Hugging Face's logo

saswatapcs17
/
whisper-small-hi

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
whisper
Inference Endpoints
Model card Files Metrics Community
saswatapcs17 commited on
Commit
c8f2bd5
•
1 Parent(s): 8b3d47d

Upload with huggingface_hub

Browse files
Files changed (42) hide show
  1. checkpoint-1000/config.json +51 -0
  2. checkpoint-1000/generation_config.json +221 -0
  3. checkpoint-1000/optimizer.pt +3 -0
  4. checkpoint-1000/preprocessor_config.json +14 -0
  5. checkpoint-1000/pytorch_model.bin +3 -0
  6. checkpoint-1000/rng_state.pth +3 -0
  7. checkpoint-1000/scaler.pt +3 -0
  8. checkpoint-1000/scheduler.pt +3 -0
  9. checkpoint-1000/trainer_state.json +265 -0
  10. checkpoint-1000/training_args.bin +3 -0
  11. checkpoint-2000/config.json +51 -0
  12. checkpoint-2000/generation_config.json +221 -0
  13. checkpoint-2000/optimizer.pt +3 -0
  14. checkpoint-2000/preprocessor_config.json +14 -0
  15. checkpoint-2000/pytorch_model.bin +3 -0
  16. checkpoint-2000/rng_state.pth +3 -0
  17. checkpoint-2000/scaler.pt +3 -0
  18. checkpoint-2000/scheduler.pt +3 -0
  19. checkpoint-2000/trainer_state.json +514 -0
  20. checkpoint-2000/training_args.bin +3 -0
  21. checkpoint-3000/config.json +51 -0
  22. checkpoint-3000/generation_config.json +221 -0
  23. checkpoint-3000/optimizer.pt +3 -0
  24. checkpoint-3000/preprocessor_config.json +14 -0
  25. checkpoint-3000/pytorch_model.bin +3 -0
  26. checkpoint-3000/rng_state.pth +3 -0
  27. checkpoint-3000/scaler.pt +3 -0
  28. checkpoint-3000/scheduler.pt +3 -0
  29. checkpoint-3000/trainer_state.json +763 -0
  30. checkpoint-3000/training_args.bin +3 -0
  31. checkpoint-4000/config.json +51 -0
  32. checkpoint-4000/generation_config.json +221 -0
  33. checkpoint-4000/optimizer.pt +3 -0
  34. checkpoint-4000/preprocessor_config.json +14 -0
  35. checkpoint-4000/pytorch_model.bin +3 -0
  36. checkpoint-4000/rng_state.pth +3 -0
  37. checkpoint-4000/scaler.pt +3 -0
  38. checkpoint-4000/scheduler.pt +3 -0
  39. checkpoint-4000/trainer_state.json +1012 -0
  40. checkpoint-4000/training_args.bin +3 -0
  41. runs/Apr07_19-09-15_6672b5533821/1680894671.4432056/events.out.tfevents.1680894671.6672b5533821.633.1 +3 -0
  42. runs/Apr07_19-09-15_6672b5533821/events.out.tfevents.1680894671.6672b5533821.633.0 +3 -0
checkpoint-1000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.0,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.28.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.28.0.dev0"
221
+ }
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f27e147daa1fb1803753410fc34249256bc9c8647e2c090b2a8be4dca9b2bbbf
3
+ size 1934161093
checkpoint-1000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-1000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49809bd37ee754e6f7501ecb40eccc494a921229e3507f77045a4aa07d06420a
3
+ size 967102601
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4200c5f9a588ac43124b21dcf1afb78b23475566dc0dd889d318f778b0c0edb9
3
+ size 14511
checkpoint-1000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26dbe439de3dd73741374b46df74ee981e9ad3b04a0c5c8a564c625a65b8aee5
3
+ size 557
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90dd2338805bd3a69fa663686537f5e5065d09a0615f56ffc3e4278e4e48b9dd
3
+ size 627
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 92.3076923076923,
3
+ "best_model_checkpoint": "./whisper-small-hi/checkpoint-1000",
4
+ "epoch": 1000.0,
5
+ "global_step": 1000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 25.0,
12
+ "learning_rate": 4.6000000000000004e-07,
13
+ "loss": 2.7845,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 50.0,
18
+ "learning_rate": 9.600000000000001e-07,
19
+ "loss": 1.8784,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 75.0,
24
+ "learning_rate": 1.46e-06,
25
+ "loss": 0.7949,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 100.0,
30
+ "learning_rate": 1.9600000000000003e-06,
31
+ "loss": 0.2619,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 125.0,
36
+ "learning_rate": 2.46e-06,
37
+ "loss": 0.1719,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 150.0,
42
+ "learning_rate": 2.96e-06,
43
+ "loss": 0.1447,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 175.0,
48
+ "learning_rate": 3.46e-06,
49
+ "loss": 0.1234,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 200.0,
54
+ "learning_rate": 3.96e-06,
55
+ "loss": 0.0997,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 225.0,
60
+ "learning_rate": 4.4600000000000005e-06,
61
+ "loss": 0.0658,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 250.0,
66
+ "learning_rate": 4.960000000000001e-06,
67
+ "loss": 0.0054,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 275.0,
72
+ "learning_rate": 5.460000000000001e-06,
73
+ "loss": 0.0006,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 300.0,
78
+ "learning_rate": 5.9600000000000005e-06,
79
+ "loss": 0.0003,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 325.0,
84
+ "learning_rate": 6.460000000000001e-06,
85
+ "loss": 0.0002,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 350.0,
90
+ "learning_rate": 6.96e-06,
91
+ "loss": 0.0002,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 375.0,
96
+ "learning_rate": 7.4600000000000006e-06,
97
+ "loss": 0.0002,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 400.0,
102
+ "learning_rate": 7.960000000000002e-06,
103
+ "loss": 0.0001,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 425.0,
108
+ "learning_rate": 8.46e-06,
109
+ "loss": 0.0001,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 450.0,
114
+ "learning_rate": 8.96e-06,
115
+ "loss": 0.0001,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 475.0,
120
+ "learning_rate": 9.460000000000001e-06,
121
+ "loss": 0.0001,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 500.0,
126
+ "learning_rate": 9.960000000000001e-06,
127
+ "loss": 0.0001,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 525.0,
132
+ "learning_rate": 9.934285714285715e-06,
133
+ "loss": 0.0001,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 550.0,
138
+ "learning_rate": 9.862857142857144e-06,
139
+ "loss": 0.0001,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 575.0,
144
+ "learning_rate": 9.791428571428571e-06,
145
+ "loss": 0.0001,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 600.0,
150
+ "learning_rate": 9.72e-06,
151
+ "loss": 0.0001,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 625.0,
156
+ "learning_rate": 9.648571428571429e-06,
157
+ "loss": 0.0001,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 650.0,
162
+ "learning_rate": 9.577142857142858e-06,
163
+ "loss": 0.0001,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 675.0,
168
+ "learning_rate": 9.505714285714287e-06,
169
+ "loss": 0.0,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 700.0,
174
+ "learning_rate": 9.434285714285714e-06,
175
+ "loss": 0.0,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 725.0,
180
+ "learning_rate": 9.362857142857143e-06,
181
+ "loss": 0.0,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 750.0,
186
+ "learning_rate": 9.291428571428572e-06,
187
+ "loss": 0.0,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 775.0,
192
+ "learning_rate": 9.220000000000002e-06,
193
+ "loss": 0.0,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 800.0,
198
+ "learning_rate": 9.148571428571429e-06,
199
+ "loss": 0.0,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 825.0,
204
+ "learning_rate": 9.077142857142858e-06,
205
+ "loss": 0.0,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 850.0,
210
+ "learning_rate": 9.005714285714287e-06,
211
+ "loss": 0.0,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 875.0,
216
+ "learning_rate": 8.934285714285716e-06,
217
+ "loss": 0.0,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 900.0,
222
+ "learning_rate": 8.862857142857143e-06,
223
+ "loss": 0.0,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 925.0,
228
+ "learning_rate": 8.791428571428572e-06,
229
+ "loss": 0.0,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 950.0,
234
+ "learning_rate": 8.720000000000001e-06,
235
+ "loss": 0.0,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 975.0,
240
+ "learning_rate": 8.64857142857143e-06,
241
+ "loss": 0.0,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 1000.0,
246
+ "learning_rate": 8.577142857142858e-06,
247
+ "loss": 0.0,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 1000.0,
252
+ "eval_loss": 1.5850915908813477,
253
+ "eval_runtime": 3.2322,
254
+ "eval_samples_per_second": 0.619,
255
+ "eval_steps_per_second": 0.309,
256
+ "eval_wer": 92.3076923076923,
257
+ "step": 1000
258
+ }
259
+ ],
260
+ "max_steps": 4000,
261
+ "num_train_epochs": 4000,
262
+ "total_flos": 2.8858540032e+18,
263
+ "trial_name": null,
264
+ "trial_params": null
265
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6b5d0e7f2aaac3a58da2d607fd67d4ade51924fd9bf60c5e1dea221d52e1338
3
+ size 3771
checkpoint-2000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.0,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.28.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-2000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.28.0.dev0"
221
+ }
checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2186182e843846d42b2fb28e9b95c677aa9a650eb4e84068e5de79355cf8db9
3
+ size 1934161093
checkpoint-2000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-2000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee8f88daa13e0fb1015386bb80fa8ff4da3b4efa75476cfce7c627e85ae65f09
3
+ size 967102601
checkpoint-2000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8addf63c5b2d987369e8d3a9882e5db53e629ed25a75c300420b6f333075e3f3
3
+ size 14575
checkpoint-2000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93dcbb27405421069b5f31003e3f224646e375ff4c9d7ba4b0d265614f660d0f
3
+ size 557
checkpoint-2000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9536b63cd02db26062d122a42fc284c452283fd0cb8cc2a36d3aee25ee14708d
3
+ size 627
checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,514 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 84.61538461538461,
3
+ "best_model_checkpoint": "./whisper-small-hi/checkpoint-2000",
4
+ "epoch": 2000.0,
5
+ "global_step": 2000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 25.0,
12
+ "learning_rate": 4.6000000000000004e-07,
13
+ "loss": 2.7845,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 50.0,
18
+ "learning_rate": 9.600000000000001e-07,
19
+ "loss": 1.8784,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 75.0,
24
+ "learning_rate": 1.46e-06,
25
+ "loss": 0.7949,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 100.0,
30
+ "learning_rate": 1.9600000000000003e-06,
31
+ "loss": 0.2619,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 125.0,
36
+ "learning_rate": 2.46e-06,
37
+ "loss": 0.1719,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 150.0,
42
+ "learning_rate": 2.96e-06,
43
+ "loss": 0.1447,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 175.0,
48
+ "learning_rate": 3.46e-06,
49
+ "loss": 0.1234,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 200.0,
54
+ "learning_rate": 3.96e-06,
55
+ "loss": 0.0997,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 225.0,
60
+ "learning_rate": 4.4600000000000005e-06,
61
+ "loss": 0.0658,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 250.0,
66
+ "learning_rate": 4.960000000000001e-06,
67
+ "loss": 0.0054,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 275.0,
72
+ "learning_rate": 5.460000000000001e-06,
73
+ "loss": 0.0006,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 300.0,
78
+ "learning_rate": 5.9600000000000005e-06,
79
+ "loss": 0.0003,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 325.0,
84
+ "learning_rate": 6.460000000000001e-06,
85
+ "loss": 0.0002,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 350.0,
90
+ "learning_rate": 6.96e-06,
91
+ "loss": 0.0002,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 375.0,
96
+ "learning_rate": 7.4600000000000006e-06,
97
+ "loss": 0.0002,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 400.0,
102
+ "learning_rate": 7.960000000000002e-06,
103
+ "loss": 0.0001,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 425.0,
108
+ "learning_rate": 8.46e-06,
109
+ "loss": 0.0001,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 450.0,
114
+ "learning_rate": 8.96e-06,
115
+ "loss": 0.0001,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 475.0,
120
+ "learning_rate": 9.460000000000001e-06,
121
+ "loss": 0.0001,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 500.0,
126
+ "learning_rate": 9.960000000000001e-06,
127
+ "loss": 0.0001,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 525.0,
132
+ "learning_rate": 9.934285714285715e-06,
133
+ "loss": 0.0001,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 550.0,
138
+ "learning_rate": 9.862857142857144e-06,
139
+ "loss": 0.0001,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 575.0,
144
+ "learning_rate": 9.791428571428571e-06,
145
+ "loss": 0.0001,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 600.0,
150
+ "learning_rate": 9.72e-06,
151
+ "loss": 0.0001,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 625.0,
156
+ "learning_rate": 9.648571428571429e-06,
157
+ "loss": 0.0001,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 650.0,
162
+ "learning_rate": 9.577142857142858e-06,
163
+ "loss": 0.0001,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 675.0,
168
+ "learning_rate": 9.505714285714287e-06,
169
+ "loss": 0.0,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 700.0,
174
+ "learning_rate": 9.434285714285714e-06,
175
+ "loss": 0.0,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 725.0,
180
+ "learning_rate": 9.362857142857143e-06,
181
+ "loss": 0.0,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 750.0,
186
+ "learning_rate": 9.291428571428572e-06,
187
+ "loss": 0.0,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 775.0,
192
+ "learning_rate": 9.220000000000002e-06,
193
+ "loss": 0.0,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 800.0,
198
+ "learning_rate": 9.148571428571429e-06,
199
+ "loss": 0.0,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 825.0,
204
+ "learning_rate": 9.077142857142858e-06,
205
+ "loss": 0.0,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 850.0,
210
+ "learning_rate": 9.005714285714287e-06,
211
+ "loss": 0.0,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 875.0,
216
+ "learning_rate": 8.934285714285716e-06,
217
+ "loss": 0.0,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 900.0,
222
+ "learning_rate": 8.862857142857143e-06,
223
+ "loss": 0.0,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 925.0,
228
+ "learning_rate": 8.791428571428572e-06,
229
+ "loss": 0.0,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 950.0,
234
+ "learning_rate": 8.720000000000001e-06,
235
+ "loss": 0.0,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 975.0,
240
+ "learning_rate": 8.64857142857143e-06,
241
+ "loss": 0.0,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 1000.0,
246
+ "learning_rate": 8.577142857142858e-06,
247
+ "loss": 0.0,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 1000.0,
252
+ "eval_loss": 1.5850915908813477,
253
+ "eval_runtime": 3.2322,
254
+ "eval_samples_per_second": 0.619,
255
+ "eval_steps_per_second": 0.309,
256
+ "eval_wer": 92.3076923076923,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 1025.0,
261
+ "learning_rate": 8.505714285714287e-06,
262
+ "loss": 0.0,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1050.0,
267
+ "learning_rate": 8.434285714285716e-06,
268
+ "loss": 0.0,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1075.0,
273
+ "learning_rate": 8.362857142857143e-06,
274
+ "loss": 0.0,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1100.0,
279
+ "learning_rate": 8.291428571428572e-06,
280
+ "loss": 0.0,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1125.0,
285
+ "learning_rate": 8.220000000000001e-06,
286
+ "loss": 0.0,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1150.0,
291
+ "learning_rate": 8.148571428571428e-06,
292
+ "loss": 0.0,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1175.0,
297
+ "learning_rate": 8.077142857142857e-06,
298
+ "loss": 0.0,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1200.0,
303
+ "learning_rate": 8.005714285714286e-06,
304
+ "loss": 0.0,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1225.0,
309
+ "learning_rate": 7.934285714285715e-06,
310
+ "loss": 0.0,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1250.0,
315
+ "learning_rate": 7.862857142857143e-06,
316
+ "loss": 0.0,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1275.0,
321
+ "learning_rate": 7.791428571428572e-06,
322
+ "loss": 0.0,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1300.0,
327
+ "learning_rate": 7.72e-06,
328
+ "loss": 0.0,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1325.0,
333
+ "learning_rate": 7.64857142857143e-06,
334
+ "loss": 0.0,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1350.0,
339
+ "learning_rate": 7.577142857142857e-06,
340
+ "loss": 0.0,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1375.0,
345
+ "learning_rate": 7.505714285714286e-06,
346
+ "loss": 0.0,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1400.0,
351
+ "learning_rate": 7.434285714285715e-06,
352
+ "loss": 0.0,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1425.0,
357
+ "learning_rate": 7.362857142857144e-06,
358
+ "loss": 0.0,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1450.0,
363
+ "learning_rate": 7.291428571428571e-06,
364
+ "loss": 0.0,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1475.0,
369
+ "learning_rate": 7.22e-06,
370
+ "loss": 0.0,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1500.0,
375
+ "learning_rate": 7.148571428571429e-06,
376
+ "loss": 0.0,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1525.0,
381
+ "learning_rate": 7.077142857142858e-06,
382
+ "loss": 0.0,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1550.0,
387
+ "learning_rate": 7.0057142857142865e-06,
388
+ "loss": 0.0,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1575.0,
393
+ "learning_rate": 6.934285714285715e-06,
394
+ "loss": 0.0,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1600.0,
399
+ "learning_rate": 6.862857142857144e-06,
400
+ "loss": 0.0,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1625.0,
405
+ "learning_rate": 6.791428571428572e-06,
406
+ "loss": 0.0,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1650.0,
411
+ "learning_rate": 6.720000000000001e-06,
412
+ "loss": 0.0,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1675.0,
417
+ "learning_rate": 6.648571428571429e-06,
418
+ "loss": 0.0,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1700.0,
423
+ "learning_rate": 6.577142857142857e-06,
424
+ "loss": 0.0,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1725.0,
429
+ "learning_rate": 6.505714285714286e-06,
430
+ "loss": 0.0,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1750.0,
435
+ "learning_rate": 6.434285714285715e-06,
436
+ "loss": 0.0,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1775.0,
441
+ "learning_rate": 6.3628571428571426e-06,
442
+ "loss": 0.0,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1800.0,
447
+ "learning_rate": 6.2914285714285716e-06,
448
+ "loss": 0.0,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1825.0,
453
+ "learning_rate": 6.220000000000001e-06,
454
+ "loss": 0.0,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1850.0,
459
+ "learning_rate": 6.14857142857143e-06,
460
+ "loss": 0.0,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1875.0,
465
+ "learning_rate": 6.077142857142858e-06,
466
+ "loss": 0.0,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1900.0,
471
+ "learning_rate": 6.005714285714286e-06,
472
+ "loss": 0.0,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1925.0,
477
+ "learning_rate": 5.934285714285715e-06,
478
+ "loss": 0.0,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1950.0,
483
+ "learning_rate": 5.862857142857143e-06,
484
+ "loss": 0.0,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1975.0,
489
+ "learning_rate": 5.791428571428572e-06,
490
+ "loss": 0.0,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 2000.0,
495
+ "learning_rate": 5.72e-06,
496
+ "loss": 0.0,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 2000.0,
501
+ "eval_loss": 2.00115704536438,
502
+ "eval_runtime": 3.6044,
503
+ "eval_samples_per_second": 0.555,
504
+ "eval_steps_per_second": 0.277,
505
+ "eval_wer": 84.61538461538461,
506
+ "step": 2000
507
+ }
508
+ ],
509
+ "max_steps": 4000,
510
+ "num_train_epochs": 4000,
511
+ "total_flos": 5.7717080064e+18,
512
+ "trial_name": null,
513
+ "trial_params": null
514
+ }
checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6b5d0e7f2aaac3a58da2d607fd67d4ade51924fd9bf60c5e1dea221d52e1338
3
+ size 3771
checkpoint-3000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.0,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.28.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-3000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.28.0.dev0"
221
+ }
checkpoint-3000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77f1d5bad12f494c98b1e00a6c8f4d3567b7f922cb667c8be813793da8d38750
3
+ size 1934161093
checkpoint-3000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-3000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe2009c02aa2c46dcbbc65a0cc96a877c299c1c822e20405a8d48f91729cf0cc
3
+ size 967102601
checkpoint-3000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a82fcee28124ecb71e0e552883a2c66164293dff7c17b0045df92266f8d34c55
3
+ size 14575
checkpoint-3000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d2a59dd750539c973981813825488c88f87133aca172b473f2a8e03b35e2802
3
+ size 557
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7714285e54100d4ac40dc418b750401bda88dd754c30602720da851473b84959
3
+ size 627
checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,763 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 84.61538461538461,
3
+ "best_model_checkpoint": "./whisper-small-hi/checkpoint-2000",
4
+ "epoch": 3000.0,
5
+ "global_step": 3000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 25.0,
12
+ "learning_rate": 4.6000000000000004e-07,
13
+ "loss": 2.7845,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 50.0,
18
+ "learning_rate": 9.600000000000001e-07,
19
+ "loss": 1.8784,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 75.0,
24
+ "learning_rate": 1.46e-06,
25
+ "loss": 0.7949,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 100.0,
30
+ "learning_rate": 1.9600000000000003e-06,
31
+ "loss": 0.2619,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 125.0,
36
+ "learning_rate": 2.46e-06,
37
+ "loss": 0.1719,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 150.0,
42
+ "learning_rate": 2.96e-06,
43
+ "loss": 0.1447,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 175.0,
48
+ "learning_rate": 3.46e-06,
49
+ "loss": 0.1234,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 200.0,
54
+ "learning_rate": 3.96e-06,
55
+ "loss": 0.0997,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 225.0,
60
+ "learning_rate": 4.4600000000000005e-06,
61
+ "loss": 0.0658,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 250.0,
66
+ "learning_rate": 4.960000000000001e-06,
67
+ "loss": 0.0054,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 275.0,
72
+ "learning_rate": 5.460000000000001e-06,
73
+ "loss": 0.0006,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 300.0,
78
+ "learning_rate": 5.9600000000000005e-06,
79
+ "loss": 0.0003,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 325.0,
84
+ "learning_rate": 6.460000000000001e-06,
85
+ "loss": 0.0002,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 350.0,
90
+ "learning_rate": 6.96e-06,
91
+ "loss": 0.0002,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 375.0,
96
+ "learning_rate": 7.4600000000000006e-06,
97
+ "loss": 0.0002,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 400.0,
102
+ "learning_rate": 7.960000000000002e-06,
103
+ "loss": 0.0001,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 425.0,
108
+ "learning_rate": 8.46e-06,
109
+ "loss": 0.0001,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 450.0,
114
+ "learning_rate": 8.96e-06,
115
+ "loss": 0.0001,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 475.0,
120
+ "learning_rate": 9.460000000000001e-06,
121
+ "loss": 0.0001,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 500.0,
126
+ "learning_rate": 9.960000000000001e-06,
127
+ "loss": 0.0001,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 525.0,
132
+ "learning_rate": 9.934285714285715e-06,
133
+ "loss": 0.0001,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 550.0,
138
+ "learning_rate": 9.862857142857144e-06,
139
+ "loss": 0.0001,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 575.0,
144
+ "learning_rate": 9.791428571428571e-06,
145
+ "loss": 0.0001,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 600.0,
150
+ "learning_rate": 9.72e-06,
151
+ "loss": 0.0001,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 625.0,
156
+ "learning_rate": 9.648571428571429e-06,
157
+ "loss": 0.0001,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 650.0,
162
+ "learning_rate": 9.577142857142858e-06,
163
+ "loss": 0.0001,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 675.0,
168
+ "learning_rate": 9.505714285714287e-06,
169
+ "loss": 0.0,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 700.0,
174
+ "learning_rate": 9.434285714285714e-06,
175
+ "loss": 0.0,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 725.0,
180
+ "learning_rate": 9.362857142857143e-06,
181
+ "loss": 0.0,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 750.0,
186
+ "learning_rate": 9.291428571428572e-06,
187
+ "loss": 0.0,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 775.0,
192
+ "learning_rate": 9.220000000000002e-06,
193
+ "loss": 0.0,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 800.0,
198
+ "learning_rate": 9.148571428571429e-06,
199
+ "loss": 0.0,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 825.0,
204
+ "learning_rate": 9.077142857142858e-06,
205
+ "loss": 0.0,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 850.0,
210
+ "learning_rate": 9.005714285714287e-06,
211
+ "loss": 0.0,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 875.0,
216
+ "learning_rate": 8.934285714285716e-06,
217
+ "loss": 0.0,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 900.0,
222
+ "learning_rate": 8.862857142857143e-06,
223
+ "loss": 0.0,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 925.0,
228
+ "learning_rate": 8.791428571428572e-06,
229
+ "loss": 0.0,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 950.0,
234
+ "learning_rate": 8.720000000000001e-06,
235
+ "loss": 0.0,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 975.0,
240
+ "learning_rate": 8.64857142857143e-06,
241
+ "loss": 0.0,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 1000.0,
246
+ "learning_rate": 8.577142857142858e-06,
247
+ "loss": 0.0,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 1000.0,
252
+ "eval_loss": 1.5850915908813477,
253
+ "eval_runtime": 3.2322,
254
+ "eval_samples_per_second": 0.619,
255
+ "eval_steps_per_second": 0.309,
256
+ "eval_wer": 92.3076923076923,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 1025.0,
261
+ "learning_rate": 8.505714285714287e-06,
262
+ "loss": 0.0,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1050.0,
267
+ "learning_rate": 8.434285714285716e-06,
268
+ "loss": 0.0,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1075.0,
273
+ "learning_rate": 8.362857142857143e-06,
274
+ "loss": 0.0,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1100.0,
279
+ "learning_rate": 8.291428571428572e-06,
280
+ "loss": 0.0,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1125.0,
285
+ "learning_rate": 8.220000000000001e-06,
286
+ "loss": 0.0,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1150.0,
291
+ "learning_rate": 8.148571428571428e-06,
292
+ "loss": 0.0,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1175.0,
297
+ "learning_rate": 8.077142857142857e-06,
298
+ "loss": 0.0,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1200.0,
303
+ "learning_rate": 8.005714285714286e-06,
304
+ "loss": 0.0,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1225.0,
309
+ "learning_rate": 7.934285714285715e-06,
310
+ "loss": 0.0,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1250.0,
315
+ "learning_rate": 7.862857142857143e-06,
316
+ "loss": 0.0,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1275.0,
321
+ "learning_rate": 7.791428571428572e-06,
322
+ "loss": 0.0,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1300.0,
327
+ "learning_rate": 7.72e-06,
328
+ "loss": 0.0,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1325.0,
333
+ "learning_rate": 7.64857142857143e-06,
334
+ "loss": 0.0,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1350.0,
339
+ "learning_rate": 7.577142857142857e-06,
340
+ "loss": 0.0,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1375.0,
345
+ "learning_rate": 7.505714285714286e-06,
346
+ "loss": 0.0,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1400.0,
351
+ "learning_rate": 7.434285714285715e-06,
352
+ "loss": 0.0,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1425.0,
357
+ "learning_rate": 7.362857142857144e-06,
358
+ "loss": 0.0,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1450.0,
363
+ "learning_rate": 7.291428571428571e-06,
364
+ "loss": 0.0,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1475.0,
369
+ "learning_rate": 7.22e-06,
370
+ "loss": 0.0,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1500.0,
375
+ "learning_rate": 7.148571428571429e-06,
376
+ "loss": 0.0,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1525.0,
381
+ "learning_rate": 7.077142857142858e-06,
382
+ "loss": 0.0,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1550.0,
387
+ "learning_rate": 7.0057142857142865e-06,
388
+ "loss": 0.0,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1575.0,
393
+ "learning_rate": 6.934285714285715e-06,
394
+ "loss": 0.0,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1600.0,
399
+ "learning_rate": 6.862857142857144e-06,
400
+ "loss": 0.0,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1625.0,
405
+ "learning_rate": 6.791428571428572e-06,
406
+ "loss": 0.0,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1650.0,
411
+ "learning_rate": 6.720000000000001e-06,
412
+ "loss": 0.0,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1675.0,
417
+ "learning_rate": 6.648571428571429e-06,
418
+ "loss": 0.0,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1700.0,
423
+ "learning_rate": 6.577142857142857e-06,
424
+ "loss": 0.0,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1725.0,
429
+ "learning_rate": 6.505714285714286e-06,
430
+ "loss": 0.0,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1750.0,
435
+ "learning_rate": 6.434285714285715e-06,
436
+ "loss": 0.0,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1775.0,
441
+ "learning_rate": 6.3628571428571426e-06,
442
+ "loss": 0.0,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1800.0,
447
+ "learning_rate": 6.2914285714285716e-06,
448
+ "loss": 0.0,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1825.0,
453
+ "learning_rate": 6.220000000000001e-06,
454
+ "loss": 0.0,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1850.0,
459
+ "learning_rate": 6.14857142857143e-06,
460
+ "loss": 0.0,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1875.0,
465
+ "learning_rate": 6.077142857142858e-06,
466
+ "loss": 0.0,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1900.0,
471
+ "learning_rate": 6.005714285714286e-06,
472
+ "loss": 0.0,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1925.0,
477
+ "learning_rate": 5.934285714285715e-06,
478
+ "loss": 0.0,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1950.0,
483
+ "learning_rate": 5.862857142857143e-06,
484
+ "loss": 0.0,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1975.0,
489
+ "learning_rate": 5.791428571428572e-06,
490
+ "loss": 0.0,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 2000.0,
495
+ "learning_rate": 5.72e-06,
496
+ "loss": 0.0,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 2000.0,
501
+ "eval_loss": 2.00115704536438,
502
+ "eval_runtime": 3.6044,
503
+ "eval_samples_per_second": 0.555,
504
+ "eval_steps_per_second": 0.277,
505
+ "eval_wer": 84.61538461538461,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 2025.0,
510
+ "learning_rate": 5.6485714285714285e-06,
511
+ "loss": 0.0,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 2050.0,
516
+ "learning_rate": 5.5771428571428575e-06,
517
+ "loss": 0.0,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 2075.0,
522
+ "learning_rate": 5.5057142857142865e-06,
523
+ "loss": 0.0,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 2100.0,
528
+ "learning_rate": 5.4342857142857155e-06,
529
+ "loss": 0.0,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 2125.0,
534
+ "learning_rate": 5.362857142857143e-06,
535
+ "loss": 0.0,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 2150.0,
540
+ "learning_rate": 5.291428571428572e-06,
541
+ "loss": 0.0,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 2175.0,
546
+ "learning_rate": 5.220000000000001e-06,
547
+ "loss": 0.0,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 2200.0,
552
+ "learning_rate": 5.14857142857143e-06,
553
+ "loss": 0.0,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 2225.0,
558
+ "learning_rate": 5.077142857142857e-06,
559
+ "loss": 0.0,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 2250.0,
564
+ "learning_rate": 5.005714285714286e-06,
565
+ "loss": 0.0,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 2275.0,
570
+ "learning_rate": 4.934285714285715e-06,
571
+ "loss": 0.0,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 2300.0,
576
+ "learning_rate": 4.862857142857143e-06,
577
+ "loss": 0.0,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 2325.0,
582
+ "learning_rate": 4.7914285714285715e-06,
583
+ "loss": 0.0,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 2350.0,
588
+ "learning_rate": 4.7200000000000005e-06,
589
+ "loss": 0.0,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 2375.0,
594
+ "learning_rate": 4.648571428571429e-06,
595
+ "loss": 0.0,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 2400.0,
600
+ "learning_rate": 4.577142857142858e-06,
601
+ "loss": 0.0,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 2425.0,
606
+ "learning_rate": 4.505714285714286e-06,
607
+ "loss": 0.0,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 2450.0,
612
+ "learning_rate": 4.434285714285715e-06,
613
+ "loss": 0.0,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 2475.0,
618
+ "learning_rate": 4.362857142857143e-06,
619
+ "loss": 0.0,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 2500.0,
624
+ "learning_rate": 4.291428571428572e-06,
625
+ "loss": 0.0,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 2525.0,
630
+ "learning_rate": 4.22e-06,
631
+ "loss": 0.0,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 2550.0,
636
+ "learning_rate": 4.148571428571429e-06,
637
+ "loss": 0.0,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 2575.0,
642
+ "learning_rate": 4.0771428571428574e-06,
643
+ "loss": 0.0,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 2600.0,
648
+ "learning_rate": 4.0057142857142864e-06,
649
+ "loss": 0.0,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 2625.0,
654
+ "learning_rate": 3.934285714285715e-06,
655
+ "loss": 0.0,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 2650.0,
660
+ "learning_rate": 3.862857142857143e-06,
661
+ "loss": 0.0,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 2675.0,
666
+ "learning_rate": 3.7914285714285722e-06,
667
+ "loss": 0.0,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 2700.0,
672
+ "learning_rate": 3.7200000000000004e-06,
673
+ "loss": 0.0,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 2725.0,
678
+ "learning_rate": 3.648571428571429e-06,
679
+ "loss": 0.0,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 2750.0,
684
+ "learning_rate": 3.5771428571428576e-06,
685
+ "loss": 0.0,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 2775.0,
690
+ "learning_rate": 3.505714285714286e-06,
691
+ "loss": 0.0,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 2800.0,
696
+ "learning_rate": 3.4342857142857143e-06,
697
+ "loss": 0.0,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 2825.0,
702
+ "learning_rate": 3.3628571428571433e-06,
703
+ "loss": 0.0,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 2850.0,
708
+ "learning_rate": 3.2914285714285715e-06,
709
+ "loss": 0.0,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 2875.0,
714
+ "learning_rate": 3.2200000000000005e-06,
715
+ "loss": 0.0,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 2900.0,
720
+ "learning_rate": 3.1485714285714287e-06,
721
+ "loss": 0.0,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 2925.0,
726
+ "learning_rate": 3.0771428571428573e-06,
727
+ "loss": 0.0,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 2950.0,
732
+ "learning_rate": 3.005714285714286e-06,
733
+ "loss": 0.0,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 2975.0,
738
+ "learning_rate": 2.9342857142857144e-06,
739
+ "loss": 0.0,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 3000.0,
744
+ "learning_rate": 2.8628571428571435e-06,
745
+ "loss": 0.0,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 3000.0,
750
+ "eval_loss": 2.211437225341797,
751
+ "eval_runtime": 1.8782,
752
+ "eval_samples_per_second": 1.065,
753
+ "eval_steps_per_second": 0.532,
754
+ "eval_wer": 100.0,
755
+ "step": 3000
756
+ }
757
+ ],
758
+ "max_steps": 4000,
759
+ "num_train_epochs": 4000,
760
+ "total_flos": 8.6575620096e+18,
761
+ "trial_name": null,
762
+ "trial_params": null
763
+ }
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6b5d0e7f2aaac3a58da2d607fd67d4ade51924fd9bf60c5e1dea221d52e1338
3
+ size 3771
checkpoint-4000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.0,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.28.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-4000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.28.0.dev0"
221
+ }
checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cec62f80c33064c10676a95a3e1c353ee33c09bfc1836577d5e3dd9e58c6085
3
+ size 1934161093
checkpoint-4000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-4000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8364cb68ffaa0a78c457e9e1b3c786b98d99ee3984211e51b443d0c94b3815c0
3
+ size 967102601
checkpoint-4000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a12e9d06c2d853e014c7fd769a63ffb1151a2835c75cbbf78d2b5ecbb0f84b34
3
+ size 14575
checkpoint-4000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20b82c21fe7b9338877e01af34096cd81535fab5507b09dea08b203c45a08194
3
+ size 557
checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3be3e16e7a5628d50352078d8c3f4228325e660ea5cf0f4ec5b03b9a13573f4c
3
+ size 627
checkpoint-4000/trainer_state.json ADDED
@@ -0,0 +1,1012 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 84.61538461538461,
3
+ "best_model_checkpoint": "./whisper-small-hi/checkpoint-2000",
4
+ "epoch": 4000.0,
5
+ "global_step": 4000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 25.0,
12
+ "learning_rate": 4.6000000000000004e-07,
13
+ "loss": 2.7845,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 50.0,
18
+ "learning_rate": 9.600000000000001e-07,
19
+ "loss": 1.8784,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 75.0,
24
+ "learning_rate": 1.46e-06,
25
+ "loss": 0.7949,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 100.0,
30
+ "learning_rate": 1.9600000000000003e-06,
31
+ "loss": 0.2619,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 125.0,
36
+ "learning_rate": 2.46e-06,
37
+ "loss": 0.1719,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 150.0,
42
+ "learning_rate": 2.96e-06,
43
+ "loss": 0.1447,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 175.0,
48
+ "learning_rate": 3.46e-06,
49
+ "loss": 0.1234,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 200.0,
54
+ "learning_rate": 3.96e-06,
55
+ "loss": 0.0997,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 225.0,
60
+ "learning_rate": 4.4600000000000005e-06,
61
+ "loss": 0.0658,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 250.0,
66
+ "learning_rate": 4.960000000000001e-06,
67
+ "loss": 0.0054,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 275.0,
72
+ "learning_rate": 5.460000000000001e-06,
73
+ "loss": 0.0006,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 300.0,
78
+ "learning_rate": 5.9600000000000005e-06,
79
+ "loss": 0.0003,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 325.0,
84
+ "learning_rate": 6.460000000000001e-06,
85
+ "loss": 0.0002,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 350.0,
90
+ "learning_rate": 6.96e-06,
91
+ "loss": 0.0002,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 375.0,
96
+ "learning_rate": 7.4600000000000006e-06,
97
+ "loss": 0.0002,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 400.0,
102
+ "learning_rate": 7.960000000000002e-06,
103
+ "loss": 0.0001,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 425.0,
108
+ "learning_rate": 8.46e-06,
109
+ "loss": 0.0001,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 450.0,
114
+ "learning_rate": 8.96e-06,
115
+ "loss": 0.0001,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 475.0,
120
+ "learning_rate": 9.460000000000001e-06,
121
+ "loss": 0.0001,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 500.0,
126
+ "learning_rate": 9.960000000000001e-06,
127
+ "loss": 0.0001,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 525.0,
132
+ "learning_rate": 9.934285714285715e-06,
133
+ "loss": 0.0001,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 550.0,
138
+ "learning_rate": 9.862857142857144e-06,
139
+ "loss": 0.0001,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 575.0,
144
+ "learning_rate": 9.791428571428571e-06,
145
+ "loss": 0.0001,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 600.0,
150
+ "learning_rate": 9.72e-06,
151
+ "loss": 0.0001,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 625.0,
156
+ "learning_rate": 9.648571428571429e-06,
157
+ "loss": 0.0001,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 650.0,
162
+ "learning_rate": 9.577142857142858e-06,
163
+ "loss": 0.0001,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 675.0,
168
+ "learning_rate": 9.505714285714287e-06,
169
+ "loss": 0.0,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 700.0,
174
+ "learning_rate": 9.434285714285714e-06,
175
+ "loss": 0.0,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 725.0,
180
+ "learning_rate": 9.362857142857143e-06,
181
+ "loss": 0.0,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 750.0,
186
+ "learning_rate": 9.291428571428572e-06,
187
+ "loss": 0.0,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 775.0,
192
+ "learning_rate": 9.220000000000002e-06,
193
+ "loss": 0.0,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 800.0,
198
+ "learning_rate": 9.148571428571429e-06,
199
+ "loss": 0.0,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 825.0,
204
+ "learning_rate": 9.077142857142858e-06,
205
+ "loss": 0.0,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 850.0,
210
+ "learning_rate": 9.005714285714287e-06,
211
+ "loss": 0.0,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 875.0,
216
+ "learning_rate": 8.934285714285716e-06,
217
+ "loss": 0.0,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 900.0,
222
+ "learning_rate": 8.862857142857143e-06,
223
+ "loss": 0.0,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 925.0,
228
+ "learning_rate": 8.791428571428572e-06,
229
+ "loss": 0.0,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 950.0,
234
+ "learning_rate": 8.720000000000001e-06,
235
+ "loss": 0.0,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 975.0,
240
+ "learning_rate": 8.64857142857143e-06,
241
+ "loss": 0.0,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 1000.0,
246
+ "learning_rate": 8.577142857142858e-06,
247
+ "loss": 0.0,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 1000.0,
252
+ "eval_loss": 1.5850915908813477,
253
+ "eval_runtime": 3.2322,
254
+ "eval_samples_per_second": 0.619,
255
+ "eval_steps_per_second": 0.309,
256
+ "eval_wer": 92.3076923076923,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 1025.0,
261
+ "learning_rate": 8.505714285714287e-06,
262
+ "loss": 0.0,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1050.0,
267
+ "learning_rate": 8.434285714285716e-06,
268
+ "loss": 0.0,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1075.0,
273
+ "learning_rate": 8.362857142857143e-06,
274
+ "loss": 0.0,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1100.0,
279
+ "learning_rate": 8.291428571428572e-06,
280
+ "loss": 0.0,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1125.0,
285
+ "learning_rate": 8.220000000000001e-06,
286
+ "loss": 0.0,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1150.0,
291
+ "learning_rate": 8.148571428571428e-06,
292
+ "loss": 0.0,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1175.0,
297
+ "learning_rate": 8.077142857142857e-06,
298
+ "loss": 0.0,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1200.0,
303
+ "learning_rate": 8.005714285714286e-06,
304
+ "loss": 0.0,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1225.0,
309
+ "learning_rate": 7.934285714285715e-06,
310
+ "loss": 0.0,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1250.0,
315
+ "learning_rate": 7.862857142857143e-06,
316
+ "loss": 0.0,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1275.0,
321
+ "learning_rate": 7.791428571428572e-06,
322
+ "loss": 0.0,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1300.0,
327
+ "learning_rate": 7.72e-06,
328
+ "loss": 0.0,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1325.0,
333
+ "learning_rate": 7.64857142857143e-06,
334
+ "loss": 0.0,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1350.0,
339
+ "learning_rate": 7.577142857142857e-06,
340
+ "loss": 0.0,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1375.0,
345
+ "learning_rate": 7.505714285714286e-06,
346
+ "loss": 0.0,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1400.0,
351
+ "learning_rate": 7.434285714285715e-06,
352
+ "loss": 0.0,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1425.0,
357
+ "learning_rate": 7.362857142857144e-06,
358
+ "loss": 0.0,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1450.0,
363
+ "learning_rate": 7.291428571428571e-06,
364
+ "loss": 0.0,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1475.0,
369
+ "learning_rate": 7.22e-06,
370
+ "loss": 0.0,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1500.0,
375
+ "learning_rate": 7.148571428571429e-06,
376
+ "loss": 0.0,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1525.0,
381
+ "learning_rate": 7.077142857142858e-06,
382
+ "loss": 0.0,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1550.0,
387
+ "learning_rate": 7.0057142857142865e-06,
388
+ "loss": 0.0,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1575.0,
393
+ "learning_rate": 6.934285714285715e-06,
394
+ "loss": 0.0,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1600.0,
399
+ "learning_rate": 6.862857142857144e-06,
400
+ "loss": 0.0,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1625.0,
405
+ "learning_rate": 6.791428571428572e-06,
406
+ "loss": 0.0,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1650.0,
411
+ "learning_rate": 6.720000000000001e-06,
412
+ "loss": 0.0,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1675.0,
417
+ "learning_rate": 6.648571428571429e-06,
418
+ "loss": 0.0,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1700.0,
423
+ "learning_rate": 6.577142857142857e-06,
424
+ "loss": 0.0,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1725.0,
429
+ "learning_rate": 6.505714285714286e-06,
430
+ "loss": 0.0,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1750.0,
435
+ "learning_rate": 6.434285714285715e-06,
436
+ "loss": 0.0,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1775.0,
441
+ "learning_rate": 6.3628571428571426e-06,
442
+ "loss": 0.0,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1800.0,
447
+ "learning_rate": 6.2914285714285716e-06,
448
+ "loss": 0.0,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1825.0,
453
+ "learning_rate": 6.220000000000001e-06,
454
+ "loss": 0.0,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1850.0,
459
+ "learning_rate": 6.14857142857143e-06,
460
+ "loss": 0.0,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1875.0,
465
+ "learning_rate": 6.077142857142858e-06,
466
+ "loss": 0.0,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1900.0,
471
+ "learning_rate": 6.005714285714286e-06,
472
+ "loss": 0.0,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1925.0,
477
+ "learning_rate": 5.934285714285715e-06,
478
+ "loss": 0.0,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1950.0,
483
+ "learning_rate": 5.862857142857143e-06,
484
+ "loss": 0.0,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1975.0,
489
+ "learning_rate": 5.791428571428572e-06,
490
+ "loss": 0.0,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 2000.0,
495
+ "learning_rate": 5.72e-06,
496
+ "loss": 0.0,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 2000.0,
501
+ "eval_loss": 2.00115704536438,
502
+ "eval_runtime": 3.6044,
503
+ "eval_samples_per_second": 0.555,
504
+ "eval_steps_per_second": 0.277,
505
+ "eval_wer": 84.61538461538461,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 2025.0,
510
+ "learning_rate": 5.6485714285714285e-06,
511
+ "loss": 0.0,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 2050.0,
516
+ "learning_rate": 5.5771428571428575e-06,
517
+ "loss": 0.0,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 2075.0,
522
+ "learning_rate": 5.5057142857142865e-06,
523
+ "loss": 0.0,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 2100.0,
528
+ "learning_rate": 5.4342857142857155e-06,
529
+ "loss": 0.0,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 2125.0,
534
+ "learning_rate": 5.362857142857143e-06,
535
+ "loss": 0.0,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 2150.0,
540
+ "learning_rate": 5.291428571428572e-06,
541
+ "loss": 0.0,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 2175.0,
546
+ "learning_rate": 5.220000000000001e-06,
547
+ "loss": 0.0,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 2200.0,
552
+ "learning_rate": 5.14857142857143e-06,
553
+ "loss": 0.0,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 2225.0,
558
+ "learning_rate": 5.077142857142857e-06,
559
+ "loss": 0.0,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 2250.0,
564
+ "learning_rate": 5.005714285714286e-06,
565
+ "loss": 0.0,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 2275.0,
570
+ "learning_rate": 4.934285714285715e-06,
571
+ "loss": 0.0,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 2300.0,
576
+ "learning_rate": 4.862857142857143e-06,
577
+ "loss": 0.0,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 2325.0,
582
+ "learning_rate": 4.7914285714285715e-06,
583
+ "loss": 0.0,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 2350.0,
588
+ "learning_rate": 4.7200000000000005e-06,
589
+ "loss": 0.0,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 2375.0,
594
+ "learning_rate": 4.648571428571429e-06,
595
+ "loss": 0.0,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 2400.0,
600
+ "learning_rate": 4.577142857142858e-06,
601
+ "loss": 0.0,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 2425.0,
606
+ "learning_rate": 4.505714285714286e-06,
607
+ "loss": 0.0,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 2450.0,
612
+ "learning_rate": 4.434285714285715e-06,
613
+ "loss": 0.0,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 2475.0,
618
+ "learning_rate": 4.362857142857143e-06,
619
+ "loss": 0.0,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 2500.0,
624
+ "learning_rate": 4.291428571428572e-06,
625
+ "loss": 0.0,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 2525.0,
630
+ "learning_rate": 4.22e-06,
631
+ "loss": 0.0,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 2550.0,
636
+ "learning_rate": 4.148571428571429e-06,
637
+ "loss": 0.0,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 2575.0,
642
+ "learning_rate": 4.0771428571428574e-06,
643
+ "loss": 0.0,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 2600.0,
648
+ "learning_rate": 4.0057142857142864e-06,
649
+ "loss": 0.0,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 2625.0,
654
+ "learning_rate": 3.934285714285715e-06,
655
+ "loss": 0.0,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 2650.0,
660
+ "learning_rate": 3.862857142857143e-06,
661
+ "loss": 0.0,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 2675.0,
666
+ "learning_rate": 3.7914285714285722e-06,
667
+ "loss": 0.0,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 2700.0,
672
+ "learning_rate": 3.7200000000000004e-06,
673
+ "loss": 0.0,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 2725.0,
678
+ "learning_rate": 3.648571428571429e-06,
679
+ "loss": 0.0,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 2750.0,
684
+ "learning_rate": 3.5771428571428576e-06,
685
+ "loss": 0.0,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 2775.0,
690
+ "learning_rate": 3.505714285714286e-06,
691
+ "loss": 0.0,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 2800.0,
696
+ "learning_rate": 3.4342857142857143e-06,
697
+ "loss": 0.0,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 2825.0,
702
+ "learning_rate": 3.3628571428571433e-06,
703
+ "loss": 0.0,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 2850.0,
708
+ "learning_rate": 3.2914285714285715e-06,
709
+ "loss": 0.0,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 2875.0,
714
+ "learning_rate": 3.2200000000000005e-06,
715
+ "loss": 0.0,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 2900.0,
720
+ "learning_rate": 3.1485714285714287e-06,
721
+ "loss": 0.0,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 2925.0,
726
+ "learning_rate": 3.0771428571428573e-06,
727
+ "loss": 0.0,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 2950.0,
732
+ "learning_rate": 3.005714285714286e-06,
733
+ "loss": 0.0,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 2975.0,
738
+ "learning_rate": 2.9342857142857144e-06,
739
+ "loss": 0.0,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 3000.0,
744
+ "learning_rate": 2.8628571428571435e-06,
745
+ "loss": 0.0,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 3000.0,
750
+ "eval_loss": 2.211437225341797,
751
+ "eval_runtime": 1.8782,
752
+ "eval_samples_per_second": 1.065,
753
+ "eval_steps_per_second": 0.532,
754
+ "eval_wer": 100.0,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 3025.0,
759
+ "learning_rate": 2.7914285714285716e-06,
760
+ "loss": 0.0,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 3050.0,
765
+ "learning_rate": 2.7200000000000002e-06,
766
+ "loss": 0.0,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 3075.0,
771
+ "learning_rate": 2.648571428571429e-06,
772
+ "loss": 0.0,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 3100.0,
777
+ "learning_rate": 2.5771428571428574e-06,
778
+ "loss": 0.0,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 3125.0,
783
+ "learning_rate": 2.5057142857142856e-06,
784
+ "loss": 0.0,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 3150.0,
789
+ "learning_rate": 2.4342857142857146e-06,
790
+ "loss": 0.0,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 3175.0,
795
+ "learning_rate": 2.362857142857143e-06,
796
+ "loss": 0.0,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 3200.0,
801
+ "learning_rate": 2.2914285714285718e-06,
802
+ "loss": 0.0,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 3225.0,
807
+ "learning_rate": 2.2200000000000003e-06,
808
+ "loss": 0.0,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 3250.0,
813
+ "learning_rate": 2.148571428571429e-06,
814
+ "loss": 0.0,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 3275.0,
819
+ "learning_rate": 2.077142857142857e-06,
820
+ "loss": 0.0,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 3300.0,
825
+ "learning_rate": 2.0057142857142857e-06,
826
+ "loss": 0.0,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 3325.0,
831
+ "learning_rate": 1.9342857142857143e-06,
832
+ "loss": 0.0,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 3350.0,
837
+ "learning_rate": 1.8628571428571429e-06,
838
+ "loss": 0.0,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 3375.0,
843
+ "learning_rate": 1.7914285714285715e-06,
844
+ "loss": 0.0,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 3400.0,
849
+ "learning_rate": 1.72e-06,
850
+ "loss": 0.0,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 3425.0,
855
+ "learning_rate": 1.6485714285714289e-06,
856
+ "loss": 0.0,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 3450.0,
861
+ "learning_rate": 1.5771428571428574e-06,
862
+ "loss": 0.0,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 3475.0,
867
+ "learning_rate": 1.5057142857142858e-06,
868
+ "loss": 0.0,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 3500.0,
873
+ "learning_rate": 1.4342857142857144e-06,
874
+ "loss": 0.0,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 3525.0,
879
+ "learning_rate": 1.362857142857143e-06,
880
+ "loss": 0.0,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 3550.0,
885
+ "learning_rate": 1.2914285714285716e-06,
886
+ "loss": 0.0,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 3575.0,
891
+ "learning_rate": 1.2200000000000002e-06,
892
+ "loss": 0.0,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 3600.0,
897
+ "learning_rate": 1.1485714285714286e-06,
898
+ "loss": 0.0,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 3625.0,
903
+ "learning_rate": 1.0771428571428574e-06,
904
+ "loss": 0.0,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 3650.0,
909
+ "learning_rate": 1.0057142857142857e-06,
910
+ "loss": 0.0,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 3675.0,
915
+ "learning_rate": 9.342857142857144e-07,
916
+ "loss": 0.0,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 3700.0,
921
+ "learning_rate": 8.628571428571429e-07,
922
+ "loss": 0.0,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 3725.0,
927
+ "learning_rate": 7.914285714285715e-07,
928
+ "loss": 0.0,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 3750.0,
933
+ "learning_rate": 7.2e-07,
934
+ "loss": 0.0,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 3775.0,
939
+ "learning_rate": 6.485714285714287e-07,
940
+ "loss": 0.0,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 3800.0,
945
+ "learning_rate": 5.771428571428572e-07,
946
+ "loss": 0.0,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 3825.0,
951
+ "learning_rate": 5.057142857142858e-07,
952
+ "loss": 0.0,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 3850.0,
957
+ "learning_rate": 4.342857142857143e-07,
958
+ "loss": 0.0,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 3875.0,
963
+ "learning_rate": 3.6285714285714283e-07,
964
+ "loss": 0.0,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 3900.0,
969
+ "learning_rate": 2.914285714285715e-07,
970
+ "loss": 0.0,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 3925.0,
975
+ "learning_rate": 2.2e-07,
976
+ "loss": 0.0,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 3950.0,
981
+ "learning_rate": 1.4857142857142857e-07,
982
+ "loss": 0.0,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 3975.0,
987
+ "learning_rate": 7.714285714285715e-08,
988
+ "loss": 0.0,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 4000.0,
993
+ "learning_rate": 5.714285714285715e-09,
994
+ "loss": 0.0,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 4000.0,
999
+ "eval_loss": 2.297966957092285,
1000
+ "eval_runtime": 5.0248,
1001
+ "eval_samples_per_second": 0.398,
1002
+ "eval_steps_per_second": 0.199,
1003
+ "eval_wer": 238.46153846153845,
1004
+ "step": 4000
1005
+ }
1006
+ ],
1007
+ "max_steps": 4000,
1008
+ "num_train_epochs": 4000,
1009
+ "total_flos": 1.15434160128e+19,
1010
+ "trial_name": null,
1011
+ "trial_params": null
1012
+ }
checkpoint-4000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6b5d0e7f2aaac3a58da2d607fd67d4ade51924fd9bf60c5e1dea221d52e1338
3
+ size 3771
runs/Apr07_19-09-15_6672b5533821/1680894671.4432056/events.out.tfevents.1680894671.6672b5533821.633.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71b25104615eda169eea274979744bf918f95c0f0f875d8e75312703cf484ac4
3
+ size 6169
runs/Apr07_19-09-15_6672b5533821/events.out.tfevents.1680894671.6672b5533821.633.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b66076237ebcd4a373c1a944a861b45f811d530ec8acebed0669109f3a01373
3
+ size 31540