },
211
+ {
212
+ "epoch": 9.16,
213
+ "eval_loss": 0.5436348915100098,
214
+ "eval_runtime": 66.2104,
215
+ "eval_samples_per_second": 15.466,
216
+ "eval_steps_per_second": 1.933,
217
+ "eval_wer": 0.45457608085345313,
218
+ "step": 7000
219
+ },
220
+ {
221
+ "epoch": 9.82,
222
+ "learning_rate": 7.038321167883212e-05,
223
+ "loss": 0.2603,
224
+ "step": 7500
225
+ },
226
+ {
227
+ "epoch": 9.82,
228
+ "eval_loss": 0.5183758735656738,
229
+ "eval_runtime": 67.776,
230
+ "eval_samples_per_second": 15.109,
231
+ "eval_steps_per_second": 1.889,
232
+ "eval_wer": 0.4415496911847277,
233
+ "step": 7500
234
+ },
235
+ {
236
+ "epoch": 10.47,
237
+ "learning_rate": 6.810218978102189e-05,
238
+ "loss": 0.2438,
239
+ "step": 8000
240
+ },
241
+ {
242
+ "epoch": 10.47,
243
+ "eval_loss": 0.5246397256851196,
244
+ "eval_runtime": 66.5435,
245
+ "eval_samples_per_second": 15.388,
246
+ "eval_steps_per_second": 1.924,
247
+ "eval_wer": 0.43638405390230206,
248
+ "step": 8000
249
+ },
250
+ {
251
+ "epoch": 11.13,
252
+ "learning_rate": 6.582116788321169e-05,
253
+ "loss": 0.2396,
254
+ "step": 8500
255
+ },
256
+ {
257
+ "epoch": 11.13,
258
+ "eval_loss": 0.5663778781890869,
259
+ "eval_runtime": 66.6323,
260
+ "eval_samples_per_second": 15.368,
261
+ "eval_steps_per_second": 1.921,
262
+ "eval_wer": 0.4428972487366648,
263
+ "step": 8500
264
+ },
265
+ {
266
+ "epoch": 11.78,
267
+ "learning_rate": 6.354014598540147e-05,
268
+ "loss": 0.2229,
269
+ "step": 9000
270
+ },
271
+ {
272
+ "epoch": 11.78,
273
+ "eval_loss": 0.5550942420959473,
274
+ "eval_runtime": 65.2937,
275
+ "eval_samples_per_second": 15.683,
276
+ "eval_steps_per_second": 1.96,
277
+ "eval_wer": 0.4415496911847277,
278
+ "step": 9000
279
+ },
280
+ {
281
+ "epoch": 12.43,
282
+ "learning_rate": 6.125912408759124e-05,
283
+ "loss": 0.2088,
284
+ "step": 9500
285
+ },
286
+ {
287
+ "epoch": 12.43,
288
+ "eval_loss": 0.5503721833229065,
289
+ "eval_runtime": 66.2676,
290
+ "eval_samples_per_second": 15.453,
291
+ "eval_steps_per_second": 1.932,
292
+ "eval_wer": 0.4257158899494666,
293
+ "step": 9500
294
+ },
295
+ {
296
+ "epoch": 13.09,
297
+ "learning_rate": 5.897810218978103e-05,
298
+ "loss": 0.2042,
299
+ "step": 10000
300
+ },
301
+ {
302
+ "epoch": 13.09,
303
+ "eval_loss": 0.5646582245826721,
304
+ "eval_runtime": 66.4465,
305
+ "eval_samples_per_second": 15.411,
306
+ "eval_steps_per_second": 1.926,
307
+ "eval_wer": 0.4398652442448063,
308
+ "step": 10000
309
+ },
310
+ {
311
+ "epoch": 13.74,
312
+ "learning_rate": 5.669708029197081e-05,
313
+ "loss": 0.1975,
314
+ "step": 10500
315
+ },
316
+ {
317
+ "epoch": 13.74,
318
+ "eval_loss": 0.5424100160598755,
319
+ "eval_runtime": 66.9743,
320
+ "eval_samples_per_second": 15.289,
321
+ "eval_steps_per_second": 1.911,
322
+ "eval_wer": 0.4186412128017967,
323
+ "step": 10500
324
+ },
325
+ {
326
+ "epoch": 14.4,
327
+ "learning_rate": 5.441605839416058e-05,
328
+ "loss": 0.1856,
329
+ "step": 11000
330
+ },
331
+ {
332
+ "epoch": 14.4,
333
+ "eval_loss": 0.5583313703536987,
334
+ "eval_runtime": 66.5902,
335
+ "eval_samples_per_second": 15.378,
336
+ "eval_steps_per_second": 1.922,
337
+ "eval_wer": 0.42515440763615947,
338
+ "step": 11000
339
+ },
340
+ {
341
+ "epoch": 15.05,
342
+ "learning_rate": 5.213503649635036e-05,
343
+ "loss": 0.1872,
344
+ "step": 11500
345
+ },
346
+ {
347
+ "epoch": 15.05,
348
+ "eval_loss": 0.6118334531784058,
349
+ "eval_runtime": 66.3517,
350
+ "eval_samples_per_second": 15.433,
351
+ "eval_steps_per_second": 1.929,
352
+ "eval_wer": 0.433464345873105,
353
+ "step": 11500
354
+ },
355
+ {
356
+ "epoch": 15.71,
357
+ "learning_rate": 4.985401459854015e-05,
358
+ "loss": 0.1676,
359
+ "step": 12000
360
+ },
361
+ {
362
+ "epoch": 15.71,
363
+ "eval_loss": 0.6066869497299194,
364
+ "eval_runtime": 67.2768,
365
+ "eval_samples_per_second": 15.221,
366
+ "eval_steps_per_second": 1.903,
367
+ "eval_wer": 0.43627175743964064,
368
+ "step": 12000
369
+ },
370
+ {
371
+ "epoch": 16.36,
372
+ "learning_rate": 4.757299270072993e-05,
373
+ "loss": 0.1536,
374
+ "step": 12500
375
+ },
376
+ {
377
+ "epoch": 16.36,
378
+ "eval_loss": 0.5462669730186462,
379
+ "eval_runtime": 66.3968,
380
+ "eval_samples_per_second": 15.422,
381
+ "eval_steps_per_second": 1.928,
382
+ "eval_wer": 0.412914093206064,
383
+ "step": 12500
384
+ },
385
+ {
386
+ "epoch": 17.02,
387
+ "learning_rate": 4.5291970802919706e-05,
388
+ "loss": 0.1582,
389
+ "step": 13000
390
+ },
391
+ {
392
+ "epoch": 17.02,
393
+ "eval_loss": 0.5799742937088013,
394
+ "eval_runtime": 65.9177,
395
+ "eval_samples_per_second": 15.535,
396
+ "eval_steps_per_second": 1.942,
397
+ "eval_wer": 0.41021897810218977,
398
+ "step": 13000
399
+ },
400
+ {
401
+ "epoch": 17.67,
402
+ "learning_rate": 4.3015510948905114e-05,
403
+ "loss": 0.1429,
404
+ "step": 13500
405
+ },
406
+ {
407
+ "epoch": 17.67,
408
+ "eval_loss": 0.5899107456207275,
409
+ "eval_runtime": 66.1113,
410
+ "eval_samples_per_second": 15.489,
411
+ "eval_steps_per_second": 1.936,
412
+ "eval_wer": 0.4150477259966311,
413
+ "step": 13500
414
+ },
415
+ {
416
+ "epoch": 18.32,
417
+ "learning_rate": 4.0734489051094895e-05,
418
+ "loss": 0.1451,
419
+ "step": 14000
420
+ },
421
+ {
422
+ "epoch": 18.32,
423
+ "eval_loss": 0.6129721999168396,
424
+ "eval_runtime": 67.4516,
425
+ "eval_samples_per_second": 15.181,
426
+ "eval_steps_per_second": 1.898,
427
+ "eval_wer": 0.4138124649073554,
428
+ "step": 14000
429
+ },
430
+ {
431
+ "epoch": 18.98,
432
+ "learning_rate": 3.845346715328467e-05,
433
+ "loss": 0.1356,
434
+ "step": 14500
435
+ },
436
+ {
437
+ "epoch": 18.98,
438
+ "eval_loss": 0.6040089130401611,
439
+ "eval_runtime": 66.75,
440
+ "eval_samples_per_second": 15.341,
441
+ "eval_steps_per_second": 1.918,
442
+ "eval_wer": 0.4123526108927569,
443
+ "step": 14500
444
+ },
445
+ {
446
+ "epoch": 19.63,
447
+ "learning_rate": 3.617244525547445e-05,
448
+ "loss": 0.134,
449
+ "step": 15000
450
+ },
451
+ {
452
+ "epoch": 19.63,
453
+ "eval_loss": 0.5997006893157959,
454
+ "eval_runtime": 67.809,
455
+ "eval_samples_per_second": 15.101,
456
+ "eval_steps_per_second": 1.888,
457
+ "eval_wer": 0.40819764177428414,
458
+ "step": 15000
459
+ },
460
+ {
461
+ "epoch": 20.29,
462
+ "learning_rate": 3.389598540145986e-05,
463
+ "loss": 0.1305,
464
+ "step": 15500
465
+ },
466
+ {
467
+ "epoch": 20.29,
468
+ "eval_loss": 0.5769977569580078,
469
+ "eval_runtime": 66.9982,
470
+ "eval_samples_per_second": 15.284,
471
+ "eval_steps_per_second": 1.91,
472
+ "eval_wer": 0.40235822571588997,
473
+ "step": 15500
474
+ },
475
+ {
476
+ "epoch": 20.94,
477
+ "learning_rate": 3.161496350364963e-05,
478
+ "loss": 0.1222,
479
+ "step": 16000
480
+ },
481
+ {
482
+ "epoch": 20.94,
483
+ "eval_loss": 0.5980133414268494,
484
+ "eval_runtime": 66.1854,
485
+ "eval_samples_per_second": 15.472,
486
+ "eval_steps_per_second": 1.934,
487
+ "eval_wer": 0.391577765300393,
488
+ "step": 16000
489
+ },
490
+ {
491
+ "epoch": 21.6,
492
+ "learning_rate": 2.9333941605839417e-05,
493
+ "loss": 0.1214,
494
+ "step": 16500
495
+ },
496
+ {
497
+ "epoch": 21.6,
498
+ "eval_loss": 0.5917083024978638,
499
+ "eval_runtime": 67.5557,
500
+ "eval_samples_per_second": 15.158,
501
+ "eval_steps_per_second": 1.895,
502
+ "eval_wer": 0.4005614823133071,
503
+ "step": 16500
504
+ },
505
+ {
506
+ "epoch": 22.25,
507
+ "learning_rate": 2.70529197080292e-05,
508
+ "loss": 0.1184,
509
+ "step": 17000
510
+ },
511
+ {
512
+ "epoch": 22.25,
513
+ "eval_loss": 0.578889787197113,
514
+ "eval_runtime": 66.5407,
515
+ "eval_samples_per_second": 15.389,
516
+ "eval_steps_per_second": 1.924,
517
+ "eval_wer": 0.39460976979225154,
518
+ "step": 17000
519
+ },
520
+ {
521
+ "epoch": 22.91,
522
+ "learning_rate": 2.477189781021898e-05,
523
+ "loss": 0.1132,
524
+ "step": 17500
525
+ },
526
+ {
527
+ "epoch": 22.91,
528
+ "eval_loss": 0.5992549657821655,
529
+ "eval_runtime": 66.4759,
530
+ "eval_samples_per_second": 15.404,
531
+ "eval_steps_per_second": 1.926,
532
+ "eval_wer": 0.3892195395845031,
533
+ "step": 17500
534
+ },
535
+ {
536
+ "epoch": 23.56,
537
+ "learning_rate": 2.249087591240876e-05,
538
+ "loss": 0.1052,
539
+ "step": 18000
540
+ },
541
+ {
542
+ "epoch": 23.56,
543
+ "eval_loss": 0.5946719646453857,
544
+ "eval_runtime": 66.8036,
545
+ "eval_samples_per_second": 15.329,
546
+ "eval_steps_per_second": 1.916,
547
+ "eval_wer": 0.3889949466591802,
548
+ "step": 18000
549
+ },
550
+ {
551
+ "epoch": 24.21,
552
+ "learning_rate": 2.0209854014598544e-05,
553
+ "loss": 0.113,
554
+ "step": 18500
555
+ },
556
+ {
557
+ "epoch": 24.21,
558
+ "eval_loss": 0.6009370684623718,
559
+ "eval_runtime": 65.8523,
560
+ "eval_samples_per_second": 15.55,
561
+ "eval_steps_per_second": 1.944,
562
+ "eval_wer": 0.3841661987647389,
563
+ "step": 18500
564
+ },
565
+ {
566
+ "epoch": 24.87,
567
+ "learning_rate": 1.792883211678832e-05,
568
+ "loss": 0.1021,
569
+ "step": 19000
570
+ },
571
+ {
572
+ "epoch": 24.87,
573
+ "eval_loss": 0.6255775690078735,
574
+ "eval_runtime": 67.7689,
575
+ "eval_samples_per_second": 15.11,
576
+ "eval_steps_per_second": 1.889,
577
+ "eval_wer": 0.3841661987647389,
578
+ "step": 19000
579
+ },
580
+ {
581
+ "epoch": 25.52,
582
+ "learning_rate": 1.5652372262773723e-05,
583
+ "loss": 0.0966,
584
+ "step": 19500
585
+ },
586
+ {
587
+ "epoch": 25.52,
588
+ "eval_loss": 0.6213188171386719,
589
+ "eval_runtime": 68.5057,
590
+ "eval_samples_per_second": 14.948,
591
+ "eval_steps_per_second": 1.868,
592
+ "eval_wer": 0.3869736103312746,
593
+ "step": 19500
594
+ },
595
+ {
596
+ "epoch": 26.18,
597
+ "learning_rate": 1.3371350364963504e-05,
598
+ "loss": 0.0981,
599
+ "step": 20000
600
+ },
601
+ {
602
+ "epoch": 26.18,
603
+ "eval_loss": 0.6331676244735718,
604
+ "eval_runtime": 67.2276,
605
+ "eval_samples_per_second": 15.232,
606
+ "eval_steps_per_second": 1.904,
607
+ "eval_wer": 0.37866367209432905,
608
+ "step": 20000
609
+ },
610
+ {
611
+ "epoch": 26.83,
612
+ "learning_rate": 1.1090328467153285e-05,
613
+ "loss": 0.0983,
614
+ "step": 20500
615
+ },
616
+ {
617
+ "epoch": 26.83,
618
+ "eval_loss": 0.6225576996803284,
619
+ "eval_runtime": 66.5907,
620
+ "eval_samples_per_second": 15.378,
621
+ "eval_steps_per_second": 1.922,
622
+ "eval_wer": 0.3806850084222347,
623
+ "step": 20500
624
+ },
625
+ {
626
+ "epoch": 27.49,
627
+ "learning_rate": 8.809306569343067e-06,
628
+ "loss": 0.0915,
629
+ "step": 21000
630
+ },
631
+ {
632
+ "epoch": 27.49,
633
+ "eval_loss": 0.6176589727401733,
634
+ "eval_runtime": 66.1318,
635
+ "eval_samples_per_second": 15.484,
636
+ "eval_steps_per_second": 1.936,
637
+ "eval_wer": 0.3785513756316676,
638
+ "step": 21000
639
+ },
640
+ {
641
+ "epoch": 28.14,
642
+ "learning_rate": 6.532846715328468e-06,
643
+ "loss": 0.0898,
644
+ "step": 21500
645
+ },
646
+ {
647
+ "epoch": 28.14,
648
+ "eval_loss": 0.6252412796020508,
649
+ "eval_runtime": 66.2396,
650
+ "eval_samples_per_second": 15.459,
651
+ "eval_steps_per_second": 1.932,
652
+ "eval_wer": 0.3828186412128018,
653
+ "step": 21500
654
+ },
655
+ {
656
+ "epoch": 28.8,
657
+ "learning_rate": 4.251824817518248e-06,
658
+ "loss": 0.0916,
659
+ "step": 22000
660
+ },
661
+ {
662
+ "epoch": 28.8,
663
+ "eval_loss": 0.6163517832756042,
664
+ "eval_runtime": 66.1166,
665
+ "eval_samples_per_second": 15.488,
666
+ "eval_steps_per_second": 1.936,
667
+ "eval_wer": 0.3760808534531162,
668
+ "step": 22000
669
+ }
670
+ ],
671
+ "max_steps": 22920,
672
+ "num_train_epochs": 30,
673
+ "total_flos": 6.988254857647251e+18,
674
+ "trial_name": null,
675
+ "trial_params": null
676
+ }
wav2vec2_esp_15h/checkpoint-22000/training_args.bin ADDED
Binary file (3 kB). View file
 
wav2vec2_esp_15h/checkpoint-22500/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
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": 0.1,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
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": 0.1,
52
+ "feat_quantizer_dropout": 0.0,
53
+ "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
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": 37,
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": "4.17.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 38,
116
+ "xvector_output_dim": 512
117
+ }
wav2vec2_esp_15h/checkpoint-22500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:818ef3786510bd9b85118a60afb89c3e4c7ddb35ab3e16998f83ab3e5823033a
3
+ size 721723525
wav2vec2_esp_15h/checkpoint-22500/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
+ }
wav2vec2_esp_15h/checkpoint-22500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a06b11c51f693b8bcac67bc3aaafb31d974e586dfe30cd4ae09654564ab86df
3
+ size 377691873
wav2vec2_esp_15h/checkpoint-22500/rng_state.pth ADDED
Binary file (14.6 kB). View file
 
wav2vec2_esp_15h/checkpoint-22500/scaler.pt ADDED
Binary file (557 Bytes). View file
 
wav2vec2_esp_15h/checkpoint-22500/scheduler.pt ADDED
Binary file (627 Bytes). View file
 
wav2vec2_esp_15h/checkpoint-22500/trainer_state.json ADDED
@@ -0,0 +1,691 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.45026178010471,
5
+ "global_step": 22500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.65,
12
+ "learning_rate": 4.94e-05,
13
+ "loss": 6.5844,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.65,
18
+ "eval_loss": 2.985063076019287,
19
+ "eval_runtime": 66.7207,
20
+ "eval_samples_per_second": 15.348,
21
+ "eval_steps_per_second": 1.918,
22
+ "eval_wer": 1.0,
23
+ "step": 500
24
+ },
25
+ {
26
+ "epoch": 1.31,
27
+ "learning_rate": 9.94e-05,
28
+ "loss": 2.1717,
29
+ "step": 1000
30
+ },
31
+ {
32
+ "epoch": 1.31,
33
+ "eval_loss": 1.0769672393798828,
34
+ "eval_runtime": 66.0484,
35
+ "eval_samples_per_second": 15.504,
36
+ "eval_steps_per_second": 1.938,
37
+ "eval_wer": 0.8259404828747895,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 1.96,
42
+ "learning_rate": 9.774635036496352e-05,
43
+ "loss": 1.0058,
44
+ "step": 1500
45
+ },
46
+ {
47
+ "epoch": 1.96,
48
+ "eval_loss": 0.7441346049308777,
49
+ "eval_runtime": 66.5848,
50
+ "eval_samples_per_second": 15.379,
51
+ "eval_steps_per_second": 1.922,
52
+ "eval_wer": 0.6882650196518809,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 2.62,
57
+ "learning_rate": 9.546532846715329e-05,
58
+ "loss": 0.7316,
59
+ "step": 2000
60
+ },
61
+ {
62
+ "epoch": 2.62,
63
+ "eval_loss": 0.6406537294387817,
64
+ "eval_runtime": 65.9141,
65
+ "eval_samples_per_second": 15.535,
66
+ "eval_steps_per_second": 1.942,
67
+ "eval_wer": 0.617630544637844,
68
+ "step": 2000
69
+ },
70
+ {
71
+ "epoch": 3.27,
72
+ "learning_rate": 9.318430656934307e-05,
73
+ "loss": 0.6249,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 3.27,
78
+ "eval_loss": 0.5814340114593506,
79
+ "eval_runtime": 66.4954,
80
+ "eval_samples_per_second": 15.4,
81
+ "eval_steps_per_second": 1.925,
82
+ "eval_wer": 0.5783267827063447,
83
+ "step": 2500
84
+ },
85
+ {
86
+ "epoch": 3.93,
87
+ "learning_rate": 9.090328467153285e-05,
88
+ "loss": 0.5541,
89
+ "step": 3000
90
+ },
91
+ {
92
+ "epoch": 3.93,
93
+ "eval_loss": 0.5387497544288635,
94
+ "eval_runtime": 66.9564,
95
+ "eval_samples_per_second": 15.294,
96
+ "eval_steps_per_second": 1.912,
97
+ "eval_wer": 0.5449747332959012,
98
+ "step": 3000
99
+ },
100
+ {
101
+ "epoch": 4.58,
102
+ "learning_rate": 8.862226277372263e-05,
103
+ "loss": 0.4727,
104
+ "step": 3500
105
+ },
106
+ {
107
+ "epoch": 4.58,
108
+ "eval_loss": 0.5657380223274231,
109
+ "eval_runtime": 66.1646,
110
+ "eval_samples_per_second": 15.477,
111
+ "eval_steps_per_second": 1.935,
112
+ "eval_wer": 0.5357664233576642,
113
+ "step": 3500
114
+ },
115
+ {
116
+ "epoch": 5.24,
117
+ "learning_rate": 8.634124087591242e-05,
118
+ "loss": 0.4444,
119
+ "step": 4000
120
+ },
121
+ {
122
+ "epoch": 5.24,
123
+ "eval_loss": 0.5406317710876465,
124
+ "eval_runtime": 65.9956,
125
+ "eval_samples_per_second": 15.516,
126
+ "eval_steps_per_second": 1.94,
127
+ "eval_wer": 0.5164514317798989,
128
+ "step": 4000
129
+ },
130
+ {
131
+ "epoch": 5.89,
132
+ "learning_rate": 8.406021897810219e-05,
133
+ "loss": 0.4057,
134
+ "step": 4500
135
+ },
136
+ {
137
+ "epoch": 5.89,
138
+ "eval_loss": 0.5271017551422119,
139
+ "eval_runtime": 66.6343,
140
+ "eval_samples_per_second": 15.367,
141
+ "eval_steps_per_second": 1.921,
142
+ "eval_wer": 0.5116226838854576,
143
+ "step": 4500
144
+ },
145
+ {
146
+ "epoch": 6.54,
147
+ "learning_rate": 8.177919708029196e-05,
148
+ "loss": 0.3663,
149
+ "step": 5000
150
+ },
151
+ {
152
+ "epoch": 6.54,
153
+ "eval_loss": 0.5105618238449097,
154
+ "eval_runtime": 66.576,
155
+ "eval_samples_per_second": 15.381,
156
+ "eval_steps_per_second": 1.923,
157
+ "eval_wer": 0.4793935991016283,
158
+ "step": 5000
159
+ },
160
+ {
161
+ "epoch": 7.2,
162
+ "learning_rate": 7.949817518248176e-05,
163
+ "loss": 0.3406,
164
+ "step": 5500
165
+ },
166
+ {
167
+ "epoch": 7.2,
168
+ "eval_loss": 0.5662369132041931,
169
+ "eval_runtime": 67.3476,
170
+ "eval_samples_per_second": 15.205,
171
+ "eval_steps_per_second": 1.901,
172
+ "eval_wer": 0.48568220101066817,
173
+ "step": 5500
174
+ },
175
+ {
176
+ "epoch": 7.85,
177
+ "learning_rate": 7.721715328467154e-05,
178
+ "loss": 0.3142,
179
+ "step": 6000
180
+ },
181
+ {
182
+ "epoch": 7.85,
183
+ "eval_loss": 0.5632680654525757,
184
+ "eval_runtime": 66.2027,
185
+ "eval_samples_per_second": 15.468,
186
+ "eval_steps_per_second": 1.933,
187
+ "eval_wer": 0.4879281302638967,
188
+ "step": 6000
189
+ },
190
+ {
191
+ "epoch": 8.51,
192
+ "learning_rate": 7.493613138686131e-05,
193
+ "loss": 0.3002,
194
+ "step": 6500
195
+ },
196
+ {
197
+ "epoch": 8.51,
198
+ "eval_loss": 0.5218114852905273,
199
+ "eval_runtime": 67.5136,
200
+ "eval_samples_per_second": 15.167,
201
+ "eval_steps_per_second": 1.896,
202
+ "eval_wer": 0.45659741718135877,
203
+ "step": 6500
204
+ },
205
+ {
206
+ "epoch": 9.16,
207
+ "learning_rate": 7.265967153284671e-05,
208
+ "loss": 0.2845,
209
+ "step": 7000
210
+ },
211
+ {
212
+ "epoch": 9.16,
213
+ "eval_loss": 0.5436348915100098,
214
+ "eval_runtime": 66.2104,
215
+ "eval_samples_per_second": 15.466,
216
+ "eval_steps_per_second": 1.933,
217
+ "eval_wer": 0.45457608085345313,
218
+ "step": 7000
219
+ },
220
+ {
221
+ "epoch": 9.82,
222
+ "learning_rate": 7.038321167883212e-05,
223
+ "loss": 0.2603,
224
+ "step": 7500
225
+ },
226
+ {
227
+ "epoch": 9.82,
228
+ "eval_loss": 0.5183758735656738,
229
+ "eval_runtime": 67.776,
230
+ "eval_samples_per_second": 15.109,
231
+ "eval_steps_per_second": 1.889,
232
+ "eval_wer": 0.4415496911847277,
233
+ "step": 7500
234
+ },
235
+ {
236
+ "epoch": 10.47,
237
+ "learning_rate": 6.810218978102189e-05,
238
+ "loss": 0.2438,
239
+ "step": 8000
240
+ },
241
+ {
242
+ "epoch": 10.47,
243
+ "eval_loss": 0.5246397256851196,
244
+ "eval_runtime": 66.5435,
245
+ "eval_samples_per_second": 15.388,
246
+ "eval_steps_per_second": 1.924,
247
+ "eval_wer": 0.43638405390230206,
248
+ "step": 8000
249
+ },
250
+ {
251
+ "epoch": 11.13,
252
+ "learning_rate": 6.582116788321169e-05,
253
+ "loss": 0.2396,
254
+ "step": 8500
255
+ },
256
+ {
257
+ "epoch": 11.13,
258
+ "eval_loss": 0.5663778781890869,
259
+ "eval_runtime": 66.6323,
260
+ "eval_samples_per_second": 15.368,
261
+ "eval_steps_per_second": 1.921,
262
+ "eval_wer": 0.4428972487366648,
263
+ "step": 8500
264
+ },
265
+ {
266
+ "epoch": 11.78,
267
+ "learning_rate": 6.354014598540147e-05,
268
+ "loss": 0.2229,
269
+ "step": 9000
270
+ },
271
+ {
272
+ "epoch": 11.78,
273
+ "eval_loss": 0.5550942420959473,
274
+ "eval_runtime": 65.2937,
275
+ "eval_samples_per_second": 15.683,
276
+ "eval_steps_per_second": 1.96,
277
+ "eval_wer": 0.4415496911847277,
278
+ "step": 9000
279
+ },
280
+ {
281
+ "epoch": 12.43,
282
+ "learning_rate": 6.125912408759124e-05,
283
+ "loss": 0.2088,
284
+ "step": 9500
285
+ },
286
+ {
287
+ "epoch": 12.43,
288
+ "eval_loss": 0.5503721833229065,
289
+ "eval_runtime": 66.2676,
290
+ "eval_samples_per_second": 15.453,
291
+ "eval_steps_per_second": 1.932,
292
+ "eval_wer": 0.4257158899494666,
293
+ "step": 9500
294
+ },
295
+ {
296
+ "epoch": 13.09,
297
+ "learning_rate": 5.897810218978103e-05,
298
+ "loss": 0.2042,
299
+ "step": 10000
300
+ },
301
+ {
302
+ "epoch": 13.09,
303
+ "eval_loss": 0.5646582245826721,
304
+ "eval_runtime": 66.4465,
305
+ "eval_samples_per_second": 15.411,
306
+ "eval_steps_per_second": 1.926,
307
+ "eval_wer": 0.4398652442448063,
308
+ "step": 10000
309
+ },
310
+ {
311
+ "epoch": 13.74,
312
+ "learning_rate": 5.669708029197081e-05,
313
+ "loss": 0.1975,
314
+ "step": 10500
315
+ },
316
+ {
317
+ "epoch": 13.74,
318
+ "eval_loss": 0.5424100160598755,
319
+ "eval_runtime": 66.9743,
320
+ "eval_samples_per_second": 15.289,
321
+ "eval_steps_per_second": 1.911,
322
+ "eval_wer": 0.4186412128017967,
323
+ "step": 10500
324
+ },
325
+ {
326
+ "epoch": 14.4,
327
+ "learning_rate": 5.441605839416058e-05,
328
+ "loss": 0.1856,
329
+ "step": 11000
330
+ },
331
+ {
332
+ "epoch": 14.4,
333
+ "eval_loss": 0.5583313703536987,
334
+ "eval_runtime": 66.5902,
335
+ "eval_samples_per_second": 15.378,
336
+ "eval_steps_per_second": 1.922,
337
+ "eval_wer": 0.42515440763615947,
338
+ "step": 11000
339
+ },
340
+ {
341
+ "epoch": 15.05,
342
+ "learning_rate": 5.213503649635036e-05,
343
+ "loss": 0.1872,
344
+ "step": 11500
345
+ },
346
+ {
347
+ "epoch": 15.05,
348
+ "eval_loss": 0.6118334531784058,
349
+ "eval_runtime": 66.3517,
350
+ "eval_samples_per_second": 15.433,
351
+ "eval_steps_per_second": 1.929,
352
+ "eval_wer": 0.433464345873105,
353
+ "step": 11500
354
+ },
355
+ {
356
+ "epoch": 15.71,
357
+ "learning_rate": 4.985401459854015e-05,
358
+ "loss": 0.1676,
359
+ "step": 12000
360
+ },
361
+ {
362
+ "epoch": 15.71,
363
+ "eval_loss": 0.6066869497299194,
364
+ "eval_runtime": 67.2768,
365
+ "eval_samples_per_second": 15.221,
366
+ "eval_steps_per_second": 1.903,
367
+ "eval_wer": 0.43627175743964064,
368
+ "step": 12000
369
+ },
370
+ {
371
+ "epoch": 16.36,
372
+ "learning_rate": 4.757299270072993e-05,
373
+ "loss": 0.1536,
374
+ "step": 12500
375
+ },
376
+ {
377
+ "epoch": 16.36,
378
+ "eval_loss": 0.5462669730186462,
379
+ "eval_runtime": 66.3968,
380
+ "eval_samples_per_second": 15.422,
381
+ "eval_steps_per_second": 1.928,
382
+ "eval_wer": 0.412914093206064,
383
+ "step": 12500
384
+ },
385
+ {
386
+ "epoch": 17.02,
387
+ "learning_rate": 4.5291970802919706e-05,
388
+ "loss": 0.1582,
389
+ "step": 13000
390
+ },
391
+ {
392
+ "epoch": 17.02,
393
+ "eval_loss": 0.5799742937088013,
394
+ "eval_runtime": 65.9177,
395
+ "eval_samples_per_second": 15.535,
396
+ "eval_steps_per_second": 1.942,
397
+ "eval_wer": 0.41021897810218977,
398
+ "step": 13000
399
+ },
400
+ {
401
+ "epoch": 17.67,
402
+ "learning_rate": 4.3015510948905114e-05,
403
+ "loss": 0.1429,
404
+ "step": 13500
405
+ },
406
+ {
407
+ "epoch": 17.67,
408
+ "eval_loss": 0.5899107456207275,
409
+ "eval_runtime": 66.1113,
410
+ "eval_samples_per_second": 15.489,
411
+ "eval_steps_per_second": 1.936,
412
+ "eval_wer": 0.4150477259966311,
413
+ "step": 13500
414
+ },
415
+ {
416
+ "epoch": 18.32,
417
+ "learning_rate": 4.0734489051094895e-05,
418
+ "loss": 0.1451,
419
+ "step": 14000
420
+ },
421
+ {
422
+ "epoch": 18.32,
423
+ "eval_loss": 0.6129721999168396,
424
+ "eval_runtime": 67.4516,
425
+ "eval_samples_per_second": 15.181,
426
+ "eval_steps_per_second": 1.898,
427
+ "eval_wer": 0.4138124649073554,
428
+ "step": 14000
429
+ },
430
+ {
431
+ "epoch": 18.98,
432
+ "learning_rate": 3.845346715328467e-05,
433
+ "loss": 0.1356,
434
+ "step": 14500
435
+ },
436
+ {
437
+ "epoch": 18.98,
438
+ "eval_loss": 0.6040089130401611,
439
+ "eval_runtime": 66.75,
440
+ "eval_samples_per_second": 15.341,
441
+ "eval_steps_per_second": 1.918,
442
+ "eval_wer": 0.4123526108927569,
443
+ "step": 14500
444
+ },
445
+ {
446
+ "epoch": 19.63,
447
+ "learning_rate": 3.617244525547445e-05,
448
+ "loss": 0.134,
449
+ "step": 15000
450
+ },
451
+ {
452
+ "epoch": 19.63,
453
+ "eval_loss": 0.5997006893157959,
454
+ "eval_runtime": 67.809,
455
+ "eval_samples_per_second": 15.101,
456
+ "eval_steps_per_second": 1.888,
457
+ "eval_wer": 0.40819764177428414,
458
+ "step": 15000
459
+ },
460
+ {
461
+ "epoch": 20.29,
462
+ "learning_rate": 3.389598540145986e-05,
463
+ "loss": 0.1305,
464
+ "step": 15500
465
+ },
466
+ {
467
+ "epoch": 20.29,
468
+ "eval_loss": 0.5769977569580078,
469
+ "eval_runtime": 66.9982,
470
+ "eval_samples_per_second": 15.284,
471
+ "eval_steps_per_second": 1.91,
472
+ "eval_wer": 0.40235822571588997,
473
+ "step": 15500
474
+ },
475
+ {
476
+ "epoch": 20.94,
477
+ "learning_rate": 3.161496350364963e-05,
478
+ "loss": 0.1222,
479
+ "step": 16000
480
+ },
481
+ {
482
+ "epoch": 20.94,
483
+ "eval_loss": 0.5980133414268494,
484
+ "eval_runtime": 66.1854,
485
+ "eval_samples_per_second": 15.472,
486
+ "eval_steps_per_second": 1.934,
487
+ "eval_wer": 0.391577765300393,
488
+ "step": 16000
489
+ },
490
+ {
491
+ "epoch": 21.6,
492
+ "learning_rate": 2.9333941605839417e-05,
493
+ "loss": 0.1214,
494
+ "step": 16500
495
+ },
496
+ {
497
+ "epoch": 21.6,
498
+ "eval_loss": 0.5917083024978638,
499
+ "eval_runtime": 67.5557,
500
+ "eval_samples_per_second": 15.158,
501
+ "eval_steps_per_second": 1.895,
502
+ "eval_wer": 0.4005614823133071,
503
+ "step": 16500
504
+ },
505
+ {
506
+ "epoch": 22.25,
507
+ "learning_rate": 2.70529197080292e-05,
508
+ "loss": 0.1184,
509
+ "step": 17000
510
+ },
511
+ {
512
+ "epoch": 22.25,
513
+ "eval_loss": 0.578889787197113,
514
+ "eval_runtime": 66.5407,
515
+ "eval_samples_per_second": 15.389,
516
+ "eval_steps_per_second": 1.924,
517
+ "eval_wer": 0.39460976979225154,
518
+ "step": 17000
519
+ },
520
+ {
521
+ "epoch": 22.91,
522
+ "learning_rate": 2.477189781021898e-05,
523
+ "loss": 0.1132,
524
+ "step": 17500
525
+ },
526
+ {
527
+ "epoch": 22.91,
528
+ "eval_loss": 0.5992549657821655,
529
+ "eval_runtime": 66.4759,
530
+ "eval_samples_per_second": 15.404,
531
+ "eval_steps_per_second": 1.926,
532
+ "eval_wer": 0.3892195395845031,
533
+ "step": 17500
534
+ },
535
+ {
536
+ "epoch": 23.56,
537
+ "learning_rate": 2.249087591240876e-05,
538
+ "loss": 0.1052,
539
+ "step": 18000
540
+ },
541
+ {
542
+ "epoch": 23.56,
543
+ "eval_loss": 0.5946719646453857,
544
+ "eval_runtime": 66.8036,
545
+ "eval_samples_per_second": 15.329,
546
+ "eval_steps_per_second": 1.916,
547
+ "eval_wer": 0.3889949466591802,
548
+ "step": 18000
549
+ },
550
+ {
551
+ "epoch": 24.21,
552
+ "learning_rate": 2.0209854014598544e-05,
553
+ "loss": 0.113,
554
+ "step": 18500
555
+ },
556
+ {
557
+ "epoch": 24.21,
558
+ "eval_loss": 0.6009370684623718,
559
+ "eval_runtime": 65.8523,
560
+ "eval_samples_per_second": 15.55,
561
+ "eval_steps_per_second": 1.944,
562
+ "eval_wer": 0.3841661987647389,
563
+ "step": 18500
564
+ },
565
+ {
566
+ "epoch": 24.87,
567
+ "learning_rate": 1.792883211678832e-05,
568
+ "loss": 0.1021,
569
+ "step": 19000
570
+ },
571
+ {
572
+ "epoch": 24.87,
573
+ "eval_loss": 0.6255775690078735,
574
+ "eval_runtime": 67.7689,
575
+ "eval_samples_per_second": 15.11,
576
+ "eval_steps_per_second": 1.889,
577
+ "eval_wer": 0.3841661987647389,
578
+ "step": 19000
579
+ },
580
+ {
581
+ "epoch": 25.52,
582
+ "learning_rate": 1.5652372262773723e-05,
583
+ "loss": 0.0966,
584
+ "step": 19500
585
+ },
586
+ {
587
+ "epoch": 25.52,
588
+ "eval_loss": 0.6213188171386719,
589
+ "eval_runtime": 68.5057,
590
+ "eval_samples_per_second": 14.948,
591
+ "eval_steps_per_second": 1.868,
592
+ "eval_wer": 0.3869736103312746,
593
+ "step": 19500
594
+ },
595
+ {
596
+ "epoch": 26.18,
597
+ "learning_rate": 1.3371350364963504e-05,
598
+ "loss": 0.0981,
599
+ "step": 20000
600
+ },
601
+ {
602
+ "epoch": 26.18,
603
+ "eval_loss": 0.6331676244735718,
604
+ "eval_runtime": 67.2276,
605
+ "eval_samples_per_second": 15.232,
606
+ "eval_steps_per_second": 1.904,
607
+ "eval_wer": 0.37866367209432905,
608
+ "step": 20000
609
+ },
610
+ {
611
+ "epoch": 26.83,
612
+ "learning_rate": 1.1090328467153285e-05,
613
+ "loss": 0.0983,
614
+ "step": 20500
615
+ },
616
+ {
617
+ "epoch": 26.83,
618
+ "eval_loss": 0.6225576996803284,
619
+ "eval_runtime": 66.5907,
620
+ "eval_samples_per_second": 15.378,
621
+ "eval_steps_per_second": 1.922,
622
+ "eval_wer": 0.3806850084222347,
623
+ "step": 20500
624
+ },
625
+ {
626
+ "epoch": 27.49,
627
+ "learning_rate": 8.809306569343067e-06,
628
+ "loss": 0.0915,
629
+ "step": 21000
630
+ },
631
+ {
632
+ "epoch": 27.49,
633
+ "eval_loss": 0.6176589727401733,
634
+ "eval_runtime": 66.1318,
635
+ "eval_samples_per_second": 15.484,
636
+ "eval_steps_per_second": 1.936,
637
+ "eval_wer": 0.3785513756316676,
638
+ "step": 21000
639
+ },
640
+ {
641
+ "epoch": 28.14,
642
+ "learning_rate": 6.532846715328468e-06,
643
+ "loss": 0.0898,
644
+ "step": 21500
645
+ },
646
+ {
647
+ "epoch": 28.14,
648
+ "eval_loss": 0.6252412796020508,
649
+ "eval_runtime": 66.2396,
650
+ "eval_samples_per_second": 15.459,
651
+ "eval_steps_per_second": 1.932,
652
+ "eval_wer": 0.3828186412128018,
653
+ "step": 21500
654
+ },
655
+ {
656
+ "epoch": 28.8,
657
+ "learning_rate": 4.251824817518248e-06,
658
+ "loss": 0.0916,
659
+ "step": 22000
660
+ },
661
+ {
662
+ "epoch": 28.8,
663
+ "eval_loss": 0.6163517832756042,
664
+ "eval_runtime": 66.1166,
665
+ "eval_samples_per_second": 15.488,
666
+ "eval_steps_per_second": 1.936,
667
+ "eval_wer": 0.3760808534531162,
668
+ "step": 22000
669
+ },
670
+ {
671
+ "epoch": 29.45,
672
+ "learning_rate": 1.9708029197080293e-06,
673
+ "loss": 0.0892,
674
+ "step": 22500
675
+ },
676
+ {
677
+ "epoch": 29.45,
678
+ "eval_loss": 0.6187598705291748,
679
+ "eval_runtime": 66.847,
680
+ "eval_samples_per_second": 15.319,
681
+ "eval_steps_per_second": 1.915,
682
+ "eval_wer": 0.37113980909601346,
683
+ "step": 22500
684
+ }
685
+ ],
686
+ "max_steps": 22920,
687
+ "num_train_epochs": 30,
688
+ "total_flos": 7.146961381638424e+18,
689
+ "trial_name": null,
690
+ "trial_params": null
691
+ }
wav2vec2_esp_15h/checkpoint-22500/training_args.bin ADDED
Binary file (3 kB). View file
 
wav2vec2_esp_15h/runs/Mar08_04-04-36_mint/1678216062.1216896/events.out.tfevents.1678216062.mint ADDED
Binary file (4.57 kB). View file
 
wav2vec2_esp_15h/runs/Mar08_04-04-36_mint/events.out.tfevents.1678216062.mint ADDED
Binary file (26.8 kB). View file
 
wav2vec2_esp_30h/checkpoint-43500/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
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": 0.1,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
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": 0.1,
52
+ "feat_quantizer_dropout": 0.0,
53
+ "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
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": 37,
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": "4.17.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 38,
116
+ "xvector_output_dim": 512
117
+ }
wav2vec2_esp_30h/checkpoint-43500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c364ebd0559d3e71a017c630c7a911334c73e55398298942b26f7225f946ca95
3
+ size 721723525
wav2vec2_esp_30h/checkpoint-43500/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
+ }
wav2vec2_esp_30h/checkpoint-43500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6cb1a25d6fbe0dd18ef64cd5a9e527e3675149fb972e3b51c522be7a4e0c969
3
+ size 377691873
wav2vec2_esp_30h/checkpoint-43500/rng_state.pth ADDED
Binary file (14.6 kB). View file
 
wav2vec2_esp_30h/checkpoint-43500/scaler.pt ADDED
Binary file (557 Bytes). View file
 
wav2vec2_esp_30h/checkpoint-43500/scheduler.pt ADDED
Binary file (627 Bytes). View file
 
wav2vec2_esp_30h/checkpoint-43500/trainer_state.json ADDED
@@ -0,0 +1,1321 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.61198093941457,
5
+ "global_step": 43500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.34,
12
+ "learning_rate": 4.9500000000000004e-05,
13
+ "loss": 6.2203,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.34,
18
+ "eval_loss": 3.002925157546997,
19
+ "eval_runtime": 132.7789,
20
+ "eval_samples_per_second": 14.942,
21
+ "eval_steps_per_second": 1.868,
22
+ "eval_wer": 1.0,
23
+ "step": 500
24
+ },
25
+ {
26
+ "epoch": 0.68,
27
+ "learning_rate": 9.95e-05,
28
+ "loss": 2.2102,
29
+ "step": 1000
30
+ },
31
+ {
32
+ "epoch": 0.68,
33
+ "eval_loss": 1.1446926593780518,
34
+ "eval_runtime": 132.5997,
35
+ "eval_samples_per_second": 14.962,
36
+ "eval_steps_per_second": 1.87,
37
+ "eval_wer": 0.8542348032028871,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 1.02,
42
+ "learning_rate": 9.885070814952404e-05,
43
+ "loss": 0.9714,
44
+ "step": 1500
45
+ },
46
+ {
47
+ "epoch": 1.02,
48
+ "eval_loss": 0.7160163521766663,
49
+ "eval_runtime": 132.7305,
50
+ "eval_samples_per_second": 14.948,
51
+ "eval_steps_per_second": 1.868,
52
+ "eval_wer": 0.6600879666177963,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 1.36,
57
+ "learning_rate": 9.76898072904574e-05,
58
+ "loss": 0.7541,
59
+ "step": 2000
60
+ },
61
+ {
62
+ "epoch": 1.36,
63
+ "eval_loss": 0.6261711120605469,
64
+ "eval_runtime": 132.7267,
65
+ "eval_samples_per_second": 14.948,
66
+ "eval_steps_per_second": 1.869,
67
+ "eval_wer": 0.594733280703733,
68
+ "step": 2000
69
+ },
70
+ {
71
+ "epoch": 1.7,
72
+ "learning_rate": 9.652890643139077e-05,
73
+ "loss": 0.6838,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 1.7,
78
+ "eval_loss": 0.5789812207221985,
79
+ "eval_runtime": 131.2782,
80
+ "eval_samples_per_second": 15.113,
81
+ "eval_steps_per_second": 1.889,
82
+ "eval_wer": 0.5645088530506372,
83
+ "step": 2500
84
+ },
85
+ {
86
+ "epoch": 2.04,
87
+ "learning_rate": 9.537032737404225e-05,
88
+ "loss": 0.6287,
89
+ "step": 3000
90
+ },
91
+ {
92
+ "epoch": 2.04,
93
+ "eval_loss": 0.6015335917472839,
94
+ "eval_runtime": 133.004,
95
+ "eval_samples_per_second": 14.917,
96
+ "eval_steps_per_second": 1.865,
97
+ "eval_wer": 0.5387955340024811,
98
+ "step": 3000
99
+ },
100
+ {
101
+ "epoch": 2.38,
102
+ "learning_rate": 9.420942651497563e-05,
103
+ "loss": 0.5439,
104
+ "step": 3500
105
+ },
106
+ {
107
+ "epoch": 2.38,
108
+ "eval_loss": 0.5301734209060669,
109
+ "eval_runtime": 133.1151,
110
+ "eval_samples_per_second": 14.904,
111
+ "eval_steps_per_second": 1.863,
112
+ "eval_wer": 0.5076124957708357,
113
+ "step": 3500
114
+ },
115
+ {
116
+ "epoch": 2.72,
117
+ "learning_rate": 9.304852565590899e-05,
118
+ "loss": 0.5279,
119
+ "step": 4000
120
+ },
121
+ {
122
+ "epoch": 2.72,
123
+ "eval_loss": 0.5215476155281067,
124
+ "eval_runtime": 131.1447,
125
+ "eval_samples_per_second": 15.128,
126
+ "eval_steps_per_second": 1.891,
127
+ "eval_wer": 0.48962445020863876,
128
+ "step": 4000
129
+ },
130
+ {
131
+ "epoch": 3.06,
132
+ "learning_rate": 9.188762479684235e-05,
133
+ "loss": 0.5006,
134
+ "step": 4500
135
+ },
136
+ {
137
+ "epoch": 3.06,
138
+ "eval_loss": 0.4860161542892456,
139
+ "eval_runtime": 133.2608,
140
+ "eval_samples_per_second": 14.888,
141
+ "eval_steps_per_second": 1.861,
142
+ "eval_wer": 0.4685350174805458,
143
+ "step": 4500
144
+ },
145
+ {
146
+ "epoch": 3.4,
147
+ "learning_rate": 9.072672393777572e-05,
148
+ "loss": 0.4432,
149
+ "step": 5000
150
+ },
151
+ {
152
+ "epoch": 3.4,
153
+ "eval_loss": 0.484553724527359,
154
+ "eval_runtime": 133.4845,
155
+ "eval_samples_per_second": 14.863,
156
+ "eval_steps_per_second": 1.858,
157
+ "eval_wer": 0.46864779519566935,
158
+ "step": 5000
159
+ },
160
+ {
161
+ "epoch": 3.74,
162
+ "learning_rate": 8.956582307870908e-05,
163
+ "loss": 0.4334,
164
+ "step": 5500
165
+ },
166
+ {
167
+ "epoch": 3.74,
168
+ "eval_loss": 0.477468878030777,
169
+ "eval_runtime": 132.8856,
170
+ "eval_samples_per_second": 14.93,
171
+ "eval_steps_per_second": 1.866,
172
+ "eval_wer": 0.45415585880230064,
173
+ "step": 5500
174
+ },
175
+ {
176
+ "epoch": 4.08,
177
+ "learning_rate": 8.840492221964245e-05,
178
+ "loss": 0.4292,
179
+ "step": 6000
180
+ },
181
+ {
182
+ "epoch": 4.08,
183
+ "eval_loss": 0.451526939868927,
184
+ "eval_runtime": 132.3623,
185
+ "eval_samples_per_second": 14.989,
186
+ "eval_steps_per_second": 1.874,
187
+ "eval_wer": 0.4291192060448855,
188
+ "step": 6000
189
+ },
190
+ {
191
+ "epoch": 4.42,
192
+ "learning_rate": 8.724402136057582e-05,
193
+ "loss": 0.3779,
194
+ "step": 6500
195
+ },
196
+ {
197
+ "epoch": 4.42,
198
+ "eval_loss": 0.4495590329170227,
199
+ "eval_runtime": 132.6275,
200
+ "eval_samples_per_second": 14.959,
201
+ "eval_steps_per_second": 1.87,
202
+ "eval_wer": 0.42054809969550017,
203
+ "step": 6500
204
+ },
205
+ {
206
+ "epoch": 4.77,
207
+ "learning_rate": 8.608312050150917e-05,
208
+ "loss": 0.3783,
209
+ "step": 7000
210
+ },
211
+ {
212
+ "epoch": 4.77,
213
+ "eval_loss": 0.45758938789367676,
214
+ "eval_runtime": 132.4666,
215
+ "eval_samples_per_second": 14.977,
216
+ "eval_steps_per_second": 1.872,
217
+ "eval_wer": 0.4184053231081538,
218
+ "step": 7000
219
+ },
220
+ {
221
+ "epoch": 5.11,
222
+ "learning_rate": 8.492221964244254e-05,
223
+ "loss": 0.3622,
224
+ "step": 7500
225
+ },
226
+ {
227
+ "epoch": 5.11,
228
+ "eval_loss": 0.4782721698284149,
229
+ "eval_runtime": 133.2094,
230
+ "eval_samples_per_second": 14.894,
231
+ "eval_steps_per_second": 1.862,
232
+ "eval_wer": 0.4070147738806812,
233
+ "step": 7500
234
+ },
235
+ {
236
+ "epoch": 5.45,
237
+ "learning_rate": 8.37613187833759e-05,
238
+ "loss": 0.3278,
239
+ "step": 8000
240
+ },
241
+ {
242
+ "epoch": 5.45,
243
+ "eval_loss": 0.44267573952674866,
244
+ "eval_runtime": 133.2221,
245
+ "eval_samples_per_second": 14.892,
246
+ "eval_steps_per_second": 1.862,
247
+ "eval_wer": 0.40278560956355025,
248
+ "step": 8000
249
+ },
250
+ {
251
+ "epoch": 5.79,
252
+ "learning_rate": 8.260041792430927e-05,
253
+ "loss": 0.3304,
254
+ "step": 8500
255
+ },
256
+ {
257
+ "epoch": 5.79,
258
+ "eval_loss": 0.4482724368572235,
259
+ "eval_runtime": 132.3395,
260
+ "eval_samples_per_second": 14.992,
261
+ "eval_steps_per_second": 1.874,
262
+ "eval_wer": 0.4056050524416375,
263
+ "step": 8500
264
+ },
265
+ {
266
+ "epoch": 6.13,
267
+ "learning_rate": 8.143951706524264e-05,
268
+ "loss": 0.312,
269
+ "step": 9000
270
+ },
271
+ {
272
+ "epoch": 6.13,
273
+ "eval_loss": 0.4750816524028778,
274
+ "eval_runtime": 132.887,
275
+ "eval_samples_per_second": 14.93,
276
+ "eval_steps_per_second": 1.866,
277
+ "eval_wer": 0.3882936731701816,
278
+ "step": 9000
279
+ },
280
+ {
281
+ "epoch": 6.47,
282
+ "learning_rate": 8.027861620617599e-05,
283
+ "loss": 0.29,
284
+ "step": 9500
285
+ },
286
+ {
287
+ "epoch": 6.47,
288
+ "eval_loss": 0.4528858959674835,
289
+ "eval_runtime": 133.211,
290
+ "eval_samples_per_second": 14.894,
291
+ "eval_steps_per_second": 1.862,
292
+ "eval_wer": 0.3780872899515056,
293
+ "step": 9500
294
+ },
295
+ {
296
+ "epoch": 6.81,
297
+ "learning_rate": 7.912235895054563e-05,
298
+ "loss": 0.3057,
299
+ "step": 10000
300
+ },
301
+ {
302
+ "epoch": 6.81,
303
+ "eval_loss": 0.5861864686012268,
304
+ "eval_runtime": 133.2534,
305
+ "eval_samples_per_second": 14.889,
306
+ "eval_steps_per_second": 1.861,
307
+ "eval_wer": 0.3783128453817526,
308
+ "step": 10000
309
+ },
310
+ {
311
+ "epoch": 7.15,
312
+ "learning_rate": 7.796145809147898e-05,
313
+ "loss": 0.2971,
314
+ "step": 10500
315
+ },
316
+ {
317
+ "epoch": 7.15,
318
+ "eval_loss": 0.43457281589508057,
319
+ "eval_runtime": 132.5985,
320
+ "eval_samples_per_second": 14.962,
321
+ "eval_steps_per_second": 1.87,
322
+ "eval_wer": 0.3765084019397767,
323
+ "step": 10500
324
+ },
325
+ {
326
+ "epoch": 7.49,
327
+ "learning_rate": 7.680055723241235e-05,
328
+ "loss": 0.2684,
329
+ "step": 11000
330
+ },
331
+ {
332
+ "epoch": 7.49,
333
+ "eval_loss": 0.4560734033584595,
334
+ "eval_runtime": 132.2037,
335
+ "eval_samples_per_second": 15.007,
336
+ "eval_steps_per_second": 1.876,
337
+ "eval_wer": 0.3732942370587572,
338
+ "step": 11000
339
+ },
340
+ {
341
+ "epoch": 7.83,
342
+ "learning_rate": 7.563965637334573e-05,
343
+ "loss": 0.2622,
344
+ "step": 11500
345
+ },
346
+ {
347
+ "epoch": 7.83,
348
+ "eval_loss": 0.43240657448768616,
349
+ "eval_runtime": 133.3517,
350
+ "eval_samples_per_second": 14.878,
351
+ "eval_steps_per_second": 1.86,
352
+ "eval_wer": 0.37408368106462164,
353
+ "step": 11500
354
+ },
355
+ {
356
+ "epoch": 8.17,
357
+ "learning_rate": 7.447875551427908e-05,
358
+ "loss": 0.2635,
359
+ "step": 12000
360
+ },
361
+ {
362
+ "epoch": 8.17,
363
+ "eval_loss": 0.4556463658809662,
364
+ "eval_runtime": 133.4279,
365
+ "eval_samples_per_second": 14.869,
366
+ "eval_steps_per_second": 1.859,
367
+ "eval_wer": 0.3789895116724935,
368
+ "step": 12000
369
+ },
370
+ {
371
+ "epoch": 8.51,
372
+ "learning_rate": 7.331785465521245e-05,
373
+ "loss": 0.2363,
374
+ "step": 12500
375
+ },
376
+ {
377
+ "epoch": 8.51,
378
+ "eval_loss": 0.47497859597206116,
379
+ "eval_runtime": 134.0404,
380
+ "eval_samples_per_second": 14.802,
381
+ "eval_steps_per_second": 1.85,
382
+ "eval_wer": 0.3709822938987256,
383
+ "step": 12500
384
+ },
385
+ {
386
+ "epoch": 8.85,
387
+ "learning_rate": 7.215695379614582e-05,
388
+ "loss": 0.2516,
389
+ "step": 13000
390
+ },
391
+ {
392
+ "epoch": 8.85,
393
+ "eval_loss": 0.4297301471233368,
394
+ "eval_runtime": 133.4288,
395
+ "eval_samples_per_second": 14.869,
396
+ "eval_steps_per_second": 1.859,
397
+ "eval_wer": 0.3635389647005752,
398
+ "step": 13000
399
+ },
400
+ {
401
+ "epoch": 9.19,
402
+ "learning_rate": 7.099605293707917e-05,
403
+ "loss": 0.2291,
404
+ "step": 13500
405
+ },
406
+ {
407
+ "epoch": 9.19,
408
+ "eval_loss": 0.4862041175365448,
409
+ "eval_runtime": 133.7259,
410
+ "eval_samples_per_second": 14.836,
411
+ "eval_steps_per_second": 1.855,
412
+ "eval_wer": 0.36021202210443215,
413
+ "step": 13500
414
+ },
415
+ {
416
+ "epoch": 9.53,
417
+ "learning_rate": 6.983515207801255e-05,
418
+ "loss": 0.2266,
419
+ "step": 14000
420
+ },
421
+ {
422
+ "epoch": 9.53,
423
+ "eval_loss": 0.4703396260738373,
424
+ "eval_runtime": 133.0987,
425
+ "eval_samples_per_second": 14.906,
426
+ "eval_steps_per_second": 1.863,
427
+ "eval_wer": 0.3546295252058193,
428
+ "step": 14000
429
+ },
430
+ {
431
+ "epoch": 9.87,
432
+ "learning_rate": 6.867657302066403e-05,
433
+ "loss": 0.2281,
434
+ "step": 14500
435
+ },
436
+ {
437
+ "epoch": 9.87,
438
+ "eval_loss": 0.42882266640663147,
439
+ "eval_runtime": 133.4534,
440
+ "eval_samples_per_second": 14.867,
441
+ "eval_steps_per_second": 1.858,
442
+ "eval_wer": 0.35812563437464756,
443
+ "step": 14500
444
+ },
445
+ {
446
+ "epoch": 10.21,
447
+ "learning_rate": 6.75156721615974e-05,
448
+ "loss": 0.2126,
449
+ "step": 15000
450
+ },
451
+ {
452
+ "epoch": 10.21,
453
+ "eval_loss": 0.4615735113620758,
454
+ "eval_runtime": 132.7985,
455
+ "eval_samples_per_second": 14.94,
456
+ "eval_steps_per_second": 1.867,
457
+ "eval_wer": 0.35626480207510997,
458
+ "step": 15000
459
+ },
460
+ {
461
+ "epoch": 10.55,
462
+ "learning_rate": 6.635477130253078e-05,
463
+ "loss": 0.217,
464
+ "step": 15500
465
+ },
466
+ {
467
+ "epoch": 10.55,
468
+ "eval_loss": 0.49455586075782776,
469
+ "eval_runtime": 133.6528,
470
+ "eval_samples_per_second": 14.844,
471
+ "eval_steps_per_second": 1.856,
472
+ "eval_wer": 0.34560730799594,
473
+ "step": 15500
474
+ },
475
+ {
476
+ "epoch": 10.89,
477
+ "learning_rate": 6.519387044346413e-05,
478
+ "loss": 0.212,
479
+ "step": 16000
480
+ },
481
+ {
482
+ "epoch": 10.89,
483
+ "eval_loss": 0.43323996663093567,
484
+ "eval_runtime": 133.4419,
485
+ "eval_samples_per_second": 14.868,
486
+ "eval_steps_per_second": 1.858,
487
+ "eval_wer": 0.3452689748505695,
488
+ "step": 16000
489
+ },
490
+ {
491
+ "epoch": 11.23,
492
+ "learning_rate": 6.40329695843975e-05,
493
+ "loss": 0.1986,
494
+ "step": 16500
495
+ },
496
+ {
497
+ "epoch": 11.23,
498
+ "eval_loss": 0.4868086576461792,
499
+ "eval_runtime": 131.4603,
500
+ "eval_samples_per_second": 15.092,
501
+ "eval_steps_per_second": 1.887,
502
+ "eval_wer": 0.3399684222397654,
503
+ "step": 16500
504
+ },
505
+ {
506
+ "epoch": 11.57,
507
+ "learning_rate": 6.2874390527049e-05,
508
+ "loss": 0.2012,
509
+ "step": 17000
510
+ },
511
+ {
512
+ "epoch": 11.57,
513
+ "eval_loss": 0.4474054276943207,
514
+ "eval_runtime": 133.9018,
515
+ "eval_samples_per_second": 14.817,
516
+ "eval_steps_per_second": 1.852,
517
+ "eval_wer": 0.3460020299988722,
518
+ "step": 17000
519
+ },
520
+ {
521
+ "epoch": 11.91,
522
+ "learning_rate": 6.171348966798236e-05,
523
+ "loss": 0.1998,
524
+ "step": 17500
525
+ },
526
+ {
527
+ "epoch": 11.91,
528
+ "eval_loss": 0.45011407136917114,
529
+ "eval_runtime": 132.7933,
530
+ "eval_samples_per_second": 14.941,
531
+ "eval_steps_per_second": 1.868,
532
+ "eval_wer": 0.3362467576406902,
533
+ "step": 17500
534
+ },
535
+ {
536
+ "epoch": 12.25,
537
+ "learning_rate": 6.055258880891572e-05,
538
+ "loss": 0.1746,
539
+ "step": 18000
540
+ },
541
+ {
542
+ "epoch": 12.25,
543
+ "eval_loss": 0.4731091558933258,
544
+ "eval_runtime": 133.9219,
545
+ "eval_samples_per_second": 14.815,
546
+ "eval_steps_per_second": 1.852,
547
+ "eval_wer": 0.33630314649825194,
548
+ "step": 18000
549
+ },
550
+ {
551
+ "epoch": 12.59,
552
+ "learning_rate": 5.9391687949849086e-05,
553
+ "loss": 0.1805,
554
+ "step": 18500
555
+ },
556
+ {
557
+ "epoch": 12.59,
558
+ "eval_loss": 0.4601946771144867,
559
+ "eval_runtime": 132.0532,
560
+ "eval_samples_per_second": 15.024,
561
+ "eval_steps_per_second": 1.878,
562
+ "eval_wer": 0.33743092364948685,
563
+ "step": 18500
564
+ },
565
+ {
566
+ "epoch": 12.93,
567
+ "learning_rate": 5.823310889250059e-05,
568
+ "loss": 0.1826,
569
+ "step": 19000
570
+ },
571
+ {
572
+ "epoch": 12.93,
573
+ "eval_loss": 0.47728002071380615,
574
+ "eval_runtime": 133.9362,
575
+ "eval_samples_per_second": 14.813,
576
+ "eval_steps_per_second": 1.852,
577
+ "eval_wer": 0.3276192624337431,
578
+ "step": 19000
579
+ },
580
+ {
581
+ "epoch": 13.27,
582
+ "learning_rate": 5.707220803343395e-05,
583
+ "loss": 0.1651,
584
+ "step": 19500
585
+ },
586
+ {
587
+ "epoch": 13.27,
588
+ "eval_loss": 0.471328467130661,
589
+ "eval_runtime": 111.9462,
590
+ "eval_samples_per_second": 17.723,
591
+ "eval_steps_per_second": 2.215,
592
+ "eval_wer": 0.3304387053118304,
593
+ "step": 19500
594
+ },
595
+ {
596
+ "epoch": 13.61,
597
+ "learning_rate": 5.591130717436731e-05,
598
+ "loss": 0.1812,
599
+ "step": 20000
600
+ },
601
+ {
602
+ "epoch": 13.61,
603
+ "eval_loss": 0.4280690550804138,
604
+ "eval_runtime": 110.8492,
605
+ "eval_samples_per_second": 17.898,
606
+ "eval_steps_per_second": 2.237,
607
+ "eval_wer": 0.32276982068343296,
608
+ "step": 20000
609
+ },
610
+ {
611
+ "epoch": 13.96,
612
+ "learning_rate": 5.4750406315300676e-05,
613
+ "loss": 0.1666,
614
+ "step": 20500
615
+ },
616
+ {
617
+ "epoch": 13.96,
618
+ "eval_loss": 0.4796316623687744,
619
+ "eval_runtime": 110.5982,
620
+ "eval_samples_per_second": 17.939,
621
+ "eval_steps_per_second": 2.242,
622
+ "eval_wer": 0.32530731927371154,
623
+ "step": 20500
624
+ },
625
+ {
626
+ "epoch": 14.3,
627
+ "learning_rate": 5.359182725795218e-05,
628
+ "loss": 0.1553,
629
+ "step": 21000
630
+ },
631
+ {
632
+ "epoch": 14.3,
633
+ "eval_loss": 0.4721369445323944,
634
+ "eval_runtime": 128.2013,
635
+ "eval_samples_per_second": 15.476,
636
+ "eval_steps_per_second": 1.934,
637
+ "eval_wer": 0.3259275967068907,
638
+ "step": 21000
639
+ },
640
+ {
641
+ "epoch": 14.64,
642
+ "learning_rate": 5.243092639888554e-05,
643
+ "loss": 0.1545,
644
+ "step": 21500
645
+ },
646
+ {
647
+ "epoch": 14.64,
648
+ "eval_loss": 0.4527774453163147,
649
+ "eval_runtime": 132.8486,
650
+ "eval_samples_per_second": 14.934,
651
+ "eval_steps_per_second": 1.867,
652
+ "eval_wer": 0.32682981842787867,
653
+ "step": 21500
654
+ },
655
+ {
656
+ "epoch": 14.98,
657
+ "learning_rate": 5.12700255398189e-05,
658
+ "loss": 0.1576,
659
+ "step": 22000
660
+ },
661
+ {
662
+ "epoch": 14.98,
663
+ "eval_loss": 0.4553755819797516,
664
+ "eval_runtime": 131.6613,
665
+ "eval_samples_per_second": 15.069,
666
+ "eval_steps_per_second": 1.884,
667
+ "eval_wer": 0.3252509304161498,
668
+ "step": 22000
669
+ },
670
+ {
671
+ "epoch": 15.32,
672
+ "learning_rate": 5.0109124680752265e-05,
673
+ "loss": 0.1511,
674
+ "step": 22500
675
+ },
676
+ {
677
+ "epoch": 15.32,
678
+ "eval_loss": 0.4580444097518921,
679
+ "eval_runtime": 134.2974,
680
+ "eval_samples_per_second": 14.773,
681
+ "eval_steps_per_second": 1.847,
682
+ "eval_wer": 0.3179203789331228,
683
+ "step": 22500
684
+ },
685
+ {
686
+ "epoch": 15.66,
687
+ "learning_rate": 4.895054562340376e-05,
688
+ "loss": 0.1444,
689
+ "step": 23000
690
+ },
691
+ {
692
+ "epoch": 15.66,
693
+ "eval_loss": 0.4659278094768524,
694
+ "eval_runtime": 134.6637,
695
+ "eval_samples_per_second": 14.733,
696
+ "eval_steps_per_second": 1.842,
697
+ "eval_wer": 0.32119093267170407,
698
+ "step": 23000
699
+ },
700
+ {
701
+ "epoch": 16.0,
702
+ "learning_rate": 4.7791966566055266e-05,
703
+ "loss": 0.1496,
704
+ "step": 23500
705
+ },
706
+ {
707
+ "epoch": 16.0,
708
+ "eval_loss": 0.4660351872444153,
709
+ "eval_runtime": 133.4382,
710
+ "eval_samples_per_second": 14.868,
711
+ "eval_steps_per_second": 1.859,
712
+ "eval_wer": 0.32598398556445246,
713
+ "step": 23500
714
+ },
715
+ {
716
+ "epoch": 16.34,
717
+ "learning_rate": 4.6631065706988625e-05,
718
+ "loss": 0.1327,
719
+ "step": 24000
720
+ },
721
+ {
722
+ "epoch": 16.34,
723
+ "eval_loss": 0.4934828281402588,
724
+ "eval_runtime": 134.3743,
725
+ "eval_samples_per_second": 14.765,
726
+ "eval_steps_per_second": 1.846,
727
+ "eval_wer": 0.31825871207849327,
728
+ "step": 24000
729
+ },
730
+ {
731
+ "epoch": 16.68,
732
+ "learning_rate": 4.547016484792199e-05,
733
+ "loss": 0.1535,
734
+ "step": 24500
735
+ },
736
+ {
737
+ "epoch": 16.68,
738
+ "eval_loss": 0.46466243267059326,
739
+ "eval_runtime": 132.9271,
740
+ "eval_samples_per_second": 14.925,
741
+ "eval_steps_per_second": 1.866,
742
+ "eval_wer": 0.3113792714559603,
743
+ "step": 24500
744
+ },
745
+ {
746
+ "epoch": 17.02,
747
+ "learning_rate": 4.430926398885535e-05,
748
+ "loss": 0.1438,
749
+ "step": 25000
750
+ },
751
+ {
752
+ "epoch": 17.02,
753
+ "eval_loss": 0.5053276419639587,
754
+ "eval_runtime": 132.7443,
755
+ "eval_samples_per_second": 14.946,
756
+ "eval_steps_per_second": 1.868,
757
+ "eval_wer": 0.31487538062478854,
758
+ "step": 25000
759
+ },
760
+ {
761
+ "epoch": 17.36,
762
+ "learning_rate": 4.314836312978872e-05,
763
+ "loss": 0.1264,
764
+ "step": 25500
765
+ },
766
+ {
767
+ "epoch": 17.36,
768
+ "eval_loss": 0.48962870240211487,
769
+ "eval_runtime": 133.9594,
770
+ "eval_samples_per_second": 14.81,
771
+ "eval_steps_per_second": 1.851,
772
+ "eval_wer": 0.3131273260403744,
773
+ "step": 25500
774
+ },
775
+ {
776
+ "epoch": 17.7,
777
+ "learning_rate": 4.1987462270722084e-05,
778
+ "loss": 0.1269,
779
+ "step": 26000
780
+ },
781
+ {
782
+ "epoch": 17.7,
783
+ "eval_loss": 0.46815410256385803,
784
+ "eval_runtime": 132.7946,
785
+ "eval_samples_per_second": 14.94,
786
+ "eval_steps_per_second": 1.868,
787
+ "eval_wer": 0.30748844028419986,
788
+ "step": 26000
789
+ },
790
+ {
791
+ "epoch": 18.04,
792
+ "learning_rate": 4.0826561411655444e-05,
793
+ "loss": 0.1272,
794
+ "step": 26500
795
+ },
796
+ {
797
+ "epoch": 18.04,
798
+ "eval_loss": 0.4759911596775055,
799
+ "eval_runtime": 132.8376,
800
+ "eval_samples_per_second": 14.936,
801
+ "eval_steps_per_second": 1.867,
802
+ "eval_wer": 0.3113792714559603,
803
+ "step": 26500
804
+ },
805
+ {
806
+ "epoch": 18.38,
807
+ "learning_rate": 3.966566055258881e-05,
808
+ "loss": 0.1219,
809
+ "step": 27000
810
+ },
811
+ {
812
+ "epoch": 18.38,
813
+ "eval_loss": 0.49612942337989807,
814
+ "eval_runtime": 133.1179,
815
+ "eval_samples_per_second": 14.904,
816
+ "eval_steps_per_second": 1.863,
817
+ "eval_wer": 0.31487538062478854,
818
+ "step": 27000
819
+ },
820
+ {
821
+ "epoch": 18.72,
822
+ "learning_rate": 3.850475969352218e-05,
823
+ "loss": 0.123,
824
+ "step": 27500
825
+ },
826
+ {
827
+ "epoch": 18.72,
828
+ "eval_loss": 0.46862220764160156,
829
+ "eval_runtime": 110.6681,
830
+ "eval_samples_per_second": 17.927,
831
+ "eval_steps_per_second": 2.241,
832
+ "eval_wer": 0.3072064959963911,
833
+ "step": 27500
834
+ },
835
+ {
836
+ "epoch": 19.06,
837
+ "learning_rate": 3.734385883445554e-05,
838
+ "loss": 0.1262,
839
+ "step": 28000
840
+ },
841
+ {
842
+ "epoch": 19.06,
843
+ "eval_loss": 0.49365413188934326,
844
+ "eval_runtime": 109.7879,
845
+ "eval_samples_per_second": 18.071,
846
+ "eval_steps_per_second": 2.259,
847
+ "eval_wer": 0.30889816172324347,
848
+ "step": 28000
849
+ },
850
+ {
851
+ "epoch": 19.4,
852
+ "learning_rate": 3.61829579753889e-05,
853
+ "loss": 0.1165,
854
+ "step": 28500
855
+ },
856
+ {
857
+ "epoch": 19.4,
858
+ "eval_loss": 0.4988892376422882,
859
+ "eval_runtime": 110.8122,
860
+ "eval_samples_per_second": 17.904,
861
+ "eval_steps_per_second": 2.238,
862
+ "eval_wer": 0.30540205255441527,
863
+ "step": 28500
864
+ },
865
+ {
866
+ "epoch": 19.74,
867
+ "learning_rate": 3.50243789180404e-05,
868
+ "loss": 0.1213,
869
+ "step": 29000
870
+ },
871
+ {
872
+ "epoch": 19.74,
873
+ "eval_loss": 0.4937494993209839,
874
+ "eval_runtime": 109.7673,
875
+ "eval_samples_per_second": 18.075,
876
+ "eval_steps_per_second": 2.259,
877
+ "eval_wer": 0.30404871997293337,
878
+ "step": 29000
879
+ },
880
+ {
881
+ "epoch": 20.08,
882
+ "learning_rate": 3.3863478058973766e-05,
883
+ "loss": 0.1113,
884
+ "step": 29500
885
+ },
886
+ {
887
+ "epoch": 20.08,
888
+ "eval_loss": 0.5031572580337524,
889
+ "eval_runtime": 111.1946,
890
+ "eval_samples_per_second": 17.843,
891
+ "eval_steps_per_second": 2.23,
892
+ "eval_wer": 0.3037667756851246,
893
+ "step": 29500
894
+ },
895
+ {
896
+ "epoch": 20.42,
897
+ "learning_rate": 3.270257719990713e-05,
898
+ "loss": 0.1087,
899
+ "step": 30000
900
+ },
901
+ {
902
+ "epoch": 20.42,
903
+ "eval_loss": 0.49494612216949463,
904
+ "eval_runtime": 126.8851,
905
+ "eval_samples_per_second": 15.636,
906
+ "eval_steps_per_second": 1.955,
907
+ "eval_wer": 0.299030111649938,
908
+ "step": 30000
909
+ },
910
+ {
911
+ "epoch": 20.76,
912
+ "learning_rate": 3.154167634084049e-05,
913
+ "loss": 0.1128,
914
+ "step": 30500
915
+ },
916
+ {
917
+ "epoch": 20.76,
918
+ "eval_loss": 0.49934807419776917,
919
+ "eval_runtime": 128.8495,
920
+ "eval_samples_per_second": 15.398,
921
+ "eval_steps_per_second": 1.925,
922
+ "eval_wer": 0.3017367768129018,
923
+ "step": 30500
924
+ },
925
+ {
926
+ "epoch": 21.1,
927
+ "learning_rate": 3.038077548177386e-05,
928
+ "loss": 0.1151,
929
+ "step": 31000
930
+ },
931
+ {
932
+ "epoch": 21.1,
933
+ "eval_loss": 0.5088058114051819,
934
+ "eval_runtime": 128.5026,
935
+ "eval_samples_per_second": 15.439,
936
+ "eval_steps_per_second": 1.93,
937
+ "eval_wer": 0.30184955452802525,
938
+ "step": 31000
939
+ },
940
+ {
941
+ "epoch": 21.44,
942
+ "learning_rate": 2.922219642442536e-05,
943
+ "loss": 0.1025,
944
+ "step": 31500
945
+ },
946
+ {
947
+ "epoch": 21.44,
948
+ "eval_loss": 0.47611942887306213,
949
+ "eval_runtime": 111.1648,
950
+ "eval_samples_per_second": 17.847,
951
+ "eval_steps_per_second": 2.231,
952
+ "eval_wer": 0.29891733393481446,
953
+ "step": 31500
954
+ },
955
+ {
956
+ "epoch": 21.78,
957
+ "learning_rate": 2.8063617367076856e-05,
958
+ "loss": 0.1016,
959
+ "step": 32000
960
+ },
961
+ {
962
+ "epoch": 21.78,
963
+ "eval_loss": 0.5126113891601562,
964
+ "eval_runtime": 121.5373,
965
+ "eval_samples_per_second": 16.324,
966
+ "eval_steps_per_second": 2.041,
967
+ "eval_wer": 0.30275177624901317,
968
+ "step": 32000
969
+ },
970
+ {
971
+ "epoch": 22.12,
972
+ "learning_rate": 2.690271650801022e-05,
973
+ "loss": 0.0991,
974
+ "step": 32500
975
+ },
976
+ {
977
+ "epoch": 22.12,
978
+ "eval_loss": 0.5214097499847412,
979
+ "eval_runtime": 129.4856,
980
+ "eval_samples_per_second": 15.322,
981
+ "eval_steps_per_second": 1.915,
982
+ "eval_wer": 0.2958159467689185,
983
+ "step": 32500
984
+ },
985
+ {
986
+ "epoch": 22.46,
987
+ "learning_rate": 2.574181564894358e-05,
988
+ "loss": 0.0951,
989
+ "step": 33000
990
+ },
991
+ {
992
+ "epoch": 22.46,
993
+ "eval_loss": 0.5233449339866638,
994
+ "eval_runtime": 130.9955,
995
+ "eval_samples_per_second": 15.146,
996
+ "eval_steps_per_second": 1.893,
997
+ "eval_wer": 0.2923198376000902,
998
+ "step": 33000
999
+ },
1000
+ {
1001
+ "epoch": 22.8,
1002
+ "learning_rate": 2.4580914789876948e-05,
1003
+ "loss": 0.1049,
1004
+ "step": 33500
1005
+ },
1006
+ {
1007
+ "epoch": 22.8,
1008
+ "eval_loss": 0.48444411158561707,
1009
+ "eval_runtime": 110.8951,
1010
+ "eval_samples_per_second": 17.891,
1011
+ "eval_steps_per_second": 2.236,
1012
+ "eval_wer": 0.29677455734746816,
1013
+ "step": 33500
1014
+ },
1015
+ {
1016
+ "epoch": 23.14,
1017
+ "learning_rate": 2.342001393081031e-05,
1018
+ "loss": 0.1014,
1019
+ "step": 34000
1020
+ },
1021
+ {
1022
+ "epoch": 23.14,
1023
+ "eval_loss": 0.4997089207172394,
1024
+ "eval_runtime": 111.4563,
1025
+ "eval_samples_per_second": 17.801,
1026
+ "eval_steps_per_second": 2.225,
1027
+ "eval_wer": 0.292376226457652,
1028
+ "step": 34000
1029
+ },
1030
+ {
1031
+ "epoch": 23.49,
1032
+ "learning_rate": 2.2261434873461805e-05,
1033
+ "loss": 0.0959,
1034
+ "step": 34500
1035
+ },
1036
+ {
1037
+ "epoch": 23.49,
1038
+ "eval_loss": 0.48040756583213806,
1039
+ "eval_runtime": 129.4236,
1040
+ "eval_samples_per_second": 15.33,
1041
+ "eval_steps_per_second": 1.916,
1042
+ "eval_wer": 0.28978233900981165,
1043
+ "step": 34500
1044
+ },
1045
+ {
1046
+ "epoch": 23.83,
1047
+ "learning_rate": 2.110053401439517e-05,
1048
+ "loss": 0.098,
1049
+ "step": 35000
1050
+ },
1051
+ {
1052
+ "epoch": 23.83,
1053
+ "eval_loss": 0.501395046710968,
1054
+ "eval_runtime": 129.507,
1055
+ "eval_samples_per_second": 15.32,
1056
+ "eval_steps_per_second": 1.915,
1057
+ "eval_wer": 0.291699560166911,
1058
+ "step": 35000
1059
+ },
1060
+ {
1061
+ "epoch": 24.17,
1062
+ "learning_rate": 1.9939633155328534e-05,
1063
+ "loss": 0.0973,
1064
+ "step": 35500
1065
+ },
1066
+ {
1067
+ "epoch": 24.17,
1068
+ "eval_loss": 0.5193932056427002,
1069
+ "eval_runtime": 129.3856,
1070
+ "eval_samples_per_second": 15.334,
1071
+ "eval_steps_per_second": 1.917,
1072
+ "eval_wer": 0.28955678357956466,
1073
+ "step": 35500
1074
+ },
1075
+ {
1076
+ "epoch": 24.51,
1077
+ "learning_rate": 1.87787322962619e-05,
1078
+ "loss": 0.0876,
1079
+ "step": 36000
1080
+ },
1081
+ {
1082
+ "epoch": 24.51,
1083
+ "eval_loss": 0.5203258395195007,
1084
+ "eval_runtime": 125.0985,
1085
+ "eval_samples_per_second": 15.859,
1086
+ "eval_steps_per_second": 1.982,
1087
+ "eval_wer": 0.28718845156197137,
1088
+ "step": 36000
1089
+ },
1090
+ {
1091
+ "epoch": 24.85,
1092
+ "learning_rate": 1.7617831437195264e-05,
1093
+ "loss": 0.0902,
1094
+ "step": 36500
1095
+ },
1096
+ {
1097
+ "epoch": 24.85,
1098
+ "eval_loss": 0.536376953125,
1099
+ "eval_runtime": 113.525,
1100
+ "eval_samples_per_second": 17.476,
1101
+ "eval_steps_per_second": 2.185,
1102
+ "eval_wer": 0.29136122702154055,
1103
+ "step": 36500
1104
+ },
1105
+ {
1106
+ "epoch": 25.19,
1107
+ "learning_rate": 1.645693057812863e-05,
1108
+ "loss": 0.0889,
1109
+ "step": 37000
1110
+ },
1111
+ {
1112
+ "epoch": 25.19,
1113
+ "eval_loss": 0.531440019607544,
1114
+ "eval_runtime": 123.4292,
1115
+ "eval_samples_per_second": 16.074,
1116
+ "eval_steps_per_second": 2.009,
1117
+ "eval_wer": 0.2880906732829593,
1118
+ "step": 37000
1119
+ },
1120
+ {
1121
+ "epoch": 25.53,
1122
+ "learning_rate": 1.5296029719061993e-05,
1123
+ "loss": 0.0865,
1124
+ "step": 37500
1125
+ },
1126
+ {
1127
+ "epoch": 25.53,
1128
+ "eval_loss": 0.5107194185256958,
1129
+ "eval_runtime": 129.5881,
1130
+ "eval_samples_per_second": 15.31,
1131
+ "eval_steps_per_second": 1.914,
1132
+ "eval_wer": 0.285271230404872,
1133
+ "step": 37500
1134
+ },
1135
+ {
1136
+ "epoch": 25.87,
1137
+ "learning_rate": 1.4135128859995356e-05,
1138
+ "loss": 0.0859,
1139
+ "step": 38000
1140
+ },
1141
+ {
1142
+ "epoch": 25.87,
1143
+ "eval_loss": 0.5254319310188293,
1144
+ "eval_runtime": 128.9308,
1145
+ "eval_samples_per_second": 15.388,
1146
+ "eval_steps_per_second": 1.924,
1147
+ "eval_wer": 0.28780872899515053,
1148
+ "step": 38000
1149
+ },
1150
+ {
1151
+ "epoch": 26.21,
1152
+ "learning_rate": 1.2974228000928721e-05,
1153
+ "loss": 0.0813,
1154
+ "step": 38500
1155
+ },
1156
+ {
1157
+ "epoch": 26.21,
1158
+ "eval_loss": 0.5275471806526184,
1159
+ "eval_runtime": 110.8092,
1160
+ "eval_samples_per_second": 17.905,
1161
+ "eval_steps_per_second": 2.238,
1162
+ "eval_wer": 0.2846509529716928,
1163
+ "step": 38500
1164
+ },
1165
+ {
1166
+ "epoch": 26.55,
1167
+ "learning_rate": 1.1813327141862086e-05,
1168
+ "loss": 0.0881,
1169
+ "step": 39000
1170
+ },
1171
+ {
1172
+ "epoch": 26.55,
1173
+ "eval_loss": 0.5125020742416382,
1174
+ "eval_runtime": 113.2218,
1175
+ "eval_samples_per_second": 17.523,
1176
+ "eval_steps_per_second": 2.19,
1177
+ "eval_wer": 0.28324123153264913,
1178
+ "step": 39000
1179
+ },
1180
+ {
1181
+ "epoch": 26.89,
1182
+ "learning_rate": 1.0654748084513583e-05,
1183
+ "loss": 0.0822,
1184
+ "step": 39500
1185
+ },
1186
+ {
1187
+ "epoch": 26.89,
1188
+ "eval_loss": 0.5309813618659973,
1189
+ "eval_runtime": 129.6599,
1190
+ "eval_samples_per_second": 15.302,
1191
+ "eval_steps_per_second": 1.913,
1192
+ "eval_wer": 0.28634261869854516,
1193
+ "step": 39500
1194
+ },
1195
+ {
1196
+ "epoch": 27.23,
1197
+ "learning_rate": 9.493847225446947e-06,
1198
+ "loss": 0.0837,
1199
+ "step": 40000
1200
+ },
1201
+ {
1202
+ "epoch": 27.23,
1203
+ "eval_loss": 0.5188203454017639,
1204
+ "eval_runtime": 122.3408,
1205
+ "eval_samples_per_second": 16.217,
1206
+ "eval_steps_per_second": 2.027,
1207
+ "eval_wer": 0.28312845381752566,
1208
+ "step": 40000
1209
+ },
1210
+ {
1211
+ "epoch": 27.57,
1212
+ "learning_rate": 8.332946366380312e-06,
1213
+ "loss": 0.0823,
1214
+ "step": 40500
1215
+ },
1216
+ {
1217
+ "epoch": 27.57,
1218
+ "eval_loss": 0.5201263427734375,
1219
+ "eval_runtime": 118.4812,
1220
+ "eval_samples_per_second": 16.745,
1221
+ "eval_steps_per_second": 2.093,
1222
+ "eval_wer": 0.28132401037554977,
1223
+ "step": 40500
1224
+ },
1225
+ {
1226
+ "epoch": 27.91,
1227
+ "learning_rate": 7.172045507313675e-06,
1228
+ "loss": 0.0768,
1229
+ "step": 41000
1230
+ },
1231
+ {
1232
+ "epoch": 27.91,
1233
+ "eval_loss": 0.5197951793670654,
1234
+ "eval_runtime": 112.4253,
1235
+ "eval_samples_per_second": 17.647,
1236
+ "eval_steps_per_second": 2.206,
1237
+ "eval_wer": 0.282169843238976,
1238
+ "step": 41000
1239
+ },
1240
+ {
1241
+ "epoch": 28.25,
1242
+ "learning_rate": 6.013466449965173e-06,
1243
+ "loss": 0.0858,
1244
+ "step": 41500
1245
+ },
1246
+ {
1247
+ "epoch": 28.25,
1248
+ "eval_loss": 0.5245384573936462,
1249
+ "eval_runtime": 128.7266,
1250
+ "eval_samples_per_second": 15.413,
1251
+ "eval_steps_per_second": 1.927,
1252
+ "eval_wer": 0.2786737340701477,
1253
+ "step": 41500
1254
+ },
1255
+ {
1256
+ "epoch": 28.59,
1257
+ "learning_rate": 4.8525655908985375e-06,
1258
+ "loss": 0.0757,
1259
+ "step": 42000
1260
+ },
1261
+ {
1262
+ "epoch": 28.59,
1263
+ "eval_loss": 0.5289037227630615,
1264
+ "eval_runtime": 112.5429,
1265
+ "eval_samples_per_second": 17.629,
1266
+ "eval_steps_per_second": 2.204,
1267
+ "eval_wer": 0.2766437351979249,
1268
+ "step": 42000
1269
+ },
1270
+ {
1271
+ "epoch": 28.93,
1272
+ "learning_rate": 3.6916647318319014e-06,
1273
+ "loss": 0.0728,
1274
+ "step": 42500
1275
+ },
1276
+ {
1277
+ "epoch": 28.93,
1278
+ "eval_loss": 0.5263972282409668,
1279
+ "eval_runtime": 128.0915,
1280
+ "eval_samples_per_second": 15.489,
1281
+ "eval_steps_per_second": 1.936,
1282
+ "eval_wer": 0.2779970677794068,
1283
+ "step": 42500
1284
+ },
1285
+ {
1286
+ "epoch": 29.27,
1287
+ "learning_rate": 2.5307638727652657e-06,
1288
+ "loss": 0.0761,
1289
+ "step": 43000
1290
+ },
1291
+ {
1292
+ "epoch": 29.27,
1293
+ "eval_loss": 0.5295293927192688,
1294
+ "eval_runtime": 128.9693,
1295
+ "eval_samples_per_second": 15.384,
1296
+ "eval_steps_per_second": 1.923,
1297
+ "eval_wer": 0.2770384572008571,
1298
+ "step": 43000
1299
+ },
1300
+ {
1301
+ "epoch": 29.61,
1302
+ "learning_rate": 1.3698630136986302e-06,
1303
+ "loss": 0.0739,
1304
+ "step": 43500
1305
+ },
1306
+ {
1307
+ "epoch": 29.61,
1308
+ "eval_loss": 0.5267478823661804,
1309
+ "eval_runtime": 130.4434,
1310
+ "eval_samples_per_second": 15.21,
1311
+ "eval_steps_per_second": 1.901,
1312
+ "eval_wer": 0.2776023457764746,
1313
+ "step": 43500
1314
+ }
1315
+ ],
1316
+ "max_steps": 44070,
1317
+ "num_train_epochs": 30,
1318
+ "total_flos": 1.385946883261073e+19,
1319
+ "trial_name": null,
1320
+ "trial_params": null
1321
+ }
wav2vec2_esp_30h/checkpoint-43500/training_args.bin ADDED
Binary file (3 kB). View file
 
wav2vec2_esp_30h/checkpoint-44000/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
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": 0.1,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
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": 0.1,
52
+ "feat_quantizer_dropout": 0.0,
53
+ "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
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": 37,
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": "4.17.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 38,
116
+ "xvector_output_dim": 512
117
+ }
wav2vec2_esp_30h/checkpoint-44000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7110237357fe715569e2188e2f23724c51489b913e416952292d171c99b2fa02
3
+ size 721723525
wav2vec2_esp_30h/checkpoint-44000/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
+ }
wav2vec2_esp_30h/checkpoint-44000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ecc8f115c830b9237ad6fa177601ec8905b1b79caf32dde93e014e800a55d7f7
3
+ size 377691873
wav2vec2_esp_30h/checkpoint-44000/rng_state.pth ADDED
Binary file (14.6 kB). View file
 
wav2vec2_esp_30h/checkpoint-44000/scaler.pt ADDED
Binary file (557 Bytes). View file
 
wav2vec2_esp_30h/checkpoint-44000/scheduler.pt ADDED
Binary file (627 Bytes). View file
 
wav2vec2_esp_30h/checkpoint-44000/trainer_state.json ADDED
@@ -0,0 +1,1336 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.952348536419333,
5
+ "global_step": 44000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.34,
12
+ "learning_rate": 4.9500000000000004e-05,
13
+ "loss": 6.2203,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.34,
18
+ "eval_loss": 3.002925157546997,
19
+ "eval_runtime": 132.7789,
20
+ "eval_samples_per_second": 14.942,
21
+ "eval_steps_per_second": 1.868,
22
+ "eval_wer": 1.0,
23
+ "step": 500
24
+ },
25
+ {
26
+ "epoch": 0.68,
27
+ "learning_rate": 9.95e-05,
28
+ "loss": 2.2102,
29
+ "step": 1000
30
+ },
31
+ {
32
+ "epoch": 0.68,
33
+ "eval_loss": 1.1446926593780518,
34
+ "eval_runtime": 132.5997,
35
+ "eval_samples_per_second": 14.962,
36
+ "eval_steps_per_second": 1.87,
37
+ "eval_wer": 0.8542348032028871,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 1.02,
42
+ "learning_rate": 9.885070814952404e-05,
43
+ "loss": 0.9714,
44
+ "step": 1500
45
+ },
46
+ {
47
+ "epoch": 1.02,
48
+ "eval_loss": 0.7160163521766663,
49
+ "eval_runtime": 132.7305,
50
+ "eval_samples_per_second": 14.948,
51
+ "eval_steps_per_second": 1.868,
52
+ "eval_wer": 0.6600879666177963,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 1.36,
57
+ "learning_rate": 9.76898072904574e-05,
58
+ "loss": 0.7541,
59
+ "step": 2000
60
+ },
61
+ {
62
+ "epoch": 1.36,
63
+ "eval_loss": 0.6261711120605469,
64
+ "eval_runtime": 132.7267,
65
+ "eval_samples_per_second": 14.948,
66
+ "eval_steps_per_second": 1.869,
67
+ "eval_wer": 0.594733280703733,
68
+ "step": 2000
69
+ },
70
+ {
71
+ "epoch": 1.7,
72
+ "learning_rate": 9.652890643139077e-05,
73
+ "loss": 0.6838,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 1.7,
78
+ "eval_loss": 0.5789812207221985,
79
+ "eval_runtime": 131.2782,
80
+ "eval_samples_per_second": 15.113,
81
+ "eval_steps_per_second": 1.889,
82
+ "eval_wer": 0.5645088530506372,
83
+ "step": 2500
84
+ },
85
+ {
86
+ "epoch": 2.04,
87
+ "learning_rate": 9.537032737404225e-05,
88
+ "loss": 0.6287,
89
+ "step": 3000
90
+ },
91
+ {
92
+ "epoch": 2.04,
93
+ "eval_loss": 0.6015335917472839,
94
+ "eval_runtime": 133.004,
95
+ "eval_samples_per_second": 14.917,
96
+ "eval_steps_per_second": 1.865,
97
+ "eval_wer": 0.5387955340024811,
98
+ "step": 3000
99
+ },
100
+ {
101
+ "epoch": 2.38,
102
+ "learning_rate": 9.420942651497563e-05,
103
+ "loss": 0.5439,
104
+ "step": 3500
105
+ },
106
+ {
107
+ "epoch": 2.38,
108
+ "eval_loss": 0.5301734209060669,
109
+ "eval_runtime": 133.1151,
110
+ "eval_samples_per_second": 14.904,
111
+ "eval_steps_per_second": 1.863,
112
+ "eval_wer": 0.5076124957708357,
113
+ "step": 3500
114
+ },
115
+ {
116
+ "epoch": 2.72,
117
+ "learning_rate": 9.304852565590899e-05,
118
+ "loss": 0.5279,
119
+ "step": 4000
120
+ },
121
+ {
122
+ "epoch": 2.72,
123
+ "eval_loss": 0.5215476155281067,
124
+ "eval_runtime": 131.1447,
125
+ "eval_samples_per_second": 15.128,
126
+ "eval_steps_per_second": 1.891,
127
+ "eval_wer": 0.48962445020863876,
128
+ "step": 4000
129
+ },
130
+ {
131
+ "epoch": 3.06,
132
+ "learning_rate": 9.188762479684235e-05,
133
+ "loss": 0.5006,
134
+ "step": 4500
135
+ },
136
+ {
137
+ "epoch": 3.06,
138
+ "eval_loss": 0.4860161542892456,
139
+ "eval_runtime": 133.2608,
140
+ "eval_samples_per_second": 14.888,
141
+ "eval_steps_per_second": 1.861,
142
+ "eval_wer": 0.4685350174805458,
143
+ "step": 4500
144
+ },
145
+ {
146
+ "epoch": 3.4,
147
+ "learning_rate": 9.072672393777572e-05,
148
+ "loss": 0.4432,
149
+ "step": 5000
150
+ },
151
+ {
152
+ "epoch": 3.4,
153
+ "eval_loss": 0.484553724527359,
154
+ "eval_runtime": 133.4845,
155
+ "eval_samples_per_second": 14.863,
156
+ "eval_steps_per_second": 1.858,
157
+ "eval_wer": 0.46864779519566935,
158
+ "step": 5000
159
+ },
160
+ {
161
+ "epoch": 3.74,
162
+ "learning_rate": 8.956582307870908e-05,
163
+ "loss": 0.4334,
164
+ "step": 5500
165
+ },
166
+ {
167
+ "epoch": 3.74,
168
+ "eval_loss": 0.477468878030777,
169
+ "eval_runtime": 132.8856,
170
+ "eval_samples_per_second": 14.93,
171
+ "eval_steps_per_second": 1.866,
172
+ "eval_wer": 0.45415585880230064,
173
+ "step": 5500
174
+ },
175
+ {
176
+ "epoch": 4.08,
177
+ "learning_rate": 8.840492221964245e-05,
178
+ "loss": 0.4292,
179
+ "step": 6000
180
+ },
181
+ {
182
+ "epoch": 4.08,
183
+ "eval_loss": 0.451526939868927,
184
+ "eval_runtime": 132.3623,
185
+ "eval_samples_per_second": 14.989,
186
+ "eval_steps_per_second": 1.874,
187
+ "eval_wer": 0.4291192060448855,
188
+ "step": 6000
189
+ },
190
+ {
191
+ "epoch": 4.42,
192
+ "learning_rate": 8.724402136057582e-05,
193
+ "loss": 0.3779,
194
+ "step": 6500
195
+ },
196
+ {
197
+ "epoch": 4.42,
198
+ "eval_loss": 0.4495590329170227,
199
+ "eval_runtime": 132.6275,
200
+ "eval_samples_per_second": 14.959,
201
+ "eval_steps_per_second": 1.87,
202
+ "eval_wer": 0.42054809969550017,
203
+ "step": 6500
204
+ },
205
+ {
206
+ "epoch": 4.77,
207
+ "learning_rate": 8.608312050150917e-05,
208
+ "loss": 0.3783,
209
+ "step": 7000
210
+ },
211
+ {
212
+ "epoch": 4.77,
213
+ "eval_loss": 0.45758938789367676,
214
+ "eval_runtime": 132.4666,
215
+ "eval_samples_per_second": 14.977,
216
+ "eval_steps_per_second": 1.872,
217
+ "eval_wer": 0.4184053231081538,
218
+ "step": 7000
219
+ },
220
+ {
221
+ "epoch": 5.11,
222
+ "learning_rate": 8.492221964244254e-05,
223
+ "loss": 0.3622,
224
+ "step": 7500
225
+ },
226
+ {
227
+ "epoch": 5.11,
228
+ "eval_loss": 0.4782721698284149,
229
+ "eval_runtime": 133.2094,
230
+ "eval_samples_per_second": 14.894,
231
+ "eval_steps_per_second": 1.862,
232
+ "eval_wer": 0.4070147738806812,
233
+ "step": 7500
234
+ },
235
+ {
236
+ "epoch": 5.45,
237
+ "learning_rate": 8.37613187833759e-05,
238
+ "loss": 0.3278,
239
+ "step": 8000
240
+ },
241
+ {
242
+ "epoch": 5.45,
243
+ "eval_loss": 0.44267573952674866,
244
+ "eval_runtime": 133.2221,
245
+ "eval_samples_per_second": 14.892,
246
+ "eval_steps_per_second": 1.862,
247
+ "eval_wer": 0.40278560956355025,
248
+ "step": 8000
249
+ },
250
+ {
251
+ "epoch": 5.79,
252
+ "learning_rate": 8.260041792430927e-05,
253
+ "loss": 0.3304,
254
+ "step": 8500
255
+ },
256
+ {
257
+ "epoch": 5.79,
258
+ "eval_loss": 0.4482724368572235,
259
+ "eval_runtime": 132.3395,
260
+ "eval_samples_per_second": 14.992,
261
+ "eval_steps_per_second": 1.874,
262
+ "eval_wer": 0.4056050524416375,
263
+ "step": 8500
264
+ },
265
+ {
266
+ "epoch": 6.13,
267
+ "learning_rate": 8.143951706524264e-05,
268
+ "loss": 0.312,
269
+ "step": 9000
270
+ },
271
+ {
272
+ "epoch": 6.13,
273
+ "eval_loss": 0.4750816524028778,
274
+ "eval_runtime": 132.887,
275
+ "eval_samples_per_second": 14.93,
276
+ "eval_steps_per_second": 1.866,
277
+ "eval_wer": 0.3882936731701816,
278
+ "step": 9000
279
+ },
280
+ {
281
+ "epoch": 6.47,
282
+ "learning_rate": 8.027861620617599e-05,
283
+ "loss": 0.29,
284
+ "step": 9500
285
+ },
286
+ {
287
+ "epoch": 6.47,
288
+ "eval_loss": 0.4528858959674835,
289
+ "eval_runtime": 133.211,
290
+ "eval_samples_per_second": 14.894,
291
+ "eval_steps_per_second": 1.862,
292
+ "eval_wer": 0.3780872899515056,
293
+ "step": 9500
294
+ },
295
+ {
296
+ "epoch": 6.81,
297
+ "learning_rate": 7.912235895054563e-05,
298
+ "loss": 0.3057,
299
+ "step": 10000
300
+ },
301
+ {
302
+ "epoch": 6.81,
303
+ "eval_loss": 0.5861864686012268,
304
+ "eval_runtime": 133.2534,
305
+ "eval_samples_per_second": 14.889,
306
+ "eval_steps_per_second": 1.861,
307
+ "eval_wer": 0.3783128453817526,
308
+ "step": 10000
309
+ },
310
+ {
311
+ "epoch": 7.15,
312
+ "learning_rate": 7.796145809147898e-05,
313
+ "loss": 0.2971,
314
+ "step": 10500
315
+ },
316
+ {
317
+ "epoch": 7.15,
318
+ "eval_loss": 0.43457281589508057,
319
+ "eval_runtime": 132.5985,
320
+ "eval_samples_per_second": 14.962,
321
+ "eval_steps_per_second": 1.87,
322
+ "eval_wer": 0.3765084019397767,
323
+ "step": 10500
324
+ },
325
+ {
326
+ "epoch": 7.49,
327
+ "learning_rate": 7.680055723241235e-05,
328
+ "loss": 0.2684,
329
+ "step": 11000
330
+ },
331
+ {
332
+ "epoch": 7.49,
333
+ "eval_loss": 0.4560734033584595,
334
+ "eval_runtime": 132.2037,
335
+ "eval_samples_per_second": 15.007,
336
+ "eval_steps_per_second": 1.876,
337
+ "eval_wer": 0.3732942370587572,
338
+ "step": 11000
339
+ },
340
+ {
341
+ "epoch": 7.83,
342
+ "learning_rate": 7.563965637334573e-05,
343
+ "loss": 0.2622,
344
+ "step": 11500
345
+ },
346
+ {
347
+ "epoch": 7.83,
348
+ "eval_loss": 0.43240657448768616,
349
+ "eval_runtime": 133.3517,
350
+ "eval_samples_per_second": 14.878,
351
+ "eval_steps_per_second": 1.86,
352
+ "eval_wer": 0.37408368106462164,
353
+ "step": 11500
354
+ },
355
+ {
356
+ "epoch": 8.17,
357
+ "learning_rate": 7.447875551427908e-05,
358
+ "loss": 0.2635,
359
+ "step": 12000
360
+ },
361
+ {
362
+ "epoch": 8.17,
363
+ "eval_loss": 0.4556463658809662,
364
+ "eval_runtime": 133.4279,
365
+ "eval_samples_per_second": 14.869,
366
+ "eval_steps_per_second": 1.859,
367
+ "eval_wer": 0.3789895116724935,
368
+ "step": 12000
369
+ },
370
+ {
371
+ "epoch": 8.51,
372
+ "learning_rate": 7.331785465521245e-05,
373
+ "loss": 0.2363,
374
+ "step": 12500
375
+ },
376
+ {
377
+ "epoch": 8.51,
378
+ "eval_loss": 0.47497859597206116,
379
+ "eval_runtime": 134.0404,
380
+ "eval_samples_per_second": 14.802,
381
+ "eval_steps_per_second": 1.85,
382
+ "eval_wer": 0.3709822938987256,
383
+ "step": 12500
384
+ },
385
+ {
386
+ "epoch": 8.85,
387
+ "learning_rate": 7.215695379614582e-05,
388
+ "loss": 0.2516,
389
+ "step": 13000
390
+ },
391
+ {
392
+ "epoch": 8.85,
393
+ "eval_loss": 0.4297301471233368,
394
+ "eval_runtime": 133.4288,
395
+ "eval_samples_per_second": 14.869,
396
+ "eval_steps_per_second": 1.859,
397
+ "eval_wer": 0.3635389647005752,
398
+ "step": 13000
399
+ },
400
+ {
401
+ "epoch": 9.19,
402
+ "learning_rate": 7.099605293707917e-05,
403
+ "loss": 0.2291,
404
+ "step": 13500
405
+ },
406
+ {
407
+ "epoch": 9.19,
408
+ "eval_loss": 0.4862041175365448,
409
+ "eval_runtime": 133.7259,
410
+ "eval_samples_per_second": 14.836,
411
+ "eval_steps_per_second": 1.855,
412
+ "eval_wer": 0.36021202210443215,
413
+ "step": 13500
414
+ },
415
+ {
416
+ "epoch": 9.53,
417
+ "learning_rate": 6.983515207801255e-05,
418
+ "loss": 0.2266,
419
+ "step": 14000
420
+ },
421
+ {
422
+ "epoch": 9.53,
423
+ "eval_loss": 0.4703396260738373,
424
+ "eval_runtime": 133.0987,
425
+ "eval_samples_per_second": 14.906,
426
+ "eval_steps_per_second": 1.863,
427
+ "eval_wer": 0.3546295252058193,
428
+ "step": 14000
429
+ },
430
+ {
431
+ "epoch": 9.87,
432
+ "learning_rate": 6.867657302066403e-05,
433
+ "loss": 0.2281,
434
+ "step": 14500
435
+ },
436
+ {
437
+ "epoch": 9.87,
438
+ "eval_loss": 0.42882266640663147,
439
+ "eval_runtime": 133.4534,
440
+ "eval_samples_per_second": 14.867,
441
+ "eval_steps_per_second": 1.858,
442
+ "eval_wer": 0.35812563437464756,
443
+ "step": 14500
444
+ },
445
+ {
446
+ "epoch": 10.21,
447
+ "learning_rate": 6.75156721615974e-05,
448
+ "loss": 0.2126,
449
+ "step": 15000
450
+ },
451
+ {
452
+ "epoch": 10.21,
453
+ "eval_loss": 0.4615735113620758,
454
+ "eval_runtime": 132.7985,
455
+ "eval_samples_per_second": 14.94,
456
+ "eval_steps_per_second": 1.867,
457
+ "eval_wer": 0.35626480207510997,
458
+ "step": 15000
459
+ },
460
+ {
461
+ "epoch": 10.55,
462
+ "learning_rate": 6.635477130253078e-05,
463
+ "loss": 0.217,
464
+ "step": 15500
465
+ },
466
+ {
467
+ "epoch": 10.55,
468
+ "eval_loss": 0.49455586075782776,
469
+ "eval_runtime": 133.6528,
470
+ "eval_samples_per_second": 14.844,
471
+ "eval_steps_per_second": 1.856,
472
+ "eval_wer": 0.34560730799594,
473
+ "step": 15500
474
+ },
475
+ {
476
+ "epoch": 10.89,
477
+ "learning_rate": 6.519387044346413e-05,
478
+ "loss": 0.212,
479
+ "step": 16000
480
+ },
481
+ {
482
+ "epoch": 10.89,
483
+ "eval_loss": 0.43323996663093567,
484
+ "eval_runtime": 133.4419,
485
+ "eval_samples_per_second": 14.868,
486
+ "eval_steps_per_second": 1.858,
487
+ "eval_wer": 0.3452689748505695,
488
+ "step": 16000
489
+ },
490
+ {
491
+ "epoch": 11.23,
492
+ "learning_rate": 6.40329695843975e-05,
493
+ "loss": 0.1986,
494
+ "step": 16500
495
+ },
496
+ {
497
+ "epoch": 11.23,
498
+ "eval_loss": 0.4868086576461792,
499
+ "eval_runtime": 131.4603,
500
+ "eval_samples_per_second": 15.092,
501
+ "eval_steps_per_second": 1.887,
502
+ "eval_wer": 0.3399684222397654,
503
+ "step": 16500
504
+ },
505
+ {
506
+ "epoch": 11.57,
507
+ "learning_rate": 6.2874390527049e-05,
508
+ "loss": 0.2012,
509
+ "step": 17000
510
+ },
511
+ {
512
+ "epoch": 11.57,
513
+ "eval_loss": 0.4474054276943207,
514
+ "eval_runtime": 133.9018,
515
+ "eval_samples_per_second": 14.817,
516
+ "eval_steps_per_second": 1.852,
517
+ "eval_wer": 0.3460020299988722,
518
+ "step": 17000
519
+ },
520
+ {
521
+ "epoch": 11.91,
522
+ "learning_rate": 6.171348966798236e-05,
523
+ "loss": 0.1998,
524
+ "step": 17500
525
+ },
526
+ {
527
+ "epoch": 11.91,
528
+ "eval_loss": 0.45011407136917114,
529
+ "eval_runtime": 132.7933,
530
+ "eval_samples_per_second": 14.941,
531
+ "eval_steps_per_second": 1.868,
532
+ "eval_wer": 0.3362467576406902,
533
+ "step": 17500
534
+ },
535
+ {
536
+ "epoch": 12.25,
537
+ "learning_rate": 6.055258880891572e-05,
538
+ "loss": 0.1746,
539
+ "step": 18000
540
+ },
541
+ {
542
+ "epoch": 12.25,
543
+ "eval_loss": 0.4731091558933258,
544
+ "eval_runtime": 133.9219,
545
+ "eval_samples_per_second": 14.815,
546
+ "eval_steps_per_second": 1.852,
547
+ "eval_wer": 0.33630314649825194,
548
+ "step": 18000
549
+ },
550
+ {
551
+ "epoch": 12.59,
552
+ "learning_rate": 5.9391687949849086e-05,
553
+ "loss": 0.1805,
554
+ "step": 18500
555
+ },
556
+ {
557
+ "epoch": 12.59,
558
+ "eval_loss": 0.4601946771144867,
559
+ "eval_runtime": 132.0532,
560
+ "eval_samples_per_second": 15.024,
561
+ "eval_steps_per_second": 1.878,
562
+ "eval_wer": 0.33743092364948685,
563
+ "step": 18500
564
+ },
565
+ {
566
+ "epoch": 12.93,
567
+ "learning_rate": 5.823310889250059e-05,
568
+ "loss": 0.1826,
569
+ "step": 19000
570
+ },
571
+ {
572
+ "epoch": 12.93,
573
+ "eval_loss": 0.47728002071380615,
574
+ "eval_runtime": 133.9362,
575
+ "eval_samples_per_second": 14.813,
576
+ "eval_steps_per_second": 1.852,
577
+ "eval_wer": 0.3276192624337431,
578
+ "step": 19000
579
+ },
580
+ {
581
+ "epoch": 13.27,
582
+ "learning_rate": 5.707220803343395e-05,
583
+ "loss": 0.1651,
584
+ "step": 19500
585
+ },
586
+ {
587
+ "epoch": 13.27,
588
+ "eval_loss": 0.471328467130661,
589
+ "eval_runtime": 111.9462,
590
+ "eval_samples_per_second": 17.723,
591
+ "eval_steps_per_second": 2.215,
592
+ "eval_wer": 0.3304387053118304,
593
+ "step": 19500
594
+ },
595
+ {
596
+ "epoch": 13.61,
597
+ "learning_rate": 5.591130717436731e-05,
598
+ "loss": 0.1812,
599
+ "step": 20000
600
+ },
601
+ {
602
+ "epoch": 13.61,
603
+ "eval_loss": 0.4280690550804138,
604
+ "eval_runtime": 110.8492,
605
+ "eval_samples_per_second": 17.898,
606
+ "eval_steps_per_second": 2.237,
607
+ "eval_wer": 0.32276982068343296,
608
+ "step": 20000
609
+ },
610
+ {
611
+ "epoch": 13.96,
612
+ "learning_rate": 5.4750406315300676e-05,
613
+ "loss": 0.1666,
614
+ "step": 20500
615
+ },
616
+ {
617
+ "epoch": 13.96,
618
+ "eval_loss": 0.4796316623687744,
619
+ "eval_runtime": 110.5982,
620
+ "eval_samples_per_second": 17.939,
621
+ "eval_steps_per_second": 2.242,
622
+ "eval_wer": 0.32530731927371154,
623
+ "step": 20500
624
+ },
625
+ {
626
+ "epoch": 14.3,
627
+ "learning_rate": 5.359182725795218e-05,
628
+ "loss": 0.1553,
629
+ "step": 21000
630
+ },
631
+ {
632
+ "epoch": 14.3,
633
+ "eval_loss": 0.4721369445323944,
634
+ "eval_runtime": 128.2013,
635
+ "eval_samples_per_second": 15.476,
636
+ "eval_steps_per_second": 1.934,
637
+ "eval_wer": 0.3259275967068907,
638
+ "step": 21000
639
+ },
640
+ {
641
+ "epoch": 14.64,
642
+ "learning_rate": 5.243092639888554e-05,
643
+ "loss": 0.1545,
644
+ "step": 21500
645
+ },
646
+ {
647
+ "epoch": 14.64,
648
+ "eval_loss": 0.4527774453163147,
649
+ "eval_runtime": 132.8486,
650
+ "eval_samples_per_second": 14.934,
651
+ "eval_steps_per_second": 1.867,
652
+ "eval_wer": 0.32682981842787867,
653
+ "step": 21500
654
+ },
655
+ {
656
+ "epoch": 14.98,
657
+ "learning_rate": 5.12700255398189e-05,
658
+ "loss": 0.1576,
659
+ "step": 22000
660
+ },
661
+ {
662
+ "epoch": 14.98,
663
+ "eval_loss": 0.4553755819797516,
664
+ "eval_runtime": 131.6613,
665
+ "eval_samples_per_second": 15.069,
666
+ "eval_steps_per_second": 1.884,
667
+ "eval_wer": 0.3252509304161498,
668
+ "step": 22000
669
+ },
670
+ {
671
+ "epoch": 15.32,
672
+ "learning_rate": 5.0109124680752265e-05,
673
+ "loss": 0.1511,
674
+ "step": 22500
675
+ },
676
+ {
677
+ "epoch": 15.32,
678
+ "eval_loss": 0.4580444097518921,
679
+ "eval_runtime": 134.2974,
680
+ "eval_samples_per_second": 14.773,
681
+ "eval_steps_per_second": 1.847,
682
+ "eval_wer": 0.3179203789331228,
683
+ "step": 22500
684
+ },
685
+ {
686
+ "epoch": 15.66,
687
+ "learning_rate": 4.895054562340376e-05,
688
+ "loss": 0.1444,
689
+ "step": 23000
690
+ },
691
+ {
692
+ "epoch": 15.66,
693
+ "eval_loss": 0.4659278094768524,
694
+ "eval_runtime": 134.6637,
695
+ "eval_samples_per_second": 14.733,
696
+ "eval_steps_per_second": 1.842,
697
+ "eval_wer": 0.32119093267170407,
698
+ "step": 23000
699
+ },
700
+ {
701
+ "epoch": 16.0,
702
+ "learning_rate": 4.7791966566055266e-05,
703
+ "loss": 0.1496,
704
+ "step": 23500
705
+ },
706
+ {
707
+ "epoch": 16.0,
708
+ "eval_loss": 0.4660351872444153,
709
+ "eval_runtime": 133.4382,
710
+ "eval_samples_per_second": 14.868,
711
+ "eval_steps_per_second": 1.859,
712
+ "eval_wer": 0.32598398556445246,
713
+ "step": 23500
714
+ },
715
+ {
716
+ "epoch": 16.34,
717
+ "learning_rate": 4.6631065706988625e-05,
718
+ "loss": 0.1327,
719
+ "step": 24000
720
+ },
721
+ {
722
+ "epoch": 16.34,
723
+ "eval_loss": 0.4934828281402588,
724
+ "eval_runtime": 134.3743,
725
+ "eval_samples_per_second": 14.765,
726
+ "eval_steps_per_second": 1.846,
727
+ "eval_wer": 0.31825871207849327,
728
+ "step": 24000
729
+ },
730
+ {
731
+ "epoch": 16.68,
732
+ "learning_rate": 4.547016484792199e-05,
733
+ "loss": 0.1535,
734
+ "step": 24500
735
+ },
736
+ {
737
+ "epoch": 16.68,
738
+ "eval_loss": 0.46466243267059326,
739
+ "eval_runtime": 132.9271,
740
+ "eval_samples_per_second": 14.925,
741
+ "eval_steps_per_second": 1.866,
742
+ "eval_wer": 0.3113792714559603,
743
+ "step": 24500
744
+ },
745
+ {
746
+ "epoch": 17.02,
747
+ "learning_rate": 4.430926398885535e-05,
748
+ "loss": 0.1438,
749
+ "step": 25000
750
+ },
751
+ {
752
+ "epoch": 17.02,
753
+ "eval_loss": 0.5053276419639587,
754
+ "eval_runtime": 132.7443,
755
+ "eval_samples_per_second": 14.946,
756
+ "eval_steps_per_second": 1.868,
757
+ "eval_wer": 0.31487538062478854,
758
+ "step": 25000
759
+ },
760
+ {
761
+ "epoch": 17.36,
762
+ "learning_rate": 4.314836312978872e-05,
763
+ "loss": 0.1264,
764
+ "step": 25500
765
+ },
766
+ {
767
+ "epoch": 17.36,
768
+ "eval_loss": 0.48962870240211487,
769
+ "eval_runtime": 133.9594,
770
+ "eval_samples_per_second": 14.81,
771
+ "eval_steps_per_second": 1.851,
772
+ "eval_wer": 0.3131273260403744,
773
+ "step": 25500
774
+ },
775
+ {
776
+ "epoch": 17.7,
777
+ "learning_rate": 4.1987462270722084e-05,
778
+ "loss": 0.1269,
779
+ "step": 26000
780
+ },
781
+ {
782
+ "epoch": 17.7,
783
+ "eval_loss": 0.46815410256385803,
784
+ "eval_runtime": 132.7946,
785
+ "eval_samples_per_second": 14.94,
786
+ "eval_steps_per_second": 1.868,
787
+ "eval_wer": 0.30748844028419986,
788
+ "step": 26000
789
+ },
790
+ {
791
+ "epoch": 18.04,
792
+ "learning_rate": 4.0826561411655444e-05,
793
+ "loss": 0.1272,
794
+ "step": 26500
795
+ },
796
+ {
797
+ "epoch": 18.04,
798
+ "eval_loss": 0.4759911596775055,
799
+ "eval_runtime": 132.8376,
800
+ "eval_samples_per_second": 14.936,
801
+ "eval_steps_per_second": 1.867,
802
+ "eval_wer": 0.3113792714559603,
803
+ "step": 26500
804
+ },
805
+ {
806
+ "epoch": 18.38,
807
+ "learning_rate": 3.966566055258881e-05,
808
+ "loss": 0.1219,
809
+ "step": 27000
810
+ },
811
+ {
812
+ "epoch": 18.38,
813
+ "eval_loss": 0.49612942337989807,
814
+ "eval_runtime": 133.1179,
815
+ "eval_samples_per_second": 14.904,
816
+ "eval_steps_per_second": 1.863,
817
+ "eval_wer": 0.31487538062478854,
818
+ "step": 27000
819
+ },
820
+ {
821
+ "epoch": 18.72,
822
+ "learning_rate": 3.850475969352218e-05,
823
+ "loss": 0.123,
824
+ "step": 27500
825
+ },
826
+ {
827
+ "epoch": 18.72,
828
+ "eval_loss": 0.46862220764160156,
829
+ "eval_runtime": 110.6681,
830
+ "eval_samples_per_second": 17.927,
831
+ "eval_steps_per_second": 2.241,
832
+ "eval_wer": 0.3072064959963911,
833
+ "step": 27500
834
+ },
835
+ {
836
+ "epoch": 19.06,
837
+ "learning_rate": 3.734385883445554e-05,
838
+ "loss": 0.1262,
839
+ "step": 28000
840
+ },
841
+ {
842
+ "epoch": 19.06,
843
+ "eval_loss": 0.49365413188934326,
844
+ "eval_runtime": 109.7879,
845
+ "eval_samples_per_second": 18.071,
846
+ "eval_steps_per_second": 2.259,
847
+ "eval_wer": 0.30889816172324347,
848
+ "step": 28000
849
+ },
850
+ {
851
+ "epoch": 19.4,
852
+ "learning_rate": 3.61829579753889e-05,
853
+ "loss": 0.1165,
854
+ "step": 28500
855
+ },
856
+ {
857
+ "epoch": 19.4,
858
+ "eval_loss": 0.4988892376422882,
859
+ "eval_runtime": 110.8122,
860
+ "eval_samples_per_second": 17.904,
861
+ "eval_steps_per_second": 2.238,
862
+ "eval_wer": 0.30540205255441527,
863
+ "step": 28500
864
+ },
865
+ {
866
+ "epoch": 19.74,
867
+ "learning_rate": 3.50243789180404e-05,
868
+ "loss": 0.1213,
869
+ "step": 29000
870
+ },
871
+ {
872
+ "epoch": 19.74,
873
+ "eval_loss": 0.4937494993209839,
874
+ "eval_runtime": 109.7673,
875
+ "eval_samples_per_second": 18.075,
876
+ "eval_steps_per_second": 2.259,
877
+ "eval_wer": 0.30404871997293337,
878
+ "step": 29000
879
+ },
880
+ {
881
+ "epoch": 20.08,
882
+ "learning_rate": 3.3863478058973766e-05,
883
+ "loss": 0.1113,
884
+ "step": 29500
885
+ },
886
+ {
887
+ "epoch": 20.08,
888
+ "eval_loss": 0.5031572580337524,
889
+ "eval_runtime": 111.1946,
890
+ "eval_samples_per_second": 17.843,
891
+ "eval_steps_per_second": 2.23,
892
+ "eval_wer": 0.3037667756851246,
893
+ "step": 29500
894
+ },
895
+ {
896
+ "epoch": 20.42,
897
+ "learning_rate": 3.270257719990713e-05,
898
+ "loss": 0.1087,
899
+ "step": 30000
900
+ },
901
+ {
902
+ "epoch": 20.42,
903
+ "eval_loss": 0.49494612216949463,
904
+ "eval_runtime": 126.8851,
905
+ "eval_samples_per_second": 15.636,
906
+ "eval_steps_per_second": 1.955,
907
+ "eval_wer": 0.299030111649938,
908
+ "step": 30000
909
+ },
910
+ {
911
+ "epoch": 20.76,
912
+ "learning_rate": 3.154167634084049e-05,
913
+ "loss": 0.1128,
914
+ "step": 30500
915
+ },
916
+ {
917
+ "epoch": 20.76,
918
+ "eval_loss": 0.49934807419776917,
919
+ "eval_runtime": 128.8495,
920
+ "eval_samples_per_second": 15.398,
921
+ "eval_steps_per_second": 1.925,
922
+ "eval_wer": 0.3017367768129018,
923
+ "step": 30500
924
+ },
925
+ {
926
+ "epoch": 21.1,
927
+ "learning_rate": 3.038077548177386e-05,
928
+ "loss": 0.1151,
929
+ "step": 31000
930
+ },
931
+ {
932
+ "epoch": 21.1,
933
+ "eval_loss": 0.5088058114051819,
934
+ "eval_runtime": 128.5026,
935
+ "eval_samples_per_second": 15.439,
936
+ "eval_steps_per_second": 1.93,
937
+ "eval_wer": 0.30184955452802525,
938
+ "step": 31000
939
+ },
940
+ {
941
+ "epoch": 21.44,
942
+ "learning_rate": 2.922219642442536e-05,
943
+ "loss": 0.1025,
944
+ "step": 31500
945
+ },
946
+ {
947
+ "epoch": 21.44,
948
+ "eval_loss": 0.47611942887306213,
949
+ "eval_runtime": 111.1648,
950
+ "eval_samples_per_second": 17.847,
951
+ "eval_steps_per_second": 2.231,
952
+ "eval_wer": 0.29891733393481446,
953
+ "step": 31500
954
+ },
955
+ {
956
+ "epoch": 21.78,
957
+ "learning_rate": 2.8063617367076856e-05,
958
+ "loss": 0.1016,
959
+ "step": 32000
960
+ },
961
+ {
962
+ "epoch": 21.78,
963
+ "eval_loss": 0.5126113891601562,
964
+ "eval_runtime": 121.5373,
965
+ "eval_samples_per_second": 16.324,
966
+ "eval_steps_per_second": 2.041,
967
+ "eval_wer": 0.30275177624901317,
968
+ "step": 32000
969
+ },
970
+ {
971
+ "epoch": 22.12,
972
+ "learning_rate": 2.690271650801022e-05,
973
+ "loss": 0.0991,
974
+ "step": 32500
975
+ },
976
+ {
977
+ "epoch": 22.12,
978
+ "eval_loss": 0.5214097499847412,
979
+ "eval_runtime": 129.4856,
980
+ "eval_samples_per_second": 15.322,
981
+ "eval_steps_per_second": 1.915,
982
+ "eval_wer": 0.2958159467689185,
983
+ "step": 32500
984
+ },
985
+ {
986
+ "epoch": 22.46,
987
+ "learning_rate": 2.574181564894358e-05,
988
+ "loss": 0.0951,
989
+ "step": 33000
990
+ },
991
+ {
992
+ "epoch": 22.46,
993
+ "eval_loss": 0.5233449339866638,
994
+ "eval_runtime": 130.9955,
995
+ "eval_samples_per_second": 15.146,
996
+ "eval_steps_per_second": 1.893,
997
+ "eval_wer": 0.2923198376000902,
998
+ "step": 33000
999
+ },
1000
+ {
1001
+ "epoch": 22.8,
1002
+ "learning_rate": 2.4580914789876948e-05,
1003
+ "loss": 0.1049,
1004
+ "step": 33500
1005
+ },
1006
+ {
1007
+ "epoch": 22.8,
1008
+ "eval_loss": 0.48444411158561707,
1009
+ "eval_runtime": 110.8951,
1010
+ "eval_samples_per_second": 17.891,
1011
+ "eval_steps_per_second": 2.236,
1012
+ "eval_wer": 0.29677455734746816,
1013
+ "step": 33500
1014
+ },
1015
+ {
1016
+ "epoch": 23.14,
1017
+ "learning_rate": 2.342001393081031e-05,
1018
+ "loss": 0.1014,
1019
+ "step": 34000
1020
+ },
1021
+ {
1022
+ "epoch": 23.14,
1023
+ "eval_loss": 0.4997089207172394,
1024
+ "eval_runtime": 111.4563,
1025
+ "eval_samples_per_second": 17.801,
1026
+ "eval_steps_per_second": 2.225,
1027
+ "eval_wer": 0.292376226457652,
1028
+ "step": 34000
1029
+ },
1030
+ {
1031
+ "epoch": 23.49,
1032
+ "learning_rate": 2.2261434873461805e-05,
1033
+ "loss": 0.0959,
1034
+ "step": 34500
1035
+ },
1036
+ {
1037
+ "epoch": 23.49,
1038
+ "eval_loss": 0.48040756583213806,
1039
+ "eval_runtime": 129.4236,
1040
+ "eval_samples_per_second": 15.33,
1041
+ "eval_steps_per_second": 1.916,
1042
+ "eval_wer": 0.28978233900981165,
1043
+ "step": 34500
1044
+ },
1045
+ {
1046
+ "epoch": 23.83,
1047
+ "learning_rate": 2.110053401439517e-05,
1048
+ "loss": 0.098,
1049
+ "step": 35000
1050
+ },
1051
+ {
1052
+ "epoch": 23.83,
1053
+ "eval_loss": 0.501395046710968,
1054
+ "eval_runtime": 129.507,
1055
+ "eval_samples_per_second": 15.32,
1056
+ "eval_steps_per_second": 1.915,
1057
+ "eval_wer": 0.291699560166911,
1058
+ "step": 35000
1059
+ },
1060
+ {
1061
+ "epoch": 24.17,
1062
+ "learning_rate": 1.9939633155328534e-05,
1063
+ "loss": 0.0973,
1064
+ "step": 35500
1065
+ },
1066
+ {
1067
+ "epoch": 24.17,
1068
+ "eval_loss": 0.5193932056427002,
1069
+ "eval_runtime": 129.3856,
1070
+ "eval_samples_per_second": 15.334,
1071
+ "eval_steps_per_second": 1.917,
1072
+ "eval_wer": 0.28955678357956466,
1073
+ "step": 35500
1074
+ },
1075
+ {
1076
+ "epoch": 24.51,
1077
+ "learning_rate": 1.87787322962619e-05,
1078
+ "loss": 0.0876,
1079
+ "step": 36000
1080
+ },
1081
+ {
1082
+ "epoch": 24.51,
1083
+ "eval_loss": 0.5203258395195007,
1084
+ "eval_runtime": 125.0985,
1085
+ "eval_samples_per_second": 15.859,
1086
+ "eval_steps_per_second": 1.982,
1087
+ "eval_wer": 0.28718845156197137,
1088
+ "step": 36000
1089
+ },
1090
+ {
1091
+ "epoch": 24.85,
1092
+ "learning_rate": 1.7617831437195264e-05,
1093
+ "loss": 0.0902,
1094
+ "step": 36500
1095
+ },
1096
+ {
1097
+ "epoch": 24.85,
1098
+ "eval_loss": 0.536376953125,
1099
+ "eval_runtime": 113.525,
1100
+ "eval_samples_per_second": 17.476,
1101
+ "eval_steps_per_second": 2.185,
1102
+ "eval_wer": 0.29136122702154055,
1103
+ "step": 36500
1104
+ },
1105
+ {
1106
+ "epoch": 25.19,
1107
+ "learning_rate": 1.645693057812863e-05,
1108
+ "loss": 0.0889,
1109
+ "step": 37000
1110
+ },
1111
+ {
1112
+ "epoch": 25.19,
1113
+ "eval_loss": 0.531440019607544,
1114
+ "eval_runtime": 123.4292,
1115
+ "eval_samples_per_second": 16.074,
1116
+ "eval_steps_per_second": 2.009,
1117
+ "eval_wer": 0.2880906732829593,
1118
+ "step": 37000
1119
+ },
1120
+ {
1121
+ "epoch": 25.53,
1122
+ "learning_rate": 1.5296029719061993e-05,
1123
+ "loss": 0.0865,
1124
+ "step": 37500
1125
+ },
1126
+ {
1127
+ "epoch": 25.53,
1128
+ "eval_loss": 0.5107194185256958,
1129
+ "eval_runtime": 129.5881,
1130
+ "eval_samples_per_second": 15.31,
1131
+ "eval_steps_per_second": 1.914,
1132
+ "eval_wer": 0.285271230404872,
1133
+ "step": 37500
1134
+ },
1135
+ {
1136
+ "epoch": 25.87,
1137
+ "learning_rate": 1.4135128859995356e-05,
1138
+ "loss": 0.0859,
1139
+ "step": 38000
1140
+ },
1141
+ {
1142
+ "epoch": 25.87,
1143
+ "eval_loss": 0.5254319310188293,
1144
+ "eval_runtime": 128.9308,
1145
+ "eval_samples_per_second": 15.388,
1146
+ "eval_steps_per_second": 1.924,
1147
+ "eval_wer": 0.28780872899515053,
1148
+ "step": 38000
1149
+ },
1150
+ {
1151
+ "epoch": 26.21,
1152
+ "learning_rate": 1.2974228000928721e-05,
1153
+ "loss": 0.0813,
1154
+ "step": 38500
1155
+ },
1156
+ {
1157
+ "epoch": 26.21,
1158
+ "eval_loss": 0.5275471806526184,
1159
+ "eval_runtime": 110.8092,
1160
+ "eval_samples_per_second": 17.905,
1161
+ "eval_steps_per_second": 2.238,
1162
+ "eval_wer": 0.2846509529716928,
1163
+ "step": 38500
1164
+ },
1165
+ {
1166
+ "epoch": 26.55,
1167
+ "learning_rate": 1.1813327141862086e-05,
1168
+ "loss": 0.0881,
1169
+ "step": 39000
1170
+ },
1171
+ {
1172
+ "epoch": 26.55,
1173
+ "eval_loss": 0.5125020742416382,
1174
+ "eval_runtime": 113.2218,
1175
+ "eval_samples_per_second": 17.523,
1176
+ "eval_steps_per_second": 2.19,
1177
+ "eval_wer": 0.28324123153264913,
1178
+ "step": 39000
1179
+ },
1180
+ {
1181
+ "epoch": 26.89,
1182
+ "learning_rate": 1.0654748084513583e-05,
1183
+ "loss": 0.0822,
1184
+ "step": 39500
1185
+ },
1186
+ {
1187
+ "epoch": 26.89,
1188
+ "eval_loss": 0.5309813618659973,
1189
+ "eval_runtime": 129.6599,
1190
+ "eval_samples_per_second": 15.302,
1191
+ "eval_steps_per_second": 1.913,
1192
+ "eval_wer": 0.28634261869854516,
1193
+ "step": 39500
1194
+ },
1195
+ {
1196
+ "epoch": 27.23,
1197
+ "learning_rate": 9.493847225446947e-06,
1198
+ "loss": 0.0837,
1199
+ "step": 40000
1200
+ },
1201
+ {
1202
+ "epoch": 27.23,
1203
+ "eval_loss": 0.5188203454017639,
1204
+ "eval_runtime": 122.3408,
1205
+ "eval_samples_per_second": 16.217,
1206
+ "eval_steps_per_second": 2.027,
1207
+ "eval_wer": 0.28312845381752566,
1208
+ "step": 40000
1209
+ },
1210
+ {
1211
+ "epoch": 27.57,
1212
+ "learning_rate": 8.332946366380312e-06,
1213
+ "loss": 0.0823,
1214
+ "step": 40500
1215
+ },
1216
+ {
1217
+ "epoch": 27.57,
1218
+ "eval_loss": 0.5201263427734375,
1219
+ "eval_runtime": 118.4812,
1220
+ "eval_samples_per_second": 16.745,
1221
+ "eval_steps_per_second": 2.093,
1222
+ "eval_wer": 0.28132401037554977,
1223
+ "step": 40500
1224
+ },
1225
+ {
1226
+ "epoch": 27.91,
1227
+ "learning_rate": 7.172045507313675e-06,
1228
+ "loss": 0.0768,
1229
+ "step": 41000
1230
+ },
1231
+ {
1232
+ "epoch": 27.91,
1233
+ "eval_loss": 0.5197951793670654,
1234
+ "eval_runtime": 112.4253,
1235
+ "eval_samples_per_second": 17.647,
1236
+ "eval_steps_per_second": 2.206,
1237
+ "eval_wer": 0.282169843238976,
1238
+ "step": 41000
1239
+ },
1240
+ {
1241
+ "epoch": 28.25,
1242
+ "learning_rate": 6.013466449965173e-06,
1243
+ "loss": 0.0858,
1244
+ "step": 41500
1245
+ },
1246
+ {
1247
+ "epoch": 28.25,
1248
+ "eval_loss": 0.5245384573936462,
1249
+ "eval_runtime": 128.7266,
1250
+ "eval_samples_per_second": 15.413,
1251
+ "eval_steps_per_second": 1.927,
1252
+ "eval_wer": 0.2786737340701477,
1253
+ "step": 41500
1254
+ },
1255
+ {
1256
+ "epoch": 28.59,
1257
+ "learning_rate": 4.8525655908985375e-06,
1258
+ "loss": 0.0757,
1259
+ "step": 42000
1260
+ },
1261
+ {
1262
+ "epoch": 28.59,
1263
+ "eval_loss": 0.5289037227630615,
1264
+ "eval_runtime": 112.5429,
1265
+ "eval_samples_per_second": 17.629,
1266
+ "eval_steps_per_second": 2.204,
1267
+ "eval_wer": 0.2766437351979249,
1268
+ "step": 42000
1269
+ },
1270
+ {
1271
+ "epoch": 28.93,
1272
+ "learning_rate": 3.6916647318319014e-06,
1273
+ "loss": 0.0728,
1274
+ "step": 42500
1275
+ },
1276
+ {
1277
+ "epoch": 28.93,
1278
+ "eval_loss": 0.5263972282409668,
1279
+ "eval_runtime": 128.0915,
1280
+ "eval_samples_per_second": 15.489,
1281
+ "eval_steps_per_second": 1.936,
1282
+ "eval_wer": 0.2779970677794068,
1283
+ "step": 42500
1284
+ },
1285
+ {
1286
+ "epoch": 29.27,
1287
+ "learning_rate": 2.5307638727652657e-06,
1288
+ "loss": 0.0761,
1289
+ "step": 43000
1290
+ },
1291
+ {
1292
+ "epoch": 29.27,
1293
+ "eval_loss": 0.5295293927192688,
1294
+ "eval_runtime": 128.9693,
1295
+ "eval_samples_per_second": 15.384,
1296
+ "eval_steps_per_second": 1.923,
1297
+ "eval_wer": 0.2770384572008571,
1298
+ "step": 43000
1299
+ },
1300
+ {
1301
+ "epoch": 29.61,
1302
+ "learning_rate": 1.3698630136986302e-06,
1303
+ "loss": 0.0739,
1304
+ "step": 43500
1305
+ },
1306
+ {
1307
+ "epoch": 29.61,
1308
+ "eval_loss": 0.5267478823661804,
1309
+ "eval_runtime": 130.4434,
1310
+ "eval_samples_per_second": 15.21,
1311
+ "eval_steps_per_second": 1.901,
1312
+ "eval_wer": 0.2776023457764746,
1313
+ "step": 43500
1314
+ },
1315
+ {
1316
+ "epoch": 29.95,
1317
+ "learning_rate": 2.0896215463199446e-07,
1318
+ "loss": 0.0734,
1319
+ "step": 44000
1320
+ },
1321
+ {
1322
+ "epoch": 29.95,
1323
+ "eval_loss": 0.5276228189468384,
1324
+ "eval_runtime": 110.9362,
1325
+ "eval_samples_per_second": 17.884,
1326
+ "eval_steps_per_second": 2.236,
1327
+ "eval_wer": 0.2766437351979249,
1328
+ "step": 44000
1329
+ }
1330
+ ],
1331
+ "max_steps": 44070,
1332
+ "num_train_epochs": 30,
1333
+ "total_flos": 1.401945817013882e+19,
1334
+ "trial_name": null,
1335
+ "trial_params": null
1336
+ }
wav2vec2_esp_30h/checkpoint-44000/training_args.bin ADDED
Binary file (3 kB). View file
 
wav2vec2_esp_30h/runs/Mar08_02-48-55_mint/1678211683.9648502/events.out.tfevents.1678211683.mint ADDED
Binary file (4.57 kB). View file
 
wav2vec2_esp_30h/runs/Mar08_02-48-55_mint/events.out.tfevents.1678211683.mint ADDED
Binary file (9.71 kB). View file
 
wav2vec2_esp_30h/runs/Mar08_03-58-32_mint/1678215863.173738/events.out.tfevents.1678215863.mint ADDED
Binary file (4.57 kB). View file
 
wav2vec2_esp_30h/runs/Mar08_03-58-32_mint/events.out.tfevents.1678215863.mint ADDED
Binary file (47.6 kB). View file
 
wav2vec2_esp_5h/checkpoint-8000/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base",
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": 0.1,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
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": 0.1,
52
+ "feat_quantizer_dropout": 0.0,
53
+ "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
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": 37,
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": "4.17.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 38,
116
+ "xvector_output_dim": 512
117
+ }
wav2vec2_esp_5h/checkpoint-8000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42646df58671960375957cbd8f37fd7a40c3231761fa6deec20b76aab5d09d31
3
+ size 721723525
wav2vec2_esp_5h/checkpoint-8000/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
+ }
wav2vec2_esp_5h/checkpoint-8000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce65b43cf3bfb9430a3d7c78e0144ff466f29c79cf64d87f6c644ef8cc7b74bf
3
+ size 377691873
wav2vec2_esp_5h/checkpoint-8000/rng_state.pth ADDED
Binary file (14.6 kB). View file
 
wav2vec2_esp_5h/checkpoint-8000/scaler.pt ADDED
Binary file (557 Bytes). View file
 
wav2vec2_esp_5h/checkpoint-8000/scheduler.pt ADDED
Binary file (627 Bytes). View file