xezpeleta commited on
Commit
2810135
1 Parent(s): f1570bc

Update the learning rate hyperparameter

Browse files
Files changed (47) hide show
  1. checkpoint-1000/config.json +0 -41
  2. checkpoint-1000/optimizer.pt +0 -3
  3. checkpoint-1000/preprocessor_config.json +0 -0
  4. checkpoint-1000/pytorch_model.bin +0 -3
  5. checkpoint-1000/rng_state.pth +0 -3
  6. checkpoint-1000/scaler.pt +0 -3
  7. checkpoint-1000/scheduler.pt +0 -3
  8. checkpoint-1000/trainer_state.json +0 -265
  9. checkpoint-1000/training_args.bin +0 -3
  10. checkpoint-2000/config.json +0 -41
  11. checkpoint-2000/optimizer.pt +0 -3
  12. checkpoint-2000/preprocessor_config.json +0 -0
  13. checkpoint-2000/pytorch_model.bin +0 -3
  14. checkpoint-2000/rng_state.pth +0 -3
  15. checkpoint-2000/scaler.pt +0 -3
  16. checkpoint-2000/scheduler.pt +0 -3
  17. checkpoint-2000/trainer_state.json +0 -514
  18. checkpoint-2000/training_args.bin +0 -3
  19. checkpoint-3000/config.json +0 -41
  20. checkpoint-3000/optimizer.pt +0 -3
  21. checkpoint-3000/preprocessor_config.json +0 -0
  22. checkpoint-3000/pytorch_model.bin +0 -3
  23. checkpoint-3000/rng_state.pth +0 -3
  24. checkpoint-3000/scaler.pt +0 -3
  25. checkpoint-3000/scheduler.pt +0 -3
  26. checkpoint-3000/trainer_state.json +0 -763
  27. checkpoint-3000/training_args.bin +0 -3
  28. checkpoint-4000/config.json +0 -41
  29. checkpoint-4000/optimizer.pt +0 -3
  30. checkpoint-4000/preprocessor_config.json +0 -0
  31. checkpoint-4000/pytorch_model.bin +0 -3
  32. checkpoint-4000/rng_state.pth +0 -3
  33. checkpoint-4000/scaler.pt +0 -3
  34. checkpoint-4000/scheduler.pt +0 -3
  35. checkpoint-4000/trainer_state.json +0 -1012
  36. checkpoint-4000/training_args.bin +0 -3
  37. checkpoint-5000/config.json +0 -41
  38. checkpoint-5000/optimizer.pt +0 -3
  39. checkpoint-5000/preprocessor_config.json +0 -0
  40. checkpoint-5000/pytorch_model.bin +0 -3
  41. checkpoint-5000/rng_state.pth +0 -3
  42. checkpoint-5000/scaler.pt +0 -3
  43. checkpoint-5000/scheduler.pt +0 -3
  44. checkpoint-5000/trainer_state.json +0 -1261
  45. checkpoint-5000/training_args.bin +0 -3
  46. nohup.out +0 -0
  47. run.sh +3 -3
checkpoint-1000/config.json DELETED
@@ -1,41 +0,0 @@
1
- {
2
- "_name_or_path": "openai/whisper-medium",
3
- "activation_dropout": 0.0,
4
- "activation_function": "gelu",
5
- "architectures": [
6
- "WhisperForConditionalGeneration"
7
- ],
8
- "attention_dropout": 0.0,
9
- "begin_suppress_tokens": [
10
- 220,
11
- 50257
12
- ],
13
- "bos_token_id": 50257,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 24,
19
- "decoder_start_token_id": 50258,
20
- "dropout": 0.0,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 24,
25
- "eos_token_id": 50257,
26
- "forced_decoder_ids": null,
27
- "init_std": 0.02,
28
- "is_encoder_decoder": true,
29
- "max_length": 448,
30
- "max_source_positions": 1500,
31
- "max_target_positions": 448,
32
- "model_type": "whisper",
33
- "num_hidden_layers": 24,
34
- "num_mel_bins": 80,
35
- "pad_token_id": 50257,
36
- "scale_embedding": false,
37
- "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
- "use_cache": false,
40
- "vocab_size": 51865
41
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-1000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e445adc1c3c18c87959585f1c753c47cddbfe8321f2096b28ccdc0105f64d216
3
- size 6111428695
 
 
 
 
checkpoint-1000/preprocessor_config.json DELETED
The diff for this file is too large to render. See raw diff
 
checkpoint-1000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:ae163c1a2a53962024eae4041f8b164051aa5f9d0be4db7fd4035ffc54fcc2f7
3
- size 3055754841
 
 
 
 
checkpoint-1000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:337db3bde69262529f84673e0c1688a8983313ba44ed7b47639b3c14c033c21e
3
- size 14575
 
 
 
 
checkpoint-1000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8a52443f13b1ebc1f37fc5e2f4410a0fd95d757755d28ce4ca3377896a36d719
3
- size 557
 
 
 
 
checkpoint-1000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b864fef9e9930a8de94a02300fc770f05cb6e9d2328db736b4330cf0cb6352fb
3
- size 627
 
 
 
 
checkpoint-1000/trainer_state.json DELETED
@@ -1,265 +0,0 @@
1
- {
2
- "best_metric": 28.265624367127064,
3
- "best_model_checkpoint": "./checkpoint-1000",
4
- "epoch": 0.14285714285714285,
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": 0.0,
12
- "learning_rate": 4.6000000000000004e-07,
13
- "loss": 1.4182,
14
- "step": 25
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.400000000000001e-07,
19
- "loss": 1.292,
20
- "step": 50
21
- },
22
- {
23
- "epoch": 0.01,
24
- "learning_rate": 1.44e-06,
25
- "loss": 1.0018,
26
- "step": 75
27
- },
28
- {
29
- "epoch": 0.01,
30
- "learning_rate": 1.94e-06,
31
- "loss": 0.7765,
32
- "step": 100
33
- },
34
- {
35
- "epoch": 0.02,
36
- "learning_rate": 2.4400000000000004e-06,
37
- "loss": 0.7103,
38
- "step": 125
39
- },
40
- {
41
- "epoch": 0.02,
42
- "learning_rate": 2.9400000000000002e-06,
43
- "loss": 0.6597,
44
- "step": 150
45
- },
46
- {
47
- "epoch": 0.03,
48
- "learning_rate": 3.44e-06,
49
- "loss": 0.6657,
50
- "step": 175
51
- },
52
- {
53
- "epoch": 0.03,
54
- "learning_rate": 3.94e-06,
55
- "loss": 0.5853,
56
- "step": 200
57
- },
58
- {
59
- "epoch": 0.03,
60
- "learning_rate": 4.440000000000001e-06,
61
- "loss": 0.5273,
62
- "step": 225
63
- },
64
- {
65
- "epoch": 0.04,
66
- "learning_rate": 4.94e-06,
67
- "loss": 0.5979,
68
- "step": 250
69
- },
70
- {
71
- "epoch": 0.04,
72
- "learning_rate": 5.4400000000000004e-06,
73
- "loss": 0.5861,
74
- "step": 275
75
- },
76
- {
77
- "epoch": 0.04,
78
- "learning_rate": 5.94e-06,
79
- "loss": 0.5085,
80
- "step": 300
81
- },
82
- {
83
- "epoch": 0.05,
84
- "learning_rate": 6.440000000000001e-06,
85
- "loss": 0.4827,
86
- "step": 325
87
- },
88
- {
89
- "epoch": 0.05,
90
- "learning_rate": 6.9400000000000005e-06,
91
- "loss": 0.4909,
92
- "step": 350
93
- },
94
- {
95
- "epoch": 0.05,
96
- "learning_rate": 7.440000000000001e-06,
97
- "loss": 0.4651,
98
- "step": 375
99
- },
100
- {
101
- "epoch": 0.06,
102
- "learning_rate": 7.94e-06,
103
- "loss": 0.494,
104
- "step": 400
105
- },
106
- {
107
- "epoch": 0.06,
108
- "learning_rate": 8.44e-06,
109
- "loss": 0.4188,
110
- "step": 425
111
- },
112
- {
113
- "epoch": 0.06,
114
- "learning_rate": 8.94e-06,
115
- "loss": 0.3849,
116
- "step": 450
117
- },
118
- {
119
- "epoch": 0.07,
120
- "learning_rate": 9.440000000000001e-06,
121
- "loss": 0.4577,
122
- "step": 475
123
- },
124
- {
125
- "epoch": 0.07,
126
- "learning_rate": 9.940000000000001e-06,
127
- "loss": 0.4415,
128
- "step": 500
129
- },
130
- {
131
- "epoch": 0.07,
132
- "learning_rate": 9.966153846153847e-06,
133
- "loss": 0.4614,
134
- "step": 525
135
- },
136
- {
137
- "epoch": 0.08,
138
- "learning_rate": 9.927692307692309e-06,
139
- "loss": 0.4283,
140
- "step": 550
141
- },
142
- {
143
- "epoch": 0.08,
144
- "learning_rate": 9.88923076923077e-06,
145
- "loss": 0.4486,
146
- "step": 575
147
- },
148
- {
149
- "epoch": 0.09,
150
- "learning_rate": 9.850769230769231e-06,
151
- "loss": 0.4434,
152
- "step": 600
153
- },
154
- {
155
- "epoch": 0.09,
156
- "learning_rate": 9.812307692307694e-06,
157
- "loss": 0.4245,
158
- "step": 625
159
- },
160
- {
161
- "epoch": 0.09,
162
- "learning_rate": 9.773846153846154e-06,
163
- "loss": 0.4254,
164
- "step": 650
165
- },
166
- {
167
- "epoch": 0.1,
168
- "learning_rate": 9.735384615384616e-06,
169
- "loss": 0.4143,
170
- "step": 675
171
- },
172
- {
173
- "epoch": 0.1,
174
- "learning_rate": 9.696923076923078e-06,
175
- "loss": 0.4115,
176
- "step": 700
177
- },
178
- {
179
- "epoch": 0.1,
180
- "learning_rate": 9.658461538461539e-06,
181
- "loss": 0.3656,
182
- "step": 725
183
- },
184
- {
185
- "epoch": 0.11,
186
- "learning_rate": 9.620000000000001e-06,
187
- "loss": 0.4004,
188
- "step": 750
189
- },
190
- {
191
- "epoch": 0.11,
192
- "learning_rate": 9.581538461538462e-06,
193
- "loss": 0.388,
194
- "step": 775
195
- },
196
- {
197
- "epoch": 0.11,
198
- "learning_rate": 9.543076923076924e-06,
199
- "loss": 0.3665,
200
- "step": 800
201
- },
202
- {
203
- "epoch": 0.12,
204
- "learning_rate": 9.504615384615386e-06,
205
- "loss": 0.3777,
206
- "step": 825
207
- },
208
- {
209
- "epoch": 0.12,
210
- "learning_rate": 9.466153846153846e-06,
211
- "loss": 0.4338,
212
- "step": 850
213
- },
214
- {
215
- "epoch": 0.12,
216
- "learning_rate": 9.427692307692309e-06,
217
- "loss": 0.4017,
218
- "step": 875
219
- },
220
- {
221
- "epoch": 0.13,
222
- "learning_rate": 9.38923076923077e-06,
223
- "loss": 0.3757,
224
- "step": 900
225
- },
226
- {
227
- "epoch": 0.13,
228
- "learning_rate": 9.350769230769231e-06,
229
- "loss": 0.4084,
230
- "step": 925
231
- },
232
- {
233
- "epoch": 0.14,
234
- "learning_rate": 9.312307692307693e-06,
235
- "loss": 0.3821,
236
- "step": 950
237
- },
238
- {
239
- "epoch": 0.14,
240
- "learning_rate": 9.273846153846154e-06,
241
- "loss": 0.421,
242
- "step": 975
243
- },
244
- {
245
- "epoch": 0.14,
246
- "learning_rate": 9.235384615384616e-06,
247
- "loss": 0.4203,
248
- "step": 1000
249
- },
250
- {
251
- "epoch": 0.14,
252
- "eval_loss": 0.4127572178840637,
253
- "eval_runtime": 1802.5775,
254
- "eval_samples_per_second": 3.656,
255
- "eval_steps_per_second": 0.457,
256
- "eval_wer": 28.265624367127064,
257
- "step": 1000
258
- }
259
- ],
260
- "max_steps": 7000,
261
- "num_train_epochs": 9223372036854775807,
262
- "total_flos": 4.08241963008e+18,
263
- "trial_name": null,
264
- "trial_params": null
265
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-1000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ca27d9d1bef820c87cfd52d5adb15af5250e1ac4555cd1703a088ad0dad1a4
3
- size 3643
 
 
 
 
checkpoint-2000/config.json DELETED
@@ -1,41 +0,0 @@
1
- {
2
- "_name_or_path": "openai/whisper-medium",
3
- "activation_dropout": 0.0,
4
- "activation_function": "gelu",
5
- "architectures": [
6
- "WhisperForConditionalGeneration"
7
- ],
8
- "attention_dropout": 0.0,
9
- "begin_suppress_tokens": [
10
- 220,
11
- 50257
12
- ],
13
- "bos_token_id": 50257,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 24,
19
- "decoder_start_token_id": 50258,
20
- "dropout": 0.0,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 24,
25
- "eos_token_id": 50257,
26
- "forced_decoder_ids": null,
27
- "init_std": 0.02,
28
- "is_encoder_decoder": true,
29
- "max_length": 448,
30
- "max_source_positions": 1500,
31
- "max_target_positions": 448,
32
- "model_type": "whisper",
33
- "num_hidden_layers": 24,
34
- "num_mel_bins": 80,
35
- "pad_token_id": 50257,
36
- "scale_embedding": false,
37
- "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
- "use_cache": false,
40
- "vocab_size": 51865
41
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-2000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:a159b02fa01e4405076f9d759e53e7edd6d412b19b53a0743c660e28339c1008
3
- size 6111428695
 
 
 
 
checkpoint-2000/preprocessor_config.json DELETED
The diff for this file is too large to render. See raw diff
 
checkpoint-2000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:0fb0822dc45eb083f947413e68d9023acdfbe2dc7d6aa96a85b5e714d22ce5be
3
- size 3055754841
 
 
 
 
checkpoint-2000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:a1cbefc4297bef083a245f24c738abc29e75fca1b725711059d7702eeb3e5557
3
- size 14575
 
 
 
 
checkpoint-2000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:be1e74deedb43759bd15ec14e78ac0bbfcb7fd82b5ef5c8669deb1c4ee4fd1f2
3
- size 557
 
 
 
 
checkpoint-2000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:1f112c28e35ae8caf36bbfc9dd4dacc3130f175782da609fb5266a28a082e3cf
3
- size 627
 
 
 
 
checkpoint-2000/trainer_state.json DELETED
@@ -1,514 +0,0 @@
1
- {
2
- "best_metric": 22.052330997610273,
3
- "best_model_checkpoint": "./checkpoint-2000",
4
- "epoch": 0.2857142857142857,
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": 0.0,
12
- "learning_rate": 4.6000000000000004e-07,
13
- "loss": 1.4182,
14
- "step": 25
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.400000000000001e-07,
19
- "loss": 1.292,
20
- "step": 50
21
- },
22
- {
23
- "epoch": 0.01,
24
- "learning_rate": 1.44e-06,
25
- "loss": 1.0018,
26
- "step": 75
27
- },
28
- {
29
- "epoch": 0.01,
30
- "learning_rate": 1.94e-06,
31
- "loss": 0.7765,
32
- "step": 100
33
- },
34
- {
35
- "epoch": 0.02,
36
- "learning_rate": 2.4400000000000004e-06,
37
- "loss": 0.7103,
38
- "step": 125
39
- },
40
- {
41
- "epoch": 0.02,
42
- "learning_rate": 2.9400000000000002e-06,
43
- "loss": 0.6597,
44
- "step": 150
45
- },
46
- {
47
- "epoch": 0.03,
48
- "learning_rate": 3.44e-06,
49
- "loss": 0.6657,
50
- "step": 175
51
- },
52
- {
53
- "epoch": 0.03,
54
- "learning_rate": 3.94e-06,
55
- "loss": 0.5853,
56
- "step": 200
57
- },
58
- {
59
- "epoch": 0.03,
60
- "learning_rate": 4.440000000000001e-06,
61
- "loss": 0.5273,
62
- "step": 225
63
- },
64
- {
65
- "epoch": 0.04,
66
- "learning_rate": 4.94e-06,
67
- "loss": 0.5979,
68
- "step": 250
69
- },
70
- {
71
- "epoch": 0.04,
72
- "learning_rate": 5.4400000000000004e-06,
73
- "loss": 0.5861,
74
- "step": 275
75
- },
76
- {
77
- "epoch": 0.04,
78
- "learning_rate": 5.94e-06,
79
- "loss": 0.5085,
80
- "step": 300
81
- },
82
- {
83
- "epoch": 0.05,
84
- "learning_rate": 6.440000000000001e-06,
85
- "loss": 0.4827,
86
- "step": 325
87
- },
88
- {
89
- "epoch": 0.05,
90
- "learning_rate": 6.9400000000000005e-06,
91
- "loss": 0.4909,
92
- "step": 350
93
- },
94
- {
95
- "epoch": 0.05,
96
- "learning_rate": 7.440000000000001e-06,
97
- "loss": 0.4651,
98
- "step": 375
99
- },
100
- {
101
- "epoch": 0.06,
102
- "learning_rate": 7.94e-06,
103
- "loss": 0.494,
104
- "step": 400
105
- },
106
- {
107
- "epoch": 0.06,
108
- "learning_rate": 8.44e-06,
109
- "loss": 0.4188,
110
- "step": 425
111
- },
112
- {
113
- "epoch": 0.06,
114
- "learning_rate": 8.94e-06,
115
- "loss": 0.3849,
116
- "step": 450
117
- },
118
- {
119
- "epoch": 0.07,
120
- "learning_rate": 9.440000000000001e-06,
121
- "loss": 0.4577,
122
- "step": 475
123
- },
124
- {
125
- "epoch": 0.07,
126
- "learning_rate": 9.940000000000001e-06,
127
- "loss": 0.4415,
128
- "step": 500
129
- },
130
- {
131
- "epoch": 0.07,
132
- "learning_rate": 9.966153846153847e-06,
133
- "loss": 0.4614,
134
- "step": 525
135
- },
136
- {
137
- "epoch": 0.08,
138
- "learning_rate": 9.927692307692309e-06,
139
- "loss": 0.4283,
140
- "step": 550
141
- },
142
- {
143
- "epoch": 0.08,
144
- "learning_rate": 9.88923076923077e-06,
145
- "loss": 0.4486,
146
- "step": 575
147
- },
148
- {
149
- "epoch": 0.09,
150
- "learning_rate": 9.850769230769231e-06,
151
- "loss": 0.4434,
152
- "step": 600
153
- },
154
- {
155
- "epoch": 0.09,
156
- "learning_rate": 9.812307692307694e-06,
157
- "loss": 0.4245,
158
- "step": 625
159
- },
160
- {
161
- "epoch": 0.09,
162
- "learning_rate": 9.773846153846154e-06,
163
- "loss": 0.4254,
164
- "step": 650
165
- },
166
- {
167
- "epoch": 0.1,
168
- "learning_rate": 9.735384615384616e-06,
169
- "loss": 0.4143,
170
- "step": 675
171
- },
172
- {
173
- "epoch": 0.1,
174
- "learning_rate": 9.696923076923078e-06,
175
- "loss": 0.4115,
176
- "step": 700
177
- },
178
- {
179
- "epoch": 0.1,
180
- "learning_rate": 9.658461538461539e-06,
181
- "loss": 0.3656,
182
- "step": 725
183
- },
184
- {
185
- "epoch": 0.11,
186
- "learning_rate": 9.620000000000001e-06,
187
- "loss": 0.4004,
188
- "step": 750
189
- },
190
- {
191
- "epoch": 0.11,
192
- "learning_rate": 9.581538461538462e-06,
193
- "loss": 0.388,
194
- "step": 775
195
- },
196
- {
197
- "epoch": 0.11,
198
- "learning_rate": 9.543076923076924e-06,
199
- "loss": 0.3665,
200
- "step": 800
201
- },
202
- {
203
- "epoch": 0.12,
204
- "learning_rate": 9.504615384615386e-06,
205
- "loss": 0.3777,
206
- "step": 825
207
- },
208
- {
209
- "epoch": 0.12,
210
- "learning_rate": 9.466153846153846e-06,
211
- "loss": 0.4338,
212
- "step": 850
213
- },
214
- {
215
- "epoch": 0.12,
216
- "learning_rate": 9.427692307692309e-06,
217
- "loss": 0.4017,
218
- "step": 875
219
- },
220
- {
221
- "epoch": 0.13,
222
- "learning_rate": 9.38923076923077e-06,
223
- "loss": 0.3757,
224
- "step": 900
225
- },
226
- {
227
- "epoch": 0.13,
228
- "learning_rate": 9.350769230769231e-06,
229
- "loss": 0.4084,
230
- "step": 925
231
- },
232
- {
233
- "epoch": 0.14,
234
- "learning_rate": 9.312307692307693e-06,
235
- "loss": 0.3821,
236
- "step": 950
237
- },
238
- {
239
- "epoch": 0.14,
240
- "learning_rate": 9.273846153846154e-06,
241
- "loss": 0.421,
242
- "step": 975
243
- },
244
- {
245
- "epoch": 0.14,
246
- "learning_rate": 9.235384615384616e-06,
247
- "loss": 0.4203,
248
- "step": 1000
249
- },
250
- {
251
- "epoch": 0.14,
252
- "eval_loss": 0.4127572178840637,
253
- "eval_runtime": 1802.5775,
254
- "eval_samples_per_second": 3.656,
255
- "eval_steps_per_second": 0.457,
256
- "eval_wer": 28.265624367127064,
257
- "step": 1000
258
- },
259
- {
260
- "epoch": 0.15,
261
- "learning_rate": 9.196923076923078e-06,
262
- "loss": 0.4094,
263
- "step": 1025
264
- },
265
- {
266
- "epoch": 0.15,
267
- "learning_rate": 9.158461538461539e-06,
268
- "loss": 0.3674,
269
- "step": 1050
270
- },
271
- {
272
- "epoch": 0.15,
273
- "learning_rate": 9.12e-06,
274
- "loss": 0.3613,
275
- "step": 1075
276
- },
277
- {
278
- "epoch": 0.16,
279
- "learning_rate": 9.081538461538463e-06,
280
- "loss": 0.3904,
281
- "step": 1100
282
- },
283
- {
284
- "epoch": 0.16,
285
- "learning_rate": 9.043076923076923e-06,
286
- "loss": 0.3865,
287
- "step": 1125
288
- },
289
- {
290
- "epoch": 0.16,
291
- "learning_rate": 9.004615384615386e-06,
292
- "loss": 0.3677,
293
- "step": 1150
294
- },
295
- {
296
- "epoch": 0.17,
297
- "learning_rate": 8.966153846153846e-06,
298
- "loss": 0.3295,
299
- "step": 1175
300
- },
301
- {
302
- "epoch": 0.17,
303
- "learning_rate": 8.927692307692308e-06,
304
- "loss": 0.3207,
305
- "step": 1200
306
- },
307
- {
308
- "epoch": 0.17,
309
- "learning_rate": 8.88923076923077e-06,
310
- "loss": 0.2817,
311
- "step": 1225
312
- },
313
- {
314
- "epoch": 0.18,
315
- "learning_rate": 8.850769230769231e-06,
316
- "loss": 0.3218,
317
- "step": 1250
318
- },
319
- {
320
- "epoch": 0.18,
321
- "learning_rate": 8.812307692307693e-06,
322
- "loss": 0.3502,
323
- "step": 1275
324
- },
325
- {
326
- "epoch": 0.19,
327
- "learning_rate": 8.773846153846155e-06,
328
- "loss": 0.3432,
329
- "step": 1300
330
- },
331
- {
332
- "epoch": 0.19,
333
- "learning_rate": 8.735384615384616e-06,
334
- "loss": 0.3244,
335
- "step": 1325
336
- },
337
- {
338
- "epoch": 0.19,
339
- "learning_rate": 8.696923076923078e-06,
340
- "loss": 0.2607,
341
- "step": 1350
342
- },
343
- {
344
- "epoch": 0.2,
345
- "learning_rate": 8.658461538461538e-06,
346
- "loss": 0.3469,
347
- "step": 1375
348
- },
349
- {
350
- "epoch": 0.2,
351
- "learning_rate": 8.62e-06,
352
- "loss": 0.2756,
353
- "step": 1400
354
- },
355
- {
356
- "epoch": 0.2,
357
- "learning_rate": 8.581538461538463e-06,
358
- "loss": 0.294,
359
- "step": 1425
360
- },
361
- {
362
- "epoch": 0.21,
363
- "learning_rate": 8.543076923076923e-06,
364
- "loss": 0.3915,
365
- "step": 1450
366
- },
367
- {
368
- "epoch": 0.21,
369
- "learning_rate": 8.504615384615385e-06,
370
- "loss": 0.3237,
371
- "step": 1475
372
- },
373
- {
374
- "epoch": 0.21,
375
- "learning_rate": 8.466153846153847e-06,
376
- "loss": 0.2723,
377
- "step": 1500
378
- },
379
- {
380
- "epoch": 0.22,
381
- "learning_rate": 8.427692307692308e-06,
382
- "loss": 0.2917,
383
- "step": 1525
384
- },
385
- {
386
- "epoch": 0.22,
387
- "learning_rate": 8.38923076923077e-06,
388
- "loss": 0.2921,
389
- "step": 1550
390
- },
391
- {
392
- "epoch": 0.23,
393
- "learning_rate": 8.35076923076923e-06,
394
- "loss": 0.3724,
395
- "step": 1575
396
- },
397
- {
398
- "epoch": 0.23,
399
- "learning_rate": 8.312307692307693e-06,
400
- "loss": 0.3364,
401
- "step": 1600
402
- },
403
- {
404
- "epoch": 0.23,
405
- "learning_rate": 8.273846153846155e-06,
406
- "loss": 0.3661,
407
- "step": 1625
408
- },
409
- {
410
- "epoch": 0.24,
411
- "learning_rate": 8.235384615384615e-06,
412
- "loss": 0.3307,
413
- "step": 1650
414
- },
415
- {
416
- "epoch": 0.24,
417
- "learning_rate": 8.196923076923078e-06,
418
- "loss": 0.3201,
419
- "step": 1675
420
- },
421
- {
422
- "epoch": 0.24,
423
- "learning_rate": 8.15846153846154e-06,
424
- "loss": 0.2973,
425
- "step": 1700
426
- },
427
- {
428
- "epoch": 0.25,
429
- "learning_rate": 8.120000000000002e-06,
430
- "loss": 0.3013,
431
- "step": 1725
432
- },
433
- {
434
- "epoch": 0.25,
435
- "learning_rate": 8.081538461538462e-06,
436
- "loss": 0.3349,
437
- "step": 1750
438
- },
439
- {
440
- "epoch": 0.25,
441
- "learning_rate": 8.043076923076923e-06,
442
- "loss": 0.3244,
443
- "step": 1775
444
- },
445
- {
446
- "epoch": 0.26,
447
- "learning_rate": 8.004615384615385e-06,
448
- "loss": 0.3331,
449
- "step": 1800
450
- },
451
- {
452
- "epoch": 0.26,
453
- "learning_rate": 7.966153846153847e-06,
454
- "loss": 0.2723,
455
- "step": 1825
456
- },
457
- {
458
- "epoch": 0.26,
459
- "learning_rate": 7.927692307692308e-06,
460
- "loss": 0.316,
461
- "step": 1850
462
- },
463
- {
464
- "epoch": 0.27,
465
- "learning_rate": 7.88923076923077e-06,
466
- "loss": 0.2648,
467
- "step": 1875
468
- },
469
- {
470
- "epoch": 0.27,
471
- "learning_rate": 7.850769230769232e-06,
472
- "loss": 0.2965,
473
- "step": 1900
474
- },
475
- {
476
- "epoch": 0.28,
477
- "learning_rate": 7.812307692307694e-06,
478
- "loss": 0.256,
479
- "step": 1925
480
- },
481
- {
482
- "epoch": 0.28,
483
- "learning_rate": 7.773846153846155e-06,
484
- "loss": 0.2156,
485
- "step": 1950
486
- },
487
- {
488
- "epoch": 0.28,
489
- "learning_rate": 7.735384615384615e-06,
490
- "loss": 0.2418,
491
- "step": 1975
492
- },
493
- {
494
- "epoch": 0.29,
495
- "learning_rate": 7.696923076923077e-06,
496
- "loss": 0.2693,
497
- "step": 2000
498
- },
499
- {
500
- "epoch": 0.29,
501
- "eval_loss": 0.3239877223968506,
502
- "eval_runtime": 1817.7748,
503
- "eval_samples_per_second": 3.626,
504
- "eval_steps_per_second": 0.453,
505
- "eval_wer": 22.052330997610273,
506
- "step": 2000
507
- }
508
- ],
509
- "max_steps": 7000,
510
- "num_train_epochs": 9223372036854775807,
511
- "total_flos": 8.16483926016e+18,
512
- "trial_name": null,
513
- "trial_params": null
514
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-2000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ca27d9d1bef820c87cfd52d5adb15af5250e1ac4555cd1703a088ad0dad1a4
3
- size 3643
 
 
 
 
checkpoint-3000/config.json DELETED
@@ -1,41 +0,0 @@
1
- {
2
- "_name_or_path": "openai/whisper-medium",
3
- "activation_dropout": 0.0,
4
- "activation_function": "gelu",
5
- "architectures": [
6
- "WhisperForConditionalGeneration"
7
- ],
8
- "attention_dropout": 0.0,
9
- "begin_suppress_tokens": [
10
- 220,
11
- 50257
12
- ],
13
- "bos_token_id": 50257,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 24,
19
- "decoder_start_token_id": 50258,
20
- "dropout": 0.0,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 24,
25
- "eos_token_id": 50257,
26
- "forced_decoder_ids": null,
27
- "init_std": 0.02,
28
- "is_encoder_decoder": true,
29
- "max_length": 448,
30
- "max_source_positions": 1500,
31
- "max_target_positions": 448,
32
- "model_type": "whisper",
33
- "num_hidden_layers": 24,
34
- "num_mel_bins": 80,
35
- "pad_token_id": 50257,
36
- "scale_embedding": false,
37
- "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
- "use_cache": false,
40
- "vocab_size": 51865
41
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-3000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:204df11caa5a4cc0c8bc1ea18995b04587f1d457c412b5cf195a5a895f83074c
3
- size 6111428695
 
 
 
 
checkpoint-3000/preprocessor_config.json DELETED
The diff for this file is too large to render. See raw diff
 
checkpoint-3000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:4e8631f8c8b760ece8b67999d400c1670f308714ab9ab1de0f37457e497096b3
3
- size 3055754841
 
 
 
 
checkpoint-3000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3be0ae51fe09627683807728545588affaec3370727df148b33392ba8055e58f
3
- size 14575
 
 
 
 
checkpoint-3000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:eed3cdc5e39a2cfa8de4e0045b0e1a83c149d3d1d12677c2a9824efb5d14e59c
3
- size 557
 
 
 
 
checkpoint-3000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:dd7319c75e8411b4e2ae5a8b7fcdd7bb253d178b5a5770d7926097398cb526f8
3
- size 627
 
 
 
 
checkpoint-3000/trainer_state.json DELETED
@@ -1,763 +0,0 @@
1
- {
2
- "best_metric": 18.14370772408765,
3
- "best_model_checkpoint": "./checkpoint-3000",
4
- "epoch": 0.42857142857142855,
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": 0.0,
12
- "learning_rate": 4.6000000000000004e-07,
13
- "loss": 1.4182,
14
- "step": 25
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.400000000000001e-07,
19
- "loss": 1.292,
20
- "step": 50
21
- },
22
- {
23
- "epoch": 0.01,
24
- "learning_rate": 1.44e-06,
25
- "loss": 1.0018,
26
- "step": 75
27
- },
28
- {
29
- "epoch": 0.01,
30
- "learning_rate": 1.94e-06,
31
- "loss": 0.7765,
32
- "step": 100
33
- },
34
- {
35
- "epoch": 0.02,
36
- "learning_rate": 2.4400000000000004e-06,
37
- "loss": 0.7103,
38
- "step": 125
39
- },
40
- {
41
- "epoch": 0.02,
42
- "learning_rate": 2.9400000000000002e-06,
43
- "loss": 0.6597,
44
- "step": 150
45
- },
46
- {
47
- "epoch": 0.03,
48
- "learning_rate": 3.44e-06,
49
- "loss": 0.6657,
50
- "step": 175
51
- },
52
- {
53
- "epoch": 0.03,
54
- "learning_rate": 3.94e-06,
55
- "loss": 0.5853,
56
- "step": 200
57
- },
58
- {
59
- "epoch": 0.03,
60
- "learning_rate": 4.440000000000001e-06,
61
- "loss": 0.5273,
62
- "step": 225
63
- },
64
- {
65
- "epoch": 0.04,
66
- "learning_rate": 4.94e-06,
67
- "loss": 0.5979,
68
- "step": 250
69
- },
70
- {
71
- "epoch": 0.04,
72
- "learning_rate": 5.4400000000000004e-06,
73
- "loss": 0.5861,
74
- "step": 275
75
- },
76
- {
77
- "epoch": 0.04,
78
- "learning_rate": 5.94e-06,
79
- "loss": 0.5085,
80
- "step": 300
81
- },
82
- {
83
- "epoch": 0.05,
84
- "learning_rate": 6.440000000000001e-06,
85
- "loss": 0.4827,
86
- "step": 325
87
- },
88
- {
89
- "epoch": 0.05,
90
- "learning_rate": 6.9400000000000005e-06,
91
- "loss": 0.4909,
92
- "step": 350
93
- },
94
- {
95
- "epoch": 0.05,
96
- "learning_rate": 7.440000000000001e-06,
97
- "loss": 0.4651,
98
- "step": 375
99
- },
100
- {
101
- "epoch": 0.06,
102
- "learning_rate": 7.94e-06,
103
- "loss": 0.494,
104
- "step": 400
105
- },
106
- {
107
- "epoch": 0.06,
108
- "learning_rate": 8.44e-06,
109
- "loss": 0.4188,
110
- "step": 425
111
- },
112
- {
113
- "epoch": 0.06,
114
- "learning_rate": 8.94e-06,
115
- "loss": 0.3849,
116
- "step": 450
117
- },
118
- {
119
- "epoch": 0.07,
120
- "learning_rate": 9.440000000000001e-06,
121
- "loss": 0.4577,
122
- "step": 475
123
- },
124
- {
125
- "epoch": 0.07,
126
- "learning_rate": 9.940000000000001e-06,
127
- "loss": 0.4415,
128
- "step": 500
129
- },
130
- {
131
- "epoch": 0.07,
132
- "learning_rate": 9.966153846153847e-06,
133
- "loss": 0.4614,
134
- "step": 525
135
- },
136
- {
137
- "epoch": 0.08,
138
- "learning_rate": 9.927692307692309e-06,
139
- "loss": 0.4283,
140
- "step": 550
141
- },
142
- {
143
- "epoch": 0.08,
144
- "learning_rate": 9.88923076923077e-06,
145
- "loss": 0.4486,
146
- "step": 575
147
- },
148
- {
149
- "epoch": 0.09,
150
- "learning_rate": 9.850769230769231e-06,
151
- "loss": 0.4434,
152
- "step": 600
153
- },
154
- {
155
- "epoch": 0.09,
156
- "learning_rate": 9.812307692307694e-06,
157
- "loss": 0.4245,
158
- "step": 625
159
- },
160
- {
161
- "epoch": 0.09,
162
- "learning_rate": 9.773846153846154e-06,
163
- "loss": 0.4254,
164
- "step": 650
165
- },
166
- {
167
- "epoch": 0.1,
168
- "learning_rate": 9.735384615384616e-06,
169
- "loss": 0.4143,
170
- "step": 675
171
- },
172
- {
173
- "epoch": 0.1,
174
- "learning_rate": 9.696923076923078e-06,
175
- "loss": 0.4115,
176
- "step": 700
177
- },
178
- {
179
- "epoch": 0.1,
180
- "learning_rate": 9.658461538461539e-06,
181
- "loss": 0.3656,
182
- "step": 725
183
- },
184
- {
185
- "epoch": 0.11,
186
- "learning_rate": 9.620000000000001e-06,
187
- "loss": 0.4004,
188
- "step": 750
189
- },
190
- {
191
- "epoch": 0.11,
192
- "learning_rate": 9.581538461538462e-06,
193
- "loss": 0.388,
194
- "step": 775
195
- },
196
- {
197
- "epoch": 0.11,
198
- "learning_rate": 9.543076923076924e-06,
199
- "loss": 0.3665,
200
- "step": 800
201
- },
202
- {
203
- "epoch": 0.12,
204
- "learning_rate": 9.504615384615386e-06,
205
- "loss": 0.3777,
206
- "step": 825
207
- },
208
- {
209
- "epoch": 0.12,
210
- "learning_rate": 9.466153846153846e-06,
211
- "loss": 0.4338,
212
- "step": 850
213
- },
214
- {
215
- "epoch": 0.12,
216
- "learning_rate": 9.427692307692309e-06,
217
- "loss": 0.4017,
218
- "step": 875
219
- },
220
- {
221
- "epoch": 0.13,
222
- "learning_rate": 9.38923076923077e-06,
223
- "loss": 0.3757,
224
- "step": 900
225
- },
226
- {
227
- "epoch": 0.13,
228
- "learning_rate": 9.350769230769231e-06,
229
- "loss": 0.4084,
230
- "step": 925
231
- },
232
- {
233
- "epoch": 0.14,
234
- "learning_rate": 9.312307692307693e-06,
235
- "loss": 0.3821,
236
- "step": 950
237
- },
238
- {
239
- "epoch": 0.14,
240
- "learning_rate": 9.273846153846154e-06,
241
- "loss": 0.421,
242
- "step": 975
243
- },
244
- {
245
- "epoch": 0.14,
246
- "learning_rate": 9.235384615384616e-06,
247
- "loss": 0.4203,
248
- "step": 1000
249
- },
250
- {
251
- "epoch": 0.14,
252
- "eval_loss": 0.4127572178840637,
253
- "eval_runtime": 1802.5775,
254
- "eval_samples_per_second": 3.656,
255
- "eval_steps_per_second": 0.457,
256
- "eval_wer": 28.265624367127064,
257
- "step": 1000
258
- },
259
- {
260
- "epoch": 0.15,
261
- "learning_rate": 9.196923076923078e-06,
262
- "loss": 0.4094,
263
- "step": 1025
264
- },
265
- {
266
- "epoch": 0.15,
267
- "learning_rate": 9.158461538461539e-06,
268
- "loss": 0.3674,
269
- "step": 1050
270
- },
271
- {
272
- "epoch": 0.15,
273
- "learning_rate": 9.12e-06,
274
- "loss": 0.3613,
275
- "step": 1075
276
- },
277
- {
278
- "epoch": 0.16,
279
- "learning_rate": 9.081538461538463e-06,
280
- "loss": 0.3904,
281
- "step": 1100
282
- },
283
- {
284
- "epoch": 0.16,
285
- "learning_rate": 9.043076923076923e-06,
286
- "loss": 0.3865,
287
- "step": 1125
288
- },
289
- {
290
- "epoch": 0.16,
291
- "learning_rate": 9.004615384615386e-06,
292
- "loss": 0.3677,
293
- "step": 1150
294
- },
295
- {
296
- "epoch": 0.17,
297
- "learning_rate": 8.966153846153846e-06,
298
- "loss": 0.3295,
299
- "step": 1175
300
- },
301
- {
302
- "epoch": 0.17,
303
- "learning_rate": 8.927692307692308e-06,
304
- "loss": 0.3207,
305
- "step": 1200
306
- },
307
- {
308
- "epoch": 0.17,
309
- "learning_rate": 8.88923076923077e-06,
310
- "loss": 0.2817,
311
- "step": 1225
312
- },
313
- {
314
- "epoch": 0.18,
315
- "learning_rate": 8.850769230769231e-06,
316
- "loss": 0.3218,
317
- "step": 1250
318
- },
319
- {
320
- "epoch": 0.18,
321
- "learning_rate": 8.812307692307693e-06,
322
- "loss": 0.3502,
323
- "step": 1275
324
- },
325
- {
326
- "epoch": 0.19,
327
- "learning_rate": 8.773846153846155e-06,
328
- "loss": 0.3432,
329
- "step": 1300
330
- },
331
- {
332
- "epoch": 0.19,
333
- "learning_rate": 8.735384615384616e-06,
334
- "loss": 0.3244,
335
- "step": 1325
336
- },
337
- {
338
- "epoch": 0.19,
339
- "learning_rate": 8.696923076923078e-06,
340
- "loss": 0.2607,
341
- "step": 1350
342
- },
343
- {
344
- "epoch": 0.2,
345
- "learning_rate": 8.658461538461538e-06,
346
- "loss": 0.3469,
347
- "step": 1375
348
- },
349
- {
350
- "epoch": 0.2,
351
- "learning_rate": 8.62e-06,
352
- "loss": 0.2756,
353
- "step": 1400
354
- },
355
- {
356
- "epoch": 0.2,
357
- "learning_rate": 8.581538461538463e-06,
358
- "loss": 0.294,
359
- "step": 1425
360
- },
361
- {
362
- "epoch": 0.21,
363
- "learning_rate": 8.543076923076923e-06,
364
- "loss": 0.3915,
365
- "step": 1450
366
- },
367
- {
368
- "epoch": 0.21,
369
- "learning_rate": 8.504615384615385e-06,
370
- "loss": 0.3237,
371
- "step": 1475
372
- },
373
- {
374
- "epoch": 0.21,
375
- "learning_rate": 8.466153846153847e-06,
376
- "loss": 0.2723,
377
- "step": 1500
378
- },
379
- {
380
- "epoch": 0.22,
381
- "learning_rate": 8.427692307692308e-06,
382
- "loss": 0.2917,
383
- "step": 1525
384
- },
385
- {
386
- "epoch": 0.22,
387
- "learning_rate": 8.38923076923077e-06,
388
- "loss": 0.2921,
389
- "step": 1550
390
- },
391
- {
392
- "epoch": 0.23,
393
- "learning_rate": 8.35076923076923e-06,
394
- "loss": 0.3724,
395
- "step": 1575
396
- },
397
- {
398
- "epoch": 0.23,
399
- "learning_rate": 8.312307692307693e-06,
400
- "loss": 0.3364,
401
- "step": 1600
402
- },
403
- {
404
- "epoch": 0.23,
405
- "learning_rate": 8.273846153846155e-06,
406
- "loss": 0.3661,
407
- "step": 1625
408
- },
409
- {
410
- "epoch": 0.24,
411
- "learning_rate": 8.235384615384615e-06,
412
- "loss": 0.3307,
413
- "step": 1650
414
- },
415
- {
416
- "epoch": 0.24,
417
- "learning_rate": 8.196923076923078e-06,
418
- "loss": 0.3201,
419
- "step": 1675
420
- },
421
- {
422
- "epoch": 0.24,
423
- "learning_rate": 8.15846153846154e-06,
424
- "loss": 0.2973,
425
- "step": 1700
426
- },
427
- {
428
- "epoch": 0.25,
429
- "learning_rate": 8.120000000000002e-06,
430
- "loss": 0.3013,
431
- "step": 1725
432
- },
433
- {
434
- "epoch": 0.25,
435
- "learning_rate": 8.081538461538462e-06,
436
- "loss": 0.3349,
437
- "step": 1750
438
- },
439
- {
440
- "epoch": 0.25,
441
- "learning_rate": 8.043076923076923e-06,
442
- "loss": 0.3244,
443
- "step": 1775
444
- },
445
- {
446
- "epoch": 0.26,
447
- "learning_rate": 8.004615384615385e-06,
448
- "loss": 0.3331,
449
- "step": 1800
450
- },
451
- {
452
- "epoch": 0.26,
453
- "learning_rate": 7.966153846153847e-06,
454
- "loss": 0.2723,
455
- "step": 1825
456
- },
457
- {
458
- "epoch": 0.26,
459
- "learning_rate": 7.927692307692308e-06,
460
- "loss": 0.316,
461
- "step": 1850
462
- },
463
- {
464
- "epoch": 0.27,
465
- "learning_rate": 7.88923076923077e-06,
466
- "loss": 0.2648,
467
- "step": 1875
468
- },
469
- {
470
- "epoch": 0.27,
471
- "learning_rate": 7.850769230769232e-06,
472
- "loss": 0.2965,
473
- "step": 1900
474
- },
475
- {
476
- "epoch": 0.28,
477
- "learning_rate": 7.812307692307694e-06,
478
- "loss": 0.256,
479
- "step": 1925
480
- },
481
- {
482
- "epoch": 0.28,
483
- "learning_rate": 7.773846153846155e-06,
484
- "loss": 0.2156,
485
- "step": 1950
486
- },
487
- {
488
- "epoch": 0.28,
489
- "learning_rate": 7.735384615384615e-06,
490
- "loss": 0.2418,
491
- "step": 1975
492
- },
493
- {
494
- "epoch": 0.29,
495
- "learning_rate": 7.696923076923077e-06,
496
- "loss": 0.2693,
497
- "step": 2000
498
- },
499
- {
500
- "epoch": 0.29,
501
- "eval_loss": 0.3239877223968506,
502
- "eval_runtime": 1817.7748,
503
- "eval_samples_per_second": 3.626,
504
- "eval_steps_per_second": 0.453,
505
- "eval_wer": 22.052330997610273,
506
- "step": 2000
507
- },
508
- {
509
- "epoch": 0.29,
510
- "learning_rate": 7.65846153846154e-06,
511
- "loss": 0.2191,
512
- "step": 2025
513
- },
514
- {
515
- "epoch": 0.29,
516
- "learning_rate": 7.620000000000001e-06,
517
- "loss": 0.2437,
518
- "step": 2050
519
- },
520
- {
521
- "epoch": 0.3,
522
- "learning_rate": 7.581538461538462e-06,
523
- "loss": 0.2519,
524
- "step": 2075
525
- },
526
- {
527
- "epoch": 0.3,
528
- "learning_rate": 7.544615384615386e-06,
529
- "loss": 0.3198,
530
- "step": 2100
531
- },
532
- {
533
- "epoch": 0.3,
534
- "learning_rate": 7.5061538461538465e-06,
535
- "loss": 0.2708,
536
- "step": 2125
537
- },
538
- {
539
- "epoch": 0.31,
540
- "learning_rate": 7.467692307692308e-06,
541
- "loss": 0.265,
542
- "step": 2150
543
- },
544
- {
545
- "epoch": 0.31,
546
- "learning_rate": 7.42923076923077e-06,
547
- "loss": 0.2935,
548
- "step": 2175
549
- },
550
- {
551
- "epoch": 0.31,
552
- "learning_rate": 7.390769230769231e-06,
553
- "loss": 0.2595,
554
- "step": 2200
555
- },
556
- {
557
- "epoch": 0.32,
558
- "learning_rate": 7.3523076923076935e-06,
559
- "loss": 0.2613,
560
- "step": 2225
561
- },
562
- {
563
- "epoch": 0.32,
564
- "learning_rate": 7.313846153846155e-06,
565
- "loss": 0.2229,
566
- "step": 2250
567
- },
568
- {
569
- "epoch": 0.33,
570
- "learning_rate": 7.275384615384616e-06,
571
- "loss": 0.2814,
572
- "step": 2275
573
- },
574
- {
575
- "epoch": 0.33,
576
- "learning_rate": 7.236923076923078e-06,
577
- "loss": 0.2681,
578
- "step": 2300
579
- },
580
- {
581
- "epoch": 0.33,
582
- "learning_rate": 7.198461538461539e-06,
583
- "loss": 0.2531,
584
- "step": 2325
585
- },
586
- {
587
- "epoch": 0.34,
588
- "learning_rate": 7.16e-06,
589
- "loss": 0.2104,
590
- "step": 2350
591
- },
592
- {
593
- "epoch": 0.34,
594
- "learning_rate": 7.121538461538462e-06,
595
- "loss": 0.2294,
596
- "step": 2375
597
- },
598
- {
599
- "epoch": 0.34,
600
- "learning_rate": 7.0830769230769235e-06,
601
- "loss": 0.2556,
602
- "step": 2400
603
- },
604
- {
605
- "epoch": 0.35,
606
- "learning_rate": 7.044615384615386e-06,
607
- "loss": 0.2185,
608
- "step": 2425
609
- },
610
- {
611
- "epoch": 0.35,
612
- "learning_rate": 7.006153846153847e-06,
613
- "loss": 0.2585,
614
- "step": 2450
615
- },
616
- {
617
- "epoch": 0.35,
618
- "learning_rate": 6.967692307692308e-06,
619
- "loss": 0.2325,
620
- "step": 2475
621
- },
622
- {
623
- "epoch": 0.36,
624
- "learning_rate": 6.9292307692307705e-06,
625
- "loss": 0.2068,
626
- "step": 2500
627
- },
628
- {
629
- "epoch": 0.36,
630
- "learning_rate": 6.890769230769231e-06,
631
- "loss": 0.2233,
632
- "step": 2525
633
- },
634
- {
635
- "epoch": 0.36,
636
- "learning_rate": 6.852307692307692e-06,
637
- "loss": 0.2649,
638
- "step": 2550
639
- },
640
- {
641
- "epoch": 0.37,
642
- "learning_rate": 6.8138461538461545e-06,
643
- "loss": 0.2524,
644
- "step": 2575
645
- },
646
- {
647
- "epoch": 0.37,
648
- "learning_rate": 6.775384615384616e-06,
649
- "loss": 0.2847,
650
- "step": 2600
651
- },
652
- {
653
- "epoch": 0.38,
654
- "learning_rate": 6.736923076923078e-06,
655
- "loss": 0.1954,
656
- "step": 2625
657
- },
658
- {
659
- "epoch": 0.38,
660
- "learning_rate": 6.698461538461539e-06,
661
- "loss": 0.2553,
662
- "step": 2650
663
- },
664
- {
665
- "epoch": 0.38,
666
- "learning_rate": 6.660000000000001e-06,
667
- "loss": 0.2138,
668
- "step": 2675
669
- },
670
- {
671
- "epoch": 0.39,
672
- "learning_rate": 6.621538461538463e-06,
673
- "loss": 0.2245,
674
- "step": 2700
675
- },
676
- {
677
- "epoch": 0.39,
678
- "learning_rate": 6.583076923076923e-06,
679
- "loss": 0.2137,
680
- "step": 2725
681
- },
682
- {
683
- "epoch": 0.39,
684
- "learning_rate": 6.5446153846153846e-06,
685
- "loss": 0.2558,
686
- "step": 2750
687
- },
688
- {
689
- "epoch": 0.4,
690
- "learning_rate": 6.506153846153847e-06,
691
- "loss": 0.1981,
692
- "step": 2775
693
- },
694
- {
695
- "epoch": 0.4,
696
- "learning_rate": 6.467692307692308e-06,
697
- "loss": 0.2327,
698
- "step": 2800
699
- },
700
- {
701
- "epoch": 0.4,
702
- "learning_rate": 6.42923076923077e-06,
703
- "loss": 0.219,
704
- "step": 2825
705
- },
706
- {
707
- "epoch": 0.41,
708
- "learning_rate": 6.3907692307692315e-06,
709
- "loss": 0.226,
710
- "step": 2850
711
- },
712
- {
713
- "epoch": 0.41,
714
- "learning_rate": 6.352307692307693e-06,
715
- "loss": 0.2192,
716
- "step": 2875
717
- },
718
- {
719
- "epoch": 0.41,
720
- "learning_rate": 6.313846153846155e-06,
721
- "loss": 0.2146,
722
- "step": 2900
723
- },
724
- {
725
- "epoch": 0.42,
726
- "learning_rate": 6.2753846153846155e-06,
727
- "loss": 0.2058,
728
- "step": 2925
729
- },
730
- {
731
- "epoch": 0.42,
732
- "learning_rate": 6.236923076923077e-06,
733
- "loss": 0.2369,
734
- "step": 2950
735
- },
736
- {
737
- "epoch": 0.42,
738
- "learning_rate": 6.198461538461539e-06,
739
- "loss": 0.2202,
740
- "step": 2975
741
- },
742
- {
743
- "epoch": 0.43,
744
- "learning_rate": 6.16e-06,
745
- "loss": 0.2228,
746
- "step": 3000
747
- },
748
- {
749
- "epoch": 0.43,
750
- "eval_loss": 0.27374517917633057,
751
- "eval_runtime": 1805.7724,
752
- "eval_samples_per_second": 3.65,
753
- "eval_steps_per_second": 0.456,
754
- "eval_wer": 18.14370772408765,
755
- "step": 3000
756
- }
757
- ],
758
- "max_steps": 7000,
759
- "num_train_epochs": 9223372036854775807,
760
- "total_flos": 1.224725889024e+19,
761
- "trial_name": null,
762
- "trial_params": null
763
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-3000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ca27d9d1bef820c87cfd52d5adb15af5250e1ac4555cd1703a088ad0dad1a4
3
- size 3643
 
 
 
 
checkpoint-4000/config.json DELETED
@@ -1,41 +0,0 @@
1
- {
2
- "_name_or_path": "openai/whisper-medium",
3
- "activation_dropout": 0.0,
4
- "activation_function": "gelu",
5
- "architectures": [
6
- "WhisperForConditionalGeneration"
7
- ],
8
- "attention_dropout": 0.0,
9
- "begin_suppress_tokens": [
10
- 220,
11
- 50257
12
- ],
13
- "bos_token_id": 50257,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 24,
19
- "decoder_start_token_id": 50258,
20
- "dropout": 0.0,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 24,
25
- "eos_token_id": 50257,
26
- "forced_decoder_ids": null,
27
- "init_std": 0.02,
28
- "is_encoder_decoder": true,
29
- "max_length": 448,
30
- "max_source_positions": 1500,
31
- "max_target_positions": 448,
32
- "model_type": "whisper",
33
- "num_hidden_layers": 24,
34
- "num_mel_bins": 80,
35
- "pad_token_id": 50257,
36
- "scale_embedding": false,
37
- "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
- "use_cache": false,
40
- "vocab_size": 51865
41
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:1d10d5e2e5766ead361dc8ecf77781538635af5fdd98db487bfde99823ce4a07
3
- size 6111428695
 
 
 
 
checkpoint-4000/preprocessor_config.json DELETED
The diff for this file is too large to render. See raw diff
 
checkpoint-4000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:def26fdaf433678a4a7ce858c8c045bcb0081f1a9d18211fd996f65174b4429e
3
- size 3055754841
 
 
 
 
checkpoint-4000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:139a67ed18200d5b28ad0b83b97b2bbd7bf683a055720be06a088fa3bb073509
3
- size 14575
 
 
 
 
checkpoint-4000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8b27cbf906faccb76347b161e471af6eaccfe3470862a556cf92db01b1d1b35b
3
- size 557
 
 
 
 
checkpoint-4000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8c8724f1b8669fb50a95fa5a9cd53bb72dbb12e135946c3302c5f49b3cc9f276
3
- size 627
 
 
 
 
checkpoint-4000/trainer_state.json DELETED
@@ -1,1012 +0,0 @@
1
- {
2
- "best_metric": 16.353436753209934,
3
- "best_model_checkpoint": "./checkpoint-4000",
4
- "epoch": 1.1005714285714285,
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": 0.0,
12
- "learning_rate": 4.6000000000000004e-07,
13
- "loss": 1.4182,
14
- "step": 25
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.400000000000001e-07,
19
- "loss": 1.292,
20
- "step": 50
21
- },
22
- {
23
- "epoch": 0.01,
24
- "learning_rate": 1.44e-06,
25
- "loss": 1.0018,
26
- "step": 75
27
- },
28
- {
29
- "epoch": 0.01,
30
- "learning_rate": 1.94e-06,
31
- "loss": 0.7765,
32
- "step": 100
33
- },
34
- {
35
- "epoch": 0.02,
36
- "learning_rate": 2.4400000000000004e-06,
37
- "loss": 0.7103,
38
- "step": 125
39
- },
40
- {
41
- "epoch": 0.02,
42
- "learning_rate": 2.9400000000000002e-06,
43
- "loss": 0.6597,
44
- "step": 150
45
- },
46
- {
47
- "epoch": 0.03,
48
- "learning_rate": 3.44e-06,
49
- "loss": 0.6657,
50
- "step": 175
51
- },
52
- {
53
- "epoch": 0.03,
54
- "learning_rate": 3.94e-06,
55
- "loss": 0.5853,
56
- "step": 200
57
- },
58
- {
59
- "epoch": 0.03,
60
- "learning_rate": 4.440000000000001e-06,
61
- "loss": 0.5273,
62
- "step": 225
63
- },
64
- {
65
- "epoch": 0.04,
66
- "learning_rate": 4.94e-06,
67
- "loss": 0.5979,
68
- "step": 250
69
- },
70
- {
71
- "epoch": 0.04,
72
- "learning_rate": 5.4400000000000004e-06,
73
- "loss": 0.5861,
74
- "step": 275
75
- },
76
- {
77
- "epoch": 0.04,
78
- "learning_rate": 5.94e-06,
79
- "loss": 0.5085,
80
- "step": 300
81
- },
82
- {
83
- "epoch": 0.05,
84
- "learning_rate": 6.440000000000001e-06,
85
- "loss": 0.4827,
86
- "step": 325
87
- },
88
- {
89
- "epoch": 0.05,
90
- "learning_rate": 6.9400000000000005e-06,
91
- "loss": 0.4909,
92
- "step": 350
93
- },
94
- {
95
- "epoch": 0.05,
96
- "learning_rate": 7.440000000000001e-06,
97
- "loss": 0.4651,
98
- "step": 375
99
- },
100
- {
101
- "epoch": 0.06,
102
- "learning_rate": 7.94e-06,
103
- "loss": 0.494,
104
- "step": 400
105
- },
106
- {
107
- "epoch": 0.06,
108
- "learning_rate": 8.44e-06,
109
- "loss": 0.4188,
110
- "step": 425
111
- },
112
- {
113
- "epoch": 0.06,
114
- "learning_rate": 8.94e-06,
115
- "loss": 0.3849,
116
- "step": 450
117
- },
118
- {
119
- "epoch": 0.07,
120
- "learning_rate": 9.440000000000001e-06,
121
- "loss": 0.4577,
122
- "step": 475
123
- },
124
- {
125
- "epoch": 0.07,
126
- "learning_rate": 9.940000000000001e-06,
127
- "loss": 0.4415,
128
- "step": 500
129
- },
130
- {
131
- "epoch": 0.07,
132
- "learning_rate": 9.966153846153847e-06,
133
- "loss": 0.4614,
134
- "step": 525
135
- },
136
- {
137
- "epoch": 0.08,
138
- "learning_rate": 9.927692307692309e-06,
139
- "loss": 0.4283,
140
- "step": 550
141
- },
142
- {
143
- "epoch": 0.08,
144
- "learning_rate": 9.88923076923077e-06,
145
- "loss": 0.4486,
146
- "step": 575
147
- },
148
- {
149
- "epoch": 0.09,
150
- "learning_rate": 9.850769230769231e-06,
151
- "loss": 0.4434,
152
- "step": 600
153
- },
154
- {
155
- "epoch": 0.09,
156
- "learning_rate": 9.812307692307694e-06,
157
- "loss": 0.4245,
158
- "step": 625
159
- },
160
- {
161
- "epoch": 0.09,
162
- "learning_rate": 9.773846153846154e-06,
163
- "loss": 0.4254,
164
- "step": 650
165
- },
166
- {
167
- "epoch": 0.1,
168
- "learning_rate": 9.735384615384616e-06,
169
- "loss": 0.4143,
170
- "step": 675
171
- },
172
- {
173
- "epoch": 0.1,
174
- "learning_rate": 9.696923076923078e-06,
175
- "loss": 0.4115,
176
- "step": 700
177
- },
178
- {
179
- "epoch": 0.1,
180
- "learning_rate": 9.658461538461539e-06,
181
- "loss": 0.3656,
182
- "step": 725
183
- },
184
- {
185
- "epoch": 0.11,
186
- "learning_rate": 9.620000000000001e-06,
187
- "loss": 0.4004,
188
- "step": 750
189
- },
190
- {
191
- "epoch": 0.11,
192
- "learning_rate": 9.581538461538462e-06,
193
- "loss": 0.388,
194
- "step": 775
195
- },
196
- {
197
- "epoch": 0.11,
198
- "learning_rate": 9.543076923076924e-06,
199
- "loss": 0.3665,
200
- "step": 800
201
- },
202
- {
203
- "epoch": 0.12,
204
- "learning_rate": 9.504615384615386e-06,
205
- "loss": 0.3777,
206
- "step": 825
207
- },
208
- {
209
- "epoch": 0.12,
210
- "learning_rate": 9.466153846153846e-06,
211
- "loss": 0.4338,
212
- "step": 850
213
- },
214
- {
215
- "epoch": 0.12,
216
- "learning_rate": 9.427692307692309e-06,
217
- "loss": 0.4017,
218
- "step": 875
219
- },
220
- {
221
- "epoch": 0.13,
222
- "learning_rate": 9.38923076923077e-06,
223
- "loss": 0.3757,
224
- "step": 900
225
- },
226
- {
227
- "epoch": 0.13,
228
- "learning_rate": 9.350769230769231e-06,
229
- "loss": 0.4084,
230
- "step": 925
231
- },
232
- {
233
- "epoch": 0.14,
234
- "learning_rate": 9.312307692307693e-06,
235
- "loss": 0.3821,
236
- "step": 950
237
- },
238
- {
239
- "epoch": 0.14,
240
- "learning_rate": 9.273846153846154e-06,
241
- "loss": 0.421,
242
- "step": 975
243
- },
244
- {
245
- "epoch": 0.14,
246
- "learning_rate": 9.235384615384616e-06,
247
- "loss": 0.4203,
248
- "step": 1000
249
- },
250
- {
251
- "epoch": 0.14,
252
- "eval_loss": 0.4127572178840637,
253
- "eval_runtime": 1802.5775,
254
- "eval_samples_per_second": 3.656,
255
- "eval_steps_per_second": 0.457,
256
- "eval_wer": 28.265624367127064,
257
- "step": 1000
258
- },
259
- {
260
- "epoch": 0.15,
261
- "learning_rate": 9.196923076923078e-06,
262
- "loss": 0.4094,
263
- "step": 1025
264
- },
265
- {
266
- "epoch": 0.15,
267
- "learning_rate": 9.158461538461539e-06,
268
- "loss": 0.3674,
269
- "step": 1050
270
- },
271
- {
272
- "epoch": 0.15,
273
- "learning_rate": 9.12e-06,
274
- "loss": 0.3613,
275
- "step": 1075
276
- },
277
- {
278
- "epoch": 0.16,
279
- "learning_rate": 9.081538461538463e-06,
280
- "loss": 0.3904,
281
- "step": 1100
282
- },
283
- {
284
- "epoch": 0.16,
285
- "learning_rate": 9.043076923076923e-06,
286
- "loss": 0.3865,
287
- "step": 1125
288
- },
289
- {
290
- "epoch": 0.16,
291
- "learning_rate": 9.004615384615386e-06,
292
- "loss": 0.3677,
293
- "step": 1150
294
- },
295
- {
296
- "epoch": 0.17,
297
- "learning_rate": 8.966153846153846e-06,
298
- "loss": 0.3295,
299
- "step": 1175
300
- },
301
- {
302
- "epoch": 0.17,
303
- "learning_rate": 8.927692307692308e-06,
304
- "loss": 0.3207,
305
- "step": 1200
306
- },
307
- {
308
- "epoch": 0.17,
309
- "learning_rate": 8.88923076923077e-06,
310
- "loss": 0.2817,
311
- "step": 1225
312
- },
313
- {
314
- "epoch": 0.18,
315
- "learning_rate": 8.850769230769231e-06,
316
- "loss": 0.3218,
317
- "step": 1250
318
- },
319
- {
320
- "epoch": 0.18,
321
- "learning_rate": 8.812307692307693e-06,
322
- "loss": 0.3502,
323
- "step": 1275
324
- },
325
- {
326
- "epoch": 0.19,
327
- "learning_rate": 8.773846153846155e-06,
328
- "loss": 0.3432,
329
- "step": 1300
330
- },
331
- {
332
- "epoch": 0.19,
333
- "learning_rate": 8.735384615384616e-06,
334
- "loss": 0.3244,
335
- "step": 1325
336
- },
337
- {
338
- "epoch": 0.19,
339
- "learning_rate": 8.696923076923078e-06,
340
- "loss": 0.2607,
341
- "step": 1350
342
- },
343
- {
344
- "epoch": 0.2,
345
- "learning_rate": 8.658461538461538e-06,
346
- "loss": 0.3469,
347
- "step": 1375
348
- },
349
- {
350
- "epoch": 0.2,
351
- "learning_rate": 8.62e-06,
352
- "loss": 0.2756,
353
- "step": 1400
354
- },
355
- {
356
- "epoch": 0.2,
357
- "learning_rate": 8.581538461538463e-06,
358
- "loss": 0.294,
359
- "step": 1425
360
- },
361
- {
362
- "epoch": 0.21,
363
- "learning_rate": 8.543076923076923e-06,
364
- "loss": 0.3915,
365
- "step": 1450
366
- },
367
- {
368
- "epoch": 0.21,
369
- "learning_rate": 8.504615384615385e-06,
370
- "loss": 0.3237,
371
- "step": 1475
372
- },
373
- {
374
- "epoch": 0.21,
375
- "learning_rate": 8.466153846153847e-06,
376
- "loss": 0.2723,
377
- "step": 1500
378
- },
379
- {
380
- "epoch": 0.22,
381
- "learning_rate": 8.427692307692308e-06,
382
- "loss": 0.2917,
383
- "step": 1525
384
- },
385
- {
386
- "epoch": 0.22,
387
- "learning_rate": 8.38923076923077e-06,
388
- "loss": 0.2921,
389
- "step": 1550
390
- },
391
- {
392
- "epoch": 0.23,
393
- "learning_rate": 8.35076923076923e-06,
394
- "loss": 0.3724,
395
- "step": 1575
396
- },
397
- {
398
- "epoch": 0.23,
399
- "learning_rate": 8.312307692307693e-06,
400
- "loss": 0.3364,
401
- "step": 1600
402
- },
403
- {
404
- "epoch": 0.23,
405
- "learning_rate": 8.273846153846155e-06,
406
- "loss": 0.3661,
407
- "step": 1625
408
- },
409
- {
410
- "epoch": 0.24,
411
- "learning_rate": 8.235384615384615e-06,
412
- "loss": 0.3307,
413
- "step": 1650
414
- },
415
- {
416
- "epoch": 0.24,
417
- "learning_rate": 8.196923076923078e-06,
418
- "loss": 0.3201,
419
- "step": 1675
420
- },
421
- {
422
- "epoch": 0.24,
423
- "learning_rate": 8.15846153846154e-06,
424
- "loss": 0.2973,
425
- "step": 1700
426
- },
427
- {
428
- "epoch": 0.25,
429
- "learning_rate": 8.120000000000002e-06,
430
- "loss": 0.3013,
431
- "step": 1725
432
- },
433
- {
434
- "epoch": 0.25,
435
- "learning_rate": 8.081538461538462e-06,
436
- "loss": 0.3349,
437
- "step": 1750
438
- },
439
- {
440
- "epoch": 0.25,
441
- "learning_rate": 8.043076923076923e-06,
442
- "loss": 0.3244,
443
- "step": 1775
444
- },
445
- {
446
- "epoch": 0.26,
447
- "learning_rate": 8.004615384615385e-06,
448
- "loss": 0.3331,
449
- "step": 1800
450
- },
451
- {
452
- "epoch": 0.26,
453
- "learning_rate": 7.966153846153847e-06,
454
- "loss": 0.2723,
455
- "step": 1825
456
- },
457
- {
458
- "epoch": 0.26,
459
- "learning_rate": 7.927692307692308e-06,
460
- "loss": 0.316,
461
- "step": 1850
462
- },
463
- {
464
- "epoch": 0.27,
465
- "learning_rate": 7.88923076923077e-06,
466
- "loss": 0.2648,
467
- "step": 1875
468
- },
469
- {
470
- "epoch": 0.27,
471
- "learning_rate": 7.850769230769232e-06,
472
- "loss": 0.2965,
473
- "step": 1900
474
- },
475
- {
476
- "epoch": 0.28,
477
- "learning_rate": 7.812307692307694e-06,
478
- "loss": 0.256,
479
- "step": 1925
480
- },
481
- {
482
- "epoch": 0.28,
483
- "learning_rate": 7.773846153846155e-06,
484
- "loss": 0.2156,
485
- "step": 1950
486
- },
487
- {
488
- "epoch": 0.28,
489
- "learning_rate": 7.735384615384615e-06,
490
- "loss": 0.2418,
491
- "step": 1975
492
- },
493
- {
494
- "epoch": 0.29,
495
- "learning_rate": 7.696923076923077e-06,
496
- "loss": 0.2693,
497
- "step": 2000
498
- },
499
- {
500
- "epoch": 0.29,
501
- "eval_loss": 0.3239877223968506,
502
- "eval_runtime": 1817.7748,
503
- "eval_samples_per_second": 3.626,
504
- "eval_steps_per_second": 0.453,
505
- "eval_wer": 22.052330997610273,
506
- "step": 2000
507
- },
508
- {
509
- "epoch": 0.29,
510
- "learning_rate": 7.65846153846154e-06,
511
- "loss": 0.2191,
512
- "step": 2025
513
- },
514
- {
515
- "epoch": 0.29,
516
- "learning_rate": 7.620000000000001e-06,
517
- "loss": 0.2437,
518
- "step": 2050
519
- },
520
- {
521
- "epoch": 0.3,
522
- "learning_rate": 7.581538461538462e-06,
523
- "loss": 0.2519,
524
- "step": 2075
525
- },
526
- {
527
- "epoch": 0.3,
528
- "learning_rate": 7.544615384615386e-06,
529
- "loss": 0.3198,
530
- "step": 2100
531
- },
532
- {
533
- "epoch": 0.3,
534
- "learning_rate": 7.5061538461538465e-06,
535
- "loss": 0.2708,
536
- "step": 2125
537
- },
538
- {
539
- "epoch": 0.31,
540
- "learning_rate": 7.467692307692308e-06,
541
- "loss": 0.265,
542
- "step": 2150
543
- },
544
- {
545
- "epoch": 0.31,
546
- "learning_rate": 7.42923076923077e-06,
547
- "loss": 0.2935,
548
- "step": 2175
549
- },
550
- {
551
- "epoch": 0.31,
552
- "learning_rate": 7.390769230769231e-06,
553
- "loss": 0.2595,
554
- "step": 2200
555
- },
556
- {
557
- "epoch": 0.32,
558
- "learning_rate": 7.3523076923076935e-06,
559
- "loss": 0.2613,
560
- "step": 2225
561
- },
562
- {
563
- "epoch": 0.32,
564
- "learning_rate": 7.313846153846155e-06,
565
- "loss": 0.2229,
566
- "step": 2250
567
- },
568
- {
569
- "epoch": 0.33,
570
- "learning_rate": 7.275384615384616e-06,
571
- "loss": 0.2814,
572
- "step": 2275
573
- },
574
- {
575
- "epoch": 0.33,
576
- "learning_rate": 7.236923076923078e-06,
577
- "loss": 0.2681,
578
- "step": 2300
579
- },
580
- {
581
- "epoch": 0.33,
582
- "learning_rate": 7.198461538461539e-06,
583
- "loss": 0.2531,
584
- "step": 2325
585
- },
586
- {
587
- "epoch": 0.34,
588
- "learning_rate": 7.16e-06,
589
- "loss": 0.2104,
590
- "step": 2350
591
- },
592
- {
593
- "epoch": 0.34,
594
- "learning_rate": 7.121538461538462e-06,
595
- "loss": 0.2294,
596
- "step": 2375
597
- },
598
- {
599
- "epoch": 0.34,
600
- "learning_rate": 7.0830769230769235e-06,
601
- "loss": 0.2556,
602
- "step": 2400
603
- },
604
- {
605
- "epoch": 0.35,
606
- "learning_rate": 7.044615384615386e-06,
607
- "loss": 0.2185,
608
- "step": 2425
609
- },
610
- {
611
- "epoch": 0.35,
612
- "learning_rate": 7.006153846153847e-06,
613
- "loss": 0.2585,
614
- "step": 2450
615
- },
616
- {
617
- "epoch": 0.35,
618
- "learning_rate": 6.967692307692308e-06,
619
- "loss": 0.2325,
620
- "step": 2475
621
- },
622
- {
623
- "epoch": 0.36,
624
- "learning_rate": 6.9292307692307705e-06,
625
- "loss": 0.2068,
626
- "step": 2500
627
- },
628
- {
629
- "epoch": 0.36,
630
- "learning_rate": 6.890769230769231e-06,
631
- "loss": 0.2233,
632
- "step": 2525
633
- },
634
- {
635
- "epoch": 0.36,
636
- "learning_rate": 6.852307692307692e-06,
637
- "loss": 0.2649,
638
- "step": 2550
639
- },
640
- {
641
- "epoch": 0.37,
642
- "learning_rate": 6.8138461538461545e-06,
643
- "loss": 0.2524,
644
- "step": 2575
645
- },
646
- {
647
- "epoch": 0.37,
648
- "learning_rate": 6.775384615384616e-06,
649
- "loss": 0.2847,
650
- "step": 2600
651
- },
652
- {
653
- "epoch": 0.38,
654
- "learning_rate": 6.736923076923078e-06,
655
- "loss": 0.1954,
656
- "step": 2625
657
- },
658
- {
659
- "epoch": 0.38,
660
- "learning_rate": 6.698461538461539e-06,
661
- "loss": 0.2553,
662
- "step": 2650
663
- },
664
- {
665
- "epoch": 0.38,
666
- "learning_rate": 6.660000000000001e-06,
667
- "loss": 0.2138,
668
- "step": 2675
669
- },
670
- {
671
- "epoch": 0.39,
672
- "learning_rate": 6.621538461538463e-06,
673
- "loss": 0.2245,
674
- "step": 2700
675
- },
676
- {
677
- "epoch": 0.39,
678
- "learning_rate": 6.583076923076923e-06,
679
- "loss": 0.2137,
680
- "step": 2725
681
- },
682
- {
683
- "epoch": 0.39,
684
- "learning_rate": 6.5446153846153846e-06,
685
- "loss": 0.2558,
686
- "step": 2750
687
- },
688
- {
689
- "epoch": 0.4,
690
- "learning_rate": 6.506153846153847e-06,
691
- "loss": 0.1981,
692
- "step": 2775
693
- },
694
- {
695
- "epoch": 0.4,
696
- "learning_rate": 6.467692307692308e-06,
697
- "loss": 0.2327,
698
- "step": 2800
699
- },
700
- {
701
- "epoch": 0.4,
702
- "learning_rate": 6.42923076923077e-06,
703
- "loss": 0.219,
704
- "step": 2825
705
- },
706
- {
707
- "epoch": 0.41,
708
- "learning_rate": 6.3907692307692315e-06,
709
- "loss": 0.226,
710
- "step": 2850
711
- },
712
- {
713
- "epoch": 0.41,
714
- "learning_rate": 6.352307692307693e-06,
715
- "loss": 0.2192,
716
- "step": 2875
717
- },
718
- {
719
- "epoch": 0.41,
720
- "learning_rate": 6.313846153846155e-06,
721
- "loss": 0.2146,
722
- "step": 2900
723
- },
724
- {
725
- "epoch": 0.42,
726
- "learning_rate": 6.2753846153846155e-06,
727
- "loss": 0.2058,
728
- "step": 2925
729
- },
730
- {
731
- "epoch": 0.42,
732
- "learning_rate": 6.236923076923077e-06,
733
- "loss": 0.2369,
734
- "step": 2950
735
- },
736
- {
737
- "epoch": 0.42,
738
- "learning_rate": 6.198461538461539e-06,
739
- "loss": 0.2202,
740
- "step": 2975
741
- },
742
- {
743
- "epoch": 0.43,
744
- "learning_rate": 6.16e-06,
745
- "loss": 0.2228,
746
- "step": 3000
747
- },
748
- {
749
- "epoch": 0.43,
750
- "eval_loss": 0.27374517917633057,
751
- "eval_runtime": 1805.7724,
752
- "eval_samples_per_second": 3.65,
753
- "eval_steps_per_second": 0.456,
754
- "eval_wer": 18.14370772408765,
755
- "step": 3000
756
- },
757
- {
758
- "epoch": 0.43,
759
- "learning_rate": 6.1215384615384625e-06,
760
- "loss": 0.2029,
761
- "step": 3025
762
- },
763
- {
764
- "epoch": 0.44,
765
- "learning_rate": 6.083076923076924e-06,
766
- "loss": 0.2085,
767
- "step": 3050
768
- },
769
- {
770
- "epoch": 0.44,
771
- "learning_rate": 6.044615384615385e-06,
772
- "loss": 0.196,
773
- "step": 3075
774
- },
775
- {
776
- "epoch": 0.44,
777
- "learning_rate": 6.006153846153847e-06,
778
- "loss": 0.1999,
779
- "step": 3100
780
- },
781
- {
782
- "epoch": 0.45,
783
- "learning_rate": 5.967692307692308e-06,
784
- "loss": 0.2076,
785
- "step": 3125
786
- },
787
- {
788
- "epoch": 0.45,
789
- "learning_rate": 5.929230769230769e-06,
790
- "loss": 0.1999,
791
- "step": 3150
792
- },
793
- {
794
- "epoch": 0.45,
795
- "learning_rate": 5.890769230769231e-06,
796
- "loss": 0.2361,
797
- "step": 3175
798
- },
799
- {
800
- "epoch": 0.46,
801
- "learning_rate": 5.8523076923076926e-06,
802
- "loss": 0.2003,
803
- "step": 3200
804
- },
805
- {
806
- "epoch": 0.46,
807
- "learning_rate": 5.813846153846155e-06,
808
- "loss": 0.249,
809
- "step": 3225
810
- },
811
- {
812
- "epoch": 0.46,
813
- "learning_rate": 5.775384615384616e-06,
814
- "loss": 0.2724,
815
- "step": 3250
816
- },
817
- {
818
- "epoch": 0.47,
819
- "learning_rate": 5.736923076923077e-06,
820
- "loss": 0.2001,
821
- "step": 3275
822
- },
823
- {
824
- "epoch": 1.0,
825
- "learning_rate": 5.6984615384615395e-06,
826
- "loss": 0.2355,
827
- "step": 3300
828
- },
829
- {
830
- "epoch": 1.0,
831
- "learning_rate": 5.66e-06,
832
- "loss": 0.2297,
833
- "step": 3325
834
- },
835
- {
836
- "epoch": 1.01,
837
- "learning_rate": 5.621538461538461e-06,
838
- "loss": 0.2265,
839
- "step": 3350
840
- },
841
- {
842
- "epoch": 1.01,
843
- "learning_rate": 5.5830769230769235e-06,
844
- "loss": 0.21,
845
- "step": 3375
846
- },
847
- {
848
- "epoch": 1.01,
849
- "learning_rate": 5.544615384615385e-06,
850
- "loss": 0.1944,
851
- "step": 3400
852
- },
853
- {
854
- "epoch": 1.02,
855
- "learning_rate": 5.506153846153847e-06,
856
- "loss": 0.1844,
857
- "step": 3425
858
- },
859
- {
860
- "epoch": 1.02,
861
- "learning_rate": 5.467692307692308e-06,
862
- "loss": 0.1774,
863
- "step": 3450
864
- },
865
- {
866
- "epoch": 1.03,
867
- "learning_rate": 5.42923076923077e-06,
868
- "loss": 0.1688,
869
- "step": 3475
870
- },
871
- {
872
- "epoch": 1.03,
873
- "learning_rate": 5.390769230769232e-06,
874
- "loss": 0.1473,
875
- "step": 3500
876
- },
877
- {
878
- "epoch": 1.03,
879
- "learning_rate": 5.352307692307692e-06,
880
- "loss": 0.1743,
881
- "step": 3525
882
- },
883
- {
884
- "epoch": 1.04,
885
- "learning_rate": 5.313846153846154e-06,
886
- "loss": 0.1759,
887
- "step": 3550
888
- },
889
- {
890
- "epoch": 1.04,
891
- "learning_rate": 5.275384615384616e-06,
892
- "loss": 0.1507,
893
- "step": 3575
894
- },
895
- {
896
- "epoch": 1.04,
897
- "learning_rate": 5.236923076923077e-06,
898
- "loss": 0.1556,
899
- "step": 3600
900
- },
901
- {
902
- "epoch": 1.05,
903
- "learning_rate": 5.198461538461539e-06,
904
- "loss": 0.1545,
905
- "step": 3625
906
- },
907
- {
908
- "epoch": 1.05,
909
- "learning_rate": 5.1600000000000006e-06,
910
- "loss": 0.1143,
911
- "step": 3650
912
- },
913
- {
914
- "epoch": 1.05,
915
- "learning_rate": 5.121538461538462e-06,
916
- "loss": 0.1168,
917
- "step": 3675
918
- },
919
- {
920
- "epoch": 1.06,
921
- "learning_rate": 5.083076923076924e-06,
922
- "loss": 0.1095,
923
- "step": 3700
924
- },
925
- {
926
- "epoch": 1.06,
927
- "learning_rate": 5.044615384615385e-06,
928
- "loss": 0.1098,
929
- "step": 3725
930
- },
931
- {
932
- "epoch": 1.06,
933
- "learning_rate": 5.006153846153846e-06,
934
- "loss": 0.102,
935
- "step": 3750
936
- },
937
- {
938
- "epoch": 1.07,
939
- "learning_rate": 4.967692307692308e-06,
940
- "loss": 0.1263,
941
- "step": 3775
942
- },
943
- {
944
- "epoch": 1.07,
945
- "learning_rate": 4.929230769230769e-06,
946
- "loss": 0.093,
947
- "step": 3800
948
- },
949
- {
950
- "epoch": 1.08,
951
- "learning_rate": 4.8907692307692315e-06,
952
- "loss": 0.1163,
953
- "step": 3825
954
- },
955
- {
956
- "epoch": 1.08,
957
- "learning_rate": 4.852307692307693e-06,
958
- "loss": 0.1005,
959
- "step": 3850
960
- },
961
- {
962
- "epoch": 1.08,
963
- "learning_rate": 4.813846153846154e-06,
964
- "loss": 0.1036,
965
- "step": 3875
966
- },
967
- {
968
- "epoch": 1.09,
969
- "learning_rate": 4.7753846153846155e-06,
970
- "loss": 0.1289,
971
- "step": 3900
972
- },
973
- {
974
- "epoch": 1.09,
975
- "learning_rate": 4.736923076923078e-06,
976
- "loss": 0.1238,
977
- "step": 3925
978
- },
979
- {
980
- "epoch": 1.09,
981
- "learning_rate": 4.698461538461539e-06,
982
- "loss": 0.1218,
983
- "step": 3950
984
- },
985
- {
986
- "epoch": 1.1,
987
- "learning_rate": 4.66e-06,
988
- "loss": 0.1223,
989
- "step": 3975
990
- },
991
- {
992
- "epoch": 1.1,
993
- "learning_rate": 4.621538461538462e-06,
994
- "loss": 0.1002,
995
- "step": 4000
996
- },
997
- {
998
- "epoch": 1.1,
999
- "eval_loss": 0.2553846836090088,
1000
- "eval_runtime": 1819.1705,
1001
- "eval_samples_per_second": 3.623,
1002
- "eval_steps_per_second": 0.453,
1003
- "eval_wer": 16.353436753209934,
1004
- "step": 4000
1005
- }
1006
- ],
1007
- "max_steps": 7000,
1008
- "num_train_epochs": 9223372036854775807,
1009
- "total_flos": 1.632763731050496e+19,
1010
- "trial_name": null,
1011
- "trial_params": null
1012
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-4000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ca27d9d1bef820c87cfd52d5adb15af5250e1ac4555cd1703a088ad0dad1a4
3
- size 3643
 
 
 
 
checkpoint-5000/config.json DELETED
@@ -1,41 +0,0 @@
1
- {
2
- "_name_or_path": "openai/whisper-medium",
3
- "activation_dropout": 0.0,
4
- "activation_function": "gelu",
5
- "architectures": [
6
- "WhisperForConditionalGeneration"
7
- ],
8
- "attention_dropout": 0.0,
9
- "begin_suppress_tokens": [
10
- 220,
11
- 50257
12
- ],
13
- "bos_token_id": 50257,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 24,
19
- "decoder_start_token_id": 50258,
20
- "dropout": 0.0,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 24,
25
- "eos_token_id": 50257,
26
- "forced_decoder_ids": null,
27
- "init_std": 0.02,
28
- "is_encoder_decoder": true,
29
- "max_length": 448,
30
- "max_source_positions": 1500,
31
- "max_target_positions": 448,
32
- "model_type": "whisper",
33
- "num_hidden_layers": 24,
34
- "num_mel_bins": 80,
35
- "pad_token_id": 50257,
36
- "scale_embedding": false,
37
- "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
- "use_cache": false,
40
- "vocab_size": 51865
41
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-5000/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:dae73746a9cbfc8af8231a105d0ff0c0a1569e44353a8f7f629de550c1b98935
3
- size 6111428695
 
 
 
 
checkpoint-5000/preprocessor_config.json DELETED
The diff for this file is too large to render. See raw diff
 
checkpoint-5000/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:cd978f2c702bba55edafb2fe89ef7706d888507a7bf7803905115ee1cbeee805
3
- size 3055754841
 
 
 
 
checkpoint-5000/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:507ba841a78b27a21beeb8be24ea95ea58bc2145f1c0364911d4f242151bdde4
3
- size 14511
 
 
 
 
checkpoint-5000/scaler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:949ca9c3bcb3a1db4317303e71acb4df5237388410fcf06e1c212003108bf76f
3
- size 557
 
 
 
 
checkpoint-5000/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b6704f8857293535a5120dbc970e54bdc93a29e478b654c1b6cc889e897b72c1
3
- size 627
 
 
 
 
checkpoint-5000/trainer_state.json DELETED
@@ -1,1261 +0,0 @@
1
- {
2
- "best_metric": 14.787962250394912,
3
- "best_model_checkpoint": "./checkpoint-5000",
4
- "epoch": 1.2434285714285713,
5
- "global_step": 5000,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.0,
12
- "learning_rate": 4.6000000000000004e-07,
13
- "loss": 1.4182,
14
- "step": 25
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.400000000000001e-07,
19
- "loss": 1.292,
20
- "step": 50
21
- },
22
- {
23
- "epoch": 0.01,
24
- "learning_rate": 1.44e-06,
25
- "loss": 1.0018,
26
- "step": 75
27
- },
28
- {
29
- "epoch": 0.01,
30
- "learning_rate": 1.94e-06,
31
- "loss": 0.7765,
32
- "step": 100
33
- },
34
- {
35
- "epoch": 0.02,
36
- "learning_rate": 2.4400000000000004e-06,
37
- "loss": 0.7103,
38
- "step": 125
39
- },
40
- {
41
- "epoch": 0.02,
42
- "learning_rate": 2.9400000000000002e-06,
43
- "loss": 0.6597,
44
- "step": 150
45
- },
46
- {
47
- "epoch": 0.03,
48
- "learning_rate": 3.44e-06,
49
- "loss": 0.6657,
50
- "step": 175
51
- },
52
- {
53
- "epoch": 0.03,
54
- "learning_rate": 3.94e-06,
55
- "loss": 0.5853,
56
- "step": 200
57
- },
58
- {
59
- "epoch": 0.03,
60
- "learning_rate": 4.440000000000001e-06,
61
- "loss": 0.5273,
62
- "step": 225
63
- },
64
- {
65
- "epoch": 0.04,
66
- "learning_rate": 4.94e-06,
67
- "loss": 0.5979,
68
- "step": 250
69
- },
70
- {
71
- "epoch": 0.04,
72
- "learning_rate": 5.4400000000000004e-06,
73
- "loss": 0.5861,
74
- "step": 275
75
- },
76
- {
77
- "epoch": 0.04,
78
- "learning_rate": 5.94e-06,
79
- "loss": 0.5085,
80
- "step": 300
81
- },
82
- {
83
- "epoch": 0.05,
84
- "learning_rate": 6.440000000000001e-06,
85
- "loss": 0.4827,
86
- "step": 325
87
- },
88
- {
89
- "epoch": 0.05,
90
- "learning_rate": 6.9400000000000005e-06,
91
- "loss": 0.4909,
92
- "step": 350
93
- },
94
- {
95
- "epoch": 0.05,
96
- "learning_rate": 7.440000000000001e-06,
97
- "loss": 0.4651,
98
- "step": 375
99
- },
100
- {
101
- "epoch": 0.06,
102
- "learning_rate": 7.94e-06,
103
- "loss": 0.494,
104
- "step": 400
105
- },
106
- {
107
- "epoch": 0.06,
108
- "learning_rate": 8.44e-06,
109
- "loss": 0.4188,
110
- "step": 425
111
- },
112
- {
113
- "epoch": 0.06,
114
- "learning_rate": 8.94e-06,
115
- "loss": 0.3849,
116
- "step": 450
117
- },
118
- {
119
- "epoch": 0.07,
120
- "learning_rate": 9.440000000000001e-06,
121
- "loss": 0.4577,
122
- "step": 475
123
- },
124
- {
125
- "epoch": 0.07,
126
- "learning_rate": 9.940000000000001e-06,
127
- "loss": 0.4415,
128
- "step": 500
129
- },
130
- {
131
- "epoch": 0.07,
132
- "learning_rate": 9.966153846153847e-06,
133
- "loss": 0.4614,
134
- "step": 525
135
- },
136
- {
137
- "epoch": 0.08,
138
- "learning_rate": 9.927692307692309e-06,
139
- "loss": 0.4283,
140
- "step": 550
141
- },
142
- {
143
- "epoch": 0.08,
144
- "learning_rate": 9.88923076923077e-06,
145
- "loss": 0.4486,
146
- "step": 575
147
- },
148
- {
149
- "epoch": 0.09,
150
- "learning_rate": 9.850769230769231e-06,
151
- "loss": 0.4434,
152
- "step": 600
153
- },
154
- {
155
- "epoch": 0.09,
156
- "learning_rate": 9.812307692307694e-06,
157
- "loss": 0.4245,
158
- "step": 625
159
- },
160
- {
161
- "epoch": 0.09,
162
- "learning_rate": 9.773846153846154e-06,
163
- "loss": 0.4254,
164
- "step": 650
165
- },
166
- {
167
- "epoch": 0.1,
168
- "learning_rate": 9.735384615384616e-06,
169
- "loss": 0.4143,
170
- "step": 675
171
- },
172
- {
173
- "epoch": 0.1,
174
- "learning_rate": 9.696923076923078e-06,
175
- "loss": 0.4115,
176
- "step": 700
177
- },
178
- {
179
- "epoch": 0.1,
180
- "learning_rate": 9.658461538461539e-06,
181
- "loss": 0.3656,
182
- "step": 725
183
- },
184
- {
185
- "epoch": 0.11,
186
- "learning_rate": 9.620000000000001e-06,
187
- "loss": 0.4004,
188
- "step": 750
189
- },
190
- {
191
- "epoch": 0.11,
192
- "learning_rate": 9.581538461538462e-06,
193
- "loss": 0.388,
194
- "step": 775
195
- },
196
- {
197
- "epoch": 0.11,
198
- "learning_rate": 9.543076923076924e-06,
199
- "loss": 0.3665,
200
- "step": 800
201
- },
202
- {
203
- "epoch": 0.12,
204
- "learning_rate": 9.504615384615386e-06,
205
- "loss": 0.3777,
206
- "step": 825
207
- },
208
- {
209
- "epoch": 0.12,
210
- "learning_rate": 9.466153846153846e-06,
211
- "loss": 0.4338,
212
- "step": 850
213
- },
214
- {
215
- "epoch": 0.12,
216
- "learning_rate": 9.427692307692309e-06,
217
- "loss": 0.4017,
218
- "step": 875
219
- },
220
- {
221
- "epoch": 0.13,
222
- "learning_rate": 9.38923076923077e-06,
223
- "loss": 0.3757,
224
- "step": 900
225
- },
226
- {
227
- "epoch": 0.13,
228
- "learning_rate": 9.350769230769231e-06,
229
- "loss": 0.4084,
230
- "step": 925
231
- },
232
- {
233
- "epoch": 0.14,
234
- "learning_rate": 9.312307692307693e-06,
235
- "loss": 0.3821,
236
- "step": 950
237
- },
238
- {
239
- "epoch": 0.14,
240
- "learning_rate": 9.273846153846154e-06,
241
- "loss": 0.421,
242
- "step": 975
243
- },
244
- {
245
- "epoch": 0.14,
246
- "learning_rate": 9.235384615384616e-06,
247
- "loss": 0.4203,
248
- "step": 1000
249
- },
250
- {
251
- "epoch": 0.14,
252
- "eval_loss": 0.4127572178840637,
253
- "eval_runtime": 1802.5775,
254
- "eval_samples_per_second": 3.656,
255
- "eval_steps_per_second": 0.457,
256
- "eval_wer": 28.265624367127064,
257
- "step": 1000
258
- },
259
- {
260
- "epoch": 0.15,
261
- "learning_rate": 9.196923076923078e-06,
262
- "loss": 0.4094,
263
- "step": 1025
264
- },
265
- {
266
- "epoch": 0.15,
267
- "learning_rate": 9.158461538461539e-06,
268
- "loss": 0.3674,
269
- "step": 1050
270
- },
271
- {
272
- "epoch": 0.15,
273
- "learning_rate": 9.12e-06,
274
- "loss": 0.3613,
275
- "step": 1075
276
- },
277
- {
278
- "epoch": 0.16,
279
- "learning_rate": 9.081538461538463e-06,
280
- "loss": 0.3904,
281
- "step": 1100
282
- },
283
- {
284
- "epoch": 0.16,
285
- "learning_rate": 9.043076923076923e-06,
286
- "loss": 0.3865,
287
- "step": 1125
288
- },
289
- {
290
- "epoch": 0.16,
291
- "learning_rate": 9.004615384615386e-06,
292
- "loss": 0.3677,
293
- "step": 1150
294
- },
295
- {
296
- "epoch": 0.17,
297
- "learning_rate": 8.966153846153846e-06,
298
- "loss": 0.3295,
299
- "step": 1175
300
- },
301
- {
302
- "epoch": 0.17,
303
- "learning_rate": 8.927692307692308e-06,
304
- "loss": 0.3207,
305
- "step": 1200
306
- },
307
- {
308
- "epoch": 0.17,
309
- "learning_rate": 8.88923076923077e-06,
310
- "loss": 0.2817,
311
- "step": 1225
312
- },
313
- {
314
- "epoch": 0.18,
315
- "learning_rate": 8.850769230769231e-06,
316
- "loss": 0.3218,
317
- "step": 1250
318
- },
319
- {
320
- "epoch": 0.18,
321
- "learning_rate": 8.812307692307693e-06,
322
- "loss": 0.3502,
323
- "step": 1275
324
- },
325
- {
326
- "epoch": 0.19,
327
- "learning_rate": 8.773846153846155e-06,
328
- "loss": 0.3432,
329
- "step": 1300
330
- },
331
- {
332
- "epoch": 0.19,
333
- "learning_rate": 8.735384615384616e-06,
334
- "loss": 0.3244,
335
- "step": 1325
336
- },
337
- {
338
- "epoch": 0.19,
339
- "learning_rate": 8.696923076923078e-06,
340
- "loss": 0.2607,
341
- "step": 1350
342
- },
343
- {
344
- "epoch": 0.2,
345
- "learning_rate": 8.658461538461538e-06,
346
- "loss": 0.3469,
347
- "step": 1375
348
- },
349
- {
350
- "epoch": 0.2,
351
- "learning_rate": 8.62e-06,
352
- "loss": 0.2756,
353
- "step": 1400
354
- },
355
- {
356
- "epoch": 0.2,
357
- "learning_rate": 8.581538461538463e-06,
358
- "loss": 0.294,
359
- "step": 1425
360
- },
361
- {
362
- "epoch": 0.21,
363
- "learning_rate": 8.543076923076923e-06,
364
- "loss": 0.3915,
365
- "step": 1450
366
- },
367
- {
368
- "epoch": 0.21,
369
- "learning_rate": 8.504615384615385e-06,
370
- "loss": 0.3237,
371
- "step": 1475
372
- },
373
- {
374
- "epoch": 0.21,
375
- "learning_rate": 8.466153846153847e-06,
376
- "loss": 0.2723,
377
- "step": 1500
378
- },
379
- {
380
- "epoch": 0.22,
381
- "learning_rate": 8.427692307692308e-06,
382
- "loss": 0.2917,
383
- "step": 1525
384
- },
385
- {
386
- "epoch": 0.22,
387
- "learning_rate": 8.38923076923077e-06,
388
- "loss": 0.2921,
389
- "step": 1550
390
- },
391
- {
392
- "epoch": 0.23,
393
- "learning_rate": 8.35076923076923e-06,
394
- "loss": 0.3724,
395
- "step": 1575
396
- },
397
- {
398
- "epoch": 0.23,
399
- "learning_rate": 8.312307692307693e-06,
400
- "loss": 0.3364,
401
- "step": 1600
402
- },
403
- {
404
- "epoch": 0.23,
405
- "learning_rate": 8.273846153846155e-06,
406
- "loss": 0.3661,
407
- "step": 1625
408
- },
409
- {
410
- "epoch": 0.24,
411
- "learning_rate": 8.235384615384615e-06,
412
- "loss": 0.3307,
413
- "step": 1650
414
- },
415
- {
416
- "epoch": 0.24,
417
- "learning_rate": 8.196923076923078e-06,
418
- "loss": 0.3201,
419
- "step": 1675
420
- },
421
- {
422
- "epoch": 0.24,
423
- "learning_rate": 8.15846153846154e-06,
424
- "loss": 0.2973,
425
- "step": 1700
426
- },
427
- {
428
- "epoch": 0.25,
429
- "learning_rate": 8.120000000000002e-06,
430
- "loss": 0.3013,
431
- "step": 1725
432
- },
433
- {
434
- "epoch": 0.25,
435
- "learning_rate": 8.081538461538462e-06,
436
- "loss": 0.3349,
437
- "step": 1750
438
- },
439
- {
440
- "epoch": 0.25,
441
- "learning_rate": 8.043076923076923e-06,
442
- "loss": 0.3244,
443
- "step": 1775
444
- },
445
- {
446
- "epoch": 0.26,
447
- "learning_rate": 8.004615384615385e-06,
448
- "loss": 0.3331,
449
- "step": 1800
450
- },
451
- {
452
- "epoch": 0.26,
453
- "learning_rate": 7.966153846153847e-06,
454
- "loss": 0.2723,
455
- "step": 1825
456
- },
457
- {
458
- "epoch": 0.26,
459
- "learning_rate": 7.927692307692308e-06,
460
- "loss": 0.316,
461
- "step": 1850
462
- },
463
- {
464
- "epoch": 0.27,
465
- "learning_rate": 7.88923076923077e-06,
466
- "loss": 0.2648,
467
- "step": 1875
468
- },
469
- {
470
- "epoch": 0.27,
471
- "learning_rate": 7.850769230769232e-06,
472
- "loss": 0.2965,
473
- "step": 1900
474
- },
475
- {
476
- "epoch": 0.28,
477
- "learning_rate": 7.812307692307694e-06,
478
- "loss": 0.256,
479
- "step": 1925
480
- },
481
- {
482
- "epoch": 0.28,
483
- "learning_rate": 7.773846153846155e-06,
484
- "loss": 0.2156,
485
- "step": 1950
486
- },
487
- {
488
- "epoch": 0.28,
489
- "learning_rate": 7.735384615384615e-06,
490
- "loss": 0.2418,
491
- "step": 1975
492
- },
493
- {
494
- "epoch": 0.29,
495
- "learning_rate": 7.696923076923077e-06,
496
- "loss": 0.2693,
497
- "step": 2000
498
- },
499
- {
500
- "epoch": 0.29,
501
- "eval_loss": 0.3239877223968506,
502
- "eval_runtime": 1817.7748,
503
- "eval_samples_per_second": 3.626,
504
- "eval_steps_per_second": 0.453,
505
- "eval_wer": 22.052330997610273,
506
- "step": 2000
507
- },
508
- {
509
- "epoch": 0.29,
510
- "learning_rate": 7.65846153846154e-06,
511
- "loss": 0.2191,
512
- "step": 2025
513
- },
514
- {
515
- "epoch": 0.29,
516
- "learning_rate": 7.620000000000001e-06,
517
- "loss": 0.2437,
518
- "step": 2050
519
- },
520
- {
521
- "epoch": 0.3,
522
- "learning_rate": 7.581538461538462e-06,
523
- "loss": 0.2519,
524
- "step": 2075
525
- },
526
- {
527
- "epoch": 0.3,
528
- "learning_rate": 7.544615384615386e-06,
529
- "loss": 0.3198,
530
- "step": 2100
531
- },
532
- {
533
- "epoch": 0.3,
534
- "learning_rate": 7.5061538461538465e-06,
535
- "loss": 0.2708,
536
- "step": 2125
537
- },
538
- {
539
- "epoch": 0.31,
540
- "learning_rate": 7.467692307692308e-06,
541
- "loss": 0.265,
542
- "step": 2150
543
- },
544
- {
545
- "epoch": 0.31,
546
- "learning_rate": 7.42923076923077e-06,
547
- "loss": 0.2935,
548
- "step": 2175
549
- },
550
- {
551
- "epoch": 0.31,
552
- "learning_rate": 7.390769230769231e-06,
553
- "loss": 0.2595,
554
- "step": 2200
555
- },
556
- {
557
- "epoch": 0.32,
558
- "learning_rate": 7.3523076923076935e-06,
559
- "loss": 0.2613,
560
- "step": 2225
561
- },
562
- {
563
- "epoch": 0.32,
564
- "learning_rate": 7.313846153846155e-06,
565
- "loss": 0.2229,
566
- "step": 2250
567
- },
568
- {
569
- "epoch": 0.33,
570
- "learning_rate": 7.275384615384616e-06,
571
- "loss": 0.2814,
572
- "step": 2275
573
- },
574
- {
575
- "epoch": 0.33,
576
- "learning_rate": 7.236923076923078e-06,
577
- "loss": 0.2681,
578
- "step": 2300
579
- },
580
- {
581
- "epoch": 0.33,
582
- "learning_rate": 7.198461538461539e-06,
583
- "loss": 0.2531,
584
- "step": 2325
585
- },
586
- {
587
- "epoch": 0.34,
588
- "learning_rate": 7.16e-06,
589
- "loss": 0.2104,
590
- "step": 2350
591
- },
592
- {
593
- "epoch": 0.34,
594
- "learning_rate": 7.121538461538462e-06,
595
- "loss": 0.2294,
596
- "step": 2375
597
- },
598
- {
599
- "epoch": 0.34,
600
- "learning_rate": 7.0830769230769235e-06,
601
- "loss": 0.2556,
602
- "step": 2400
603
- },
604
- {
605
- "epoch": 0.35,
606
- "learning_rate": 7.044615384615386e-06,
607
- "loss": 0.2185,
608
- "step": 2425
609
- },
610
- {
611
- "epoch": 0.35,
612
- "learning_rate": 7.006153846153847e-06,
613
- "loss": 0.2585,
614
- "step": 2450
615
- },
616
- {
617
- "epoch": 0.35,
618
- "learning_rate": 6.967692307692308e-06,
619
- "loss": 0.2325,
620
- "step": 2475
621
- },
622
- {
623
- "epoch": 0.36,
624
- "learning_rate": 6.9292307692307705e-06,
625
- "loss": 0.2068,
626
- "step": 2500
627
- },
628
- {
629
- "epoch": 0.36,
630
- "learning_rate": 6.890769230769231e-06,
631
- "loss": 0.2233,
632
- "step": 2525
633
- },
634
- {
635
- "epoch": 0.36,
636
- "learning_rate": 6.852307692307692e-06,
637
- "loss": 0.2649,
638
- "step": 2550
639
- },
640
- {
641
- "epoch": 0.37,
642
- "learning_rate": 6.8138461538461545e-06,
643
- "loss": 0.2524,
644
- "step": 2575
645
- },
646
- {
647
- "epoch": 0.37,
648
- "learning_rate": 6.775384615384616e-06,
649
- "loss": 0.2847,
650
- "step": 2600
651
- },
652
- {
653
- "epoch": 0.38,
654
- "learning_rate": 6.736923076923078e-06,
655
- "loss": 0.1954,
656
- "step": 2625
657
- },
658
- {
659
- "epoch": 0.38,
660
- "learning_rate": 6.698461538461539e-06,
661
- "loss": 0.2553,
662
- "step": 2650
663
- },
664
- {
665
- "epoch": 0.38,
666
- "learning_rate": 6.660000000000001e-06,
667
- "loss": 0.2138,
668
- "step": 2675
669
- },
670
- {
671
- "epoch": 0.39,
672
- "learning_rate": 6.621538461538463e-06,
673
- "loss": 0.2245,
674
- "step": 2700
675
- },
676
- {
677
- "epoch": 0.39,
678
- "learning_rate": 6.583076923076923e-06,
679
- "loss": 0.2137,
680
- "step": 2725
681
- },
682
- {
683
- "epoch": 0.39,
684
- "learning_rate": 6.5446153846153846e-06,
685
- "loss": 0.2558,
686
- "step": 2750
687
- },
688
- {
689
- "epoch": 0.4,
690
- "learning_rate": 6.506153846153847e-06,
691
- "loss": 0.1981,
692
- "step": 2775
693
- },
694
- {
695
- "epoch": 0.4,
696
- "learning_rate": 6.467692307692308e-06,
697
- "loss": 0.2327,
698
- "step": 2800
699
- },
700
- {
701
- "epoch": 0.4,
702
- "learning_rate": 6.42923076923077e-06,
703
- "loss": 0.219,
704
- "step": 2825
705
- },
706
- {
707
- "epoch": 0.41,
708
- "learning_rate": 6.3907692307692315e-06,
709
- "loss": 0.226,
710
- "step": 2850
711
- },
712
- {
713
- "epoch": 0.41,
714
- "learning_rate": 6.352307692307693e-06,
715
- "loss": 0.2192,
716
- "step": 2875
717
- },
718
- {
719
- "epoch": 0.41,
720
- "learning_rate": 6.313846153846155e-06,
721
- "loss": 0.2146,
722
- "step": 2900
723
- },
724
- {
725
- "epoch": 0.42,
726
- "learning_rate": 6.2753846153846155e-06,
727
- "loss": 0.2058,
728
- "step": 2925
729
- },
730
- {
731
- "epoch": 0.42,
732
- "learning_rate": 6.236923076923077e-06,
733
- "loss": 0.2369,
734
- "step": 2950
735
- },
736
- {
737
- "epoch": 0.42,
738
- "learning_rate": 6.198461538461539e-06,
739
- "loss": 0.2202,
740
- "step": 2975
741
- },
742
- {
743
- "epoch": 0.43,
744
- "learning_rate": 6.16e-06,
745
- "loss": 0.2228,
746
- "step": 3000
747
- },
748
- {
749
- "epoch": 0.43,
750
- "eval_loss": 0.27374517917633057,
751
- "eval_runtime": 1805.7724,
752
- "eval_samples_per_second": 3.65,
753
- "eval_steps_per_second": 0.456,
754
- "eval_wer": 18.14370772408765,
755
- "step": 3000
756
- },
757
- {
758
- "epoch": 0.43,
759
- "learning_rate": 6.1215384615384625e-06,
760
- "loss": 0.2029,
761
- "step": 3025
762
- },
763
- {
764
- "epoch": 0.44,
765
- "learning_rate": 6.083076923076924e-06,
766
- "loss": 0.2085,
767
- "step": 3050
768
- },
769
- {
770
- "epoch": 0.44,
771
- "learning_rate": 6.044615384615385e-06,
772
- "loss": 0.196,
773
- "step": 3075
774
- },
775
- {
776
- "epoch": 0.44,
777
- "learning_rate": 6.006153846153847e-06,
778
- "loss": 0.1999,
779
- "step": 3100
780
- },
781
- {
782
- "epoch": 0.45,
783
- "learning_rate": 5.967692307692308e-06,
784
- "loss": 0.2076,
785
- "step": 3125
786
- },
787
- {
788
- "epoch": 0.45,
789
- "learning_rate": 5.929230769230769e-06,
790
- "loss": 0.1999,
791
- "step": 3150
792
- },
793
- {
794
- "epoch": 0.45,
795
- "learning_rate": 5.890769230769231e-06,
796
- "loss": 0.2361,
797
- "step": 3175
798
- },
799
- {
800
- "epoch": 0.46,
801
- "learning_rate": 5.8523076923076926e-06,
802
- "loss": 0.2003,
803
- "step": 3200
804
- },
805
- {
806
- "epoch": 0.46,
807
- "learning_rate": 5.813846153846155e-06,
808
- "loss": 0.249,
809
- "step": 3225
810
- },
811
- {
812
- "epoch": 0.46,
813
- "learning_rate": 5.775384615384616e-06,
814
- "loss": 0.2724,
815
- "step": 3250
816
- },
817
- {
818
- "epoch": 0.47,
819
- "learning_rate": 5.736923076923077e-06,
820
- "loss": 0.2001,
821
- "step": 3275
822
- },
823
- {
824
- "epoch": 1.0,
825
- "learning_rate": 5.6984615384615395e-06,
826
- "loss": 0.2355,
827
- "step": 3300
828
- },
829
- {
830
- "epoch": 1.0,
831
- "learning_rate": 5.66e-06,
832
- "loss": 0.2297,
833
- "step": 3325
834
- },
835
- {
836
- "epoch": 1.01,
837
- "learning_rate": 5.621538461538461e-06,
838
- "loss": 0.2265,
839
- "step": 3350
840
- },
841
- {
842
- "epoch": 1.01,
843
- "learning_rate": 5.5830769230769235e-06,
844
- "loss": 0.21,
845
- "step": 3375
846
- },
847
- {
848
- "epoch": 1.01,
849
- "learning_rate": 5.544615384615385e-06,
850
- "loss": 0.1944,
851
- "step": 3400
852
- },
853
- {
854
- "epoch": 1.02,
855
- "learning_rate": 5.506153846153847e-06,
856
- "loss": 0.1844,
857
- "step": 3425
858
- },
859
- {
860
- "epoch": 1.02,
861
- "learning_rate": 5.467692307692308e-06,
862
- "loss": 0.1774,
863
- "step": 3450
864
- },
865
- {
866
- "epoch": 1.03,
867
- "learning_rate": 5.42923076923077e-06,
868
- "loss": 0.1688,
869
- "step": 3475
870
- },
871
- {
872
- "epoch": 1.03,
873
- "learning_rate": 5.390769230769232e-06,
874
- "loss": 0.1473,
875
- "step": 3500
876
- },
877
- {
878
- "epoch": 1.03,
879
- "learning_rate": 5.352307692307692e-06,
880
- "loss": 0.1743,
881
- "step": 3525
882
- },
883
- {
884
- "epoch": 1.04,
885
- "learning_rate": 5.313846153846154e-06,
886
- "loss": 0.1759,
887
- "step": 3550
888
- },
889
- {
890
- "epoch": 1.04,
891
- "learning_rate": 5.275384615384616e-06,
892
- "loss": 0.1507,
893
- "step": 3575
894
- },
895
- {
896
- "epoch": 1.04,
897
- "learning_rate": 5.236923076923077e-06,
898
- "loss": 0.1556,
899
- "step": 3600
900
- },
901
- {
902
- "epoch": 1.05,
903
- "learning_rate": 5.198461538461539e-06,
904
- "loss": 0.1545,
905
- "step": 3625
906
- },
907
- {
908
- "epoch": 1.05,
909
- "learning_rate": 5.1600000000000006e-06,
910
- "loss": 0.1143,
911
- "step": 3650
912
- },
913
- {
914
- "epoch": 1.05,
915
- "learning_rate": 5.121538461538462e-06,
916
- "loss": 0.1168,
917
- "step": 3675
918
- },
919
- {
920
- "epoch": 1.06,
921
- "learning_rate": 5.083076923076924e-06,
922
- "loss": 0.1095,
923
- "step": 3700
924
- },
925
- {
926
- "epoch": 1.06,
927
- "learning_rate": 5.044615384615385e-06,
928
- "loss": 0.1098,
929
- "step": 3725
930
- },
931
- {
932
- "epoch": 1.06,
933
- "learning_rate": 5.006153846153846e-06,
934
- "loss": 0.102,
935
- "step": 3750
936
- },
937
- {
938
- "epoch": 1.07,
939
- "learning_rate": 4.967692307692308e-06,
940
- "loss": 0.1263,
941
- "step": 3775
942
- },
943
- {
944
- "epoch": 1.07,
945
- "learning_rate": 4.929230769230769e-06,
946
- "loss": 0.093,
947
- "step": 3800
948
- },
949
- {
950
- "epoch": 1.08,
951
- "learning_rate": 4.8907692307692315e-06,
952
- "loss": 0.1163,
953
- "step": 3825
954
- },
955
- {
956
- "epoch": 1.08,
957
- "learning_rate": 4.852307692307693e-06,
958
- "loss": 0.1005,
959
- "step": 3850
960
- },
961
- {
962
- "epoch": 1.08,
963
- "learning_rate": 4.813846153846154e-06,
964
- "loss": 0.1036,
965
- "step": 3875
966
- },
967
- {
968
- "epoch": 1.09,
969
- "learning_rate": 4.7753846153846155e-06,
970
- "loss": 0.1289,
971
- "step": 3900
972
- },
973
- {
974
- "epoch": 1.09,
975
- "learning_rate": 4.736923076923078e-06,
976
- "loss": 0.1238,
977
- "step": 3925
978
- },
979
- {
980
- "epoch": 1.09,
981
- "learning_rate": 4.698461538461539e-06,
982
- "loss": 0.1218,
983
- "step": 3950
984
- },
985
- {
986
- "epoch": 1.1,
987
- "learning_rate": 4.66e-06,
988
- "loss": 0.1223,
989
- "step": 3975
990
- },
991
- {
992
- "epoch": 1.1,
993
- "learning_rate": 4.621538461538462e-06,
994
- "loss": 0.1002,
995
- "step": 4000
996
- },
997
- {
998
- "epoch": 1.1,
999
- "eval_loss": 0.2553846836090088,
1000
- "eval_runtime": 1819.1705,
1001
- "eval_samples_per_second": 3.623,
1002
- "eval_steps_per_second": 0.453,
1003
- "eval_wer": 16.353436753209934,
1004
- "step": 4000
1005
- },
1006
- {
1007
- "epoch": 1.1,
1008
- "learning_rate": 4.583076923076924e-06,
1009
- "loss": 0.1,
1010
- "step": 4025
1011
- },
1012
- {
1013
- "epoch": 1.11,
1014
- "learning_rate": 4.544615384615385e-06,
1015
- "loss": 0.0925,
1016
- "step": 4050
1017
- },
1018
- {
1019
- "epoch": 1.11,
1020
- "learning_rate": 4.506153846153846e-06,
1021
- "loss": 0.1064,
1022
- "step": 4075
1023
- },
1024
- {
1025
- "epoch": 1.11,
1026
- "learning_rate": 4.467692307692308e-06,
1027
- "loss": 0.1127,
1028
- "step": 4100
1029
- },
1030
- {
1031
- "epoch": 1.12,
1032
- "learning_rate": 4.42923076923077e-06,
1033
- "loss": 0.1036,
1034
- "step": 4125
1035
- },
1036
- {
1037
- "epoch": 1.12,
1038
- "learning_rate": 4.392307692307693e-06,
1039
- "loss": 0.091,
1040
- "step": 4150
1041
- },
1042
- {
1043
- "epoch": 1.13,
1044
- "learning_rate": 4.353846153846154e-06,
1045
- "loss": 0.105,
1046
- "step": 4175
1047
- },
1048
- {
1049
- "epoch": 1.13,
1050
- "learning_rate": 4.315384615384616e-06,
1051
- "loss": 0.1147,
1052
- "step": 4200
1053
- },
1054
- {
1055
- "epoch": 1.13,
1056
- "learning_rate": 4.276923076923078e-06,
1057
- "loss": 0.0897,
1058
- "step": 4225
1059
- },
1060
- {
1061
- "epoch": 1.14,
1062
- "learning_rate": 4.238461538461539e-06,
1063
- "loss": 0.1268,
1064
- "step": 4250
1065
- },
1066
- {
1067
- "epoch": 1.14,
1068
- "learning_rate": 4.2000000000000004e-06,
1069
- "loss": 0.1129,
1070
- "step": 4275
1071
- },
1072
- {
1073
- "epoch": 1.14,
1074
- "learning_rate": 4.161538461538462e-06,
1075
- "loss": 0.1041,
1076
- "step": 4300
1077
- },
1078
- {
1079
- "epoch": 1.15,
1080
- "learning_rate": 4.123076923076924e-06,
1081
- "loss": 0.1061,
1082
- "step": 4325
1083
- },
1084
- {
1085
- "epoch": 1.15,
1086
- "learning_rate": 4.084615384615385e-06,
1087
- "loss": 0.1059,
1088
- "step": 4350
1089
- },
1090
- {
1091
- "epoch": 1.15,
1092
- "learning_rate": 4.0461538461538466e-06,
1093
- "loss": 0.0818,
1094
- "step": 4375
1095
- },
1096
- {
1097
- "epoch": 1.16,
1098
- "learning_rate": 4.007692307692308e-06,
1099
- "loss": 0.087,
1100
- "step": 4400
1101
- },
1102
- {
1103
- "epoch": 1.16,
1104
- "learning_rate": 3.96923076923077e-06,
1105
- "loss": 0.0835,
1106
- "step": 4425
1107
- },
1108
- {
1109
- "epoch": 1.16,
1110
- "learning_rate": 3.930769230769231e-06,
1111
- "loss": 0.0713,
1112
- "step": 4450
1113
- },
1114
- {
1115
- "epoch": 1.17,
1116
- "learning_rate": 3.892307692307693e-06,
1117
- "loss": 0.0792,
1118
- "step": 4475
1119
- },
1120
- {
1121
- "epoch": 1.17,
1122
- "learning_rate": 3.853846153846154e-06,
1123
- "loss": 0.0799,
1124
- "step": 4500
1125
- },
1126
- {
1127
- "epoch": 1.18,
1128
- "learning_rate": 3.815384615384616e-06,
1129
- "loss": 0.0755,
1130
- "step": 4525
1131
- },
1132
- {
1133
- "epoch": 1.18,
1134
- "learning_rate": 3.7769230769230775e-06,
1135
- "loss": 0.0703,
1136
- "step": 4550
1137
- },
1138
- {
1139
- "epoch": 1.18,
1140
- "learning_rate": 3.7384615384615384e-06,
1141
- "loss": 0.11,
1142
- "step": 4575
1143
- },
1144
- {
1145
- "epoch": 1.19,
1146
- "learning_rate": 3.7e-06,
1147
- "loss": 0.0891,
1148
- "step": 4600
1149
- },
1150
- {
1151
- "epoch": 1.19,
1152
- "learning_rate": 3.661538461538462e-06,
1153
- "loss": 0.0757,
1154
- "step": 4625
1155
- },
1156
- {
1157
- "epoch": 1.19,
1158
- "learning_rate": 3.6230769230769236e-06,
1159
- "loss": 0.0699,
1160
- "step": 4650
1161
- },
1162
- {
1163
- "epoch": 1.2,
1164
- "learning_rate": 3.5846153846153845e-06,
1165
- "loss": 0.0839,
1166
- "step": 4675
1167
- },
1168
- {
1169
- "epoch": 1.2,
1170
- "learning_rate": 3.5461538461538463e-06,
1171
- "loss": 0.1029,
1172
- "step": 4700
1173
- },
1174
- {
1175
- "epoch": 1.2,
1176
- "learning_rate": 3.507692307692308e-06,
1177
- "loss": 0.0806,
1178
- "step": 4725
1179
- },
1180
- {
1181
- "epoch": 1.21,
1182
- "learning_rate": 3.4692307692307698e-06,
1183
- "loss": 0.0808,
1184
- "step": 4750
1185
- },
1186
- {
1187
- "epoch": 1.21,
1188
- "learning_rate": 3.4307692307692307e-06,
1189
- "loss": 0.0801,
1190
- "step": 4775
1191
- },
1192
- {
1193
- "epoch": 1.21,
1194
- "learning_rate": 3.3923076923076924e-06,
1195
- "loss": 0.0705,
1196
- "step": 4800
1197
- },
1198
- {
1199
- "epoch": 1.22,
1200
- "learning_rate": 3.353846153846154e-06,
1201
- "loss": 0.0711,
1202
- "step": 4825
1203
- },
1204
- {
1205
- "epoch": 1.22,
1206
- "learning_rate": 3.315384615384616e-06,
1207
- "loss": 0.0721,
1208
- "step": 4850
1209
- },
1210
- {
1211
- "epoch": 1.23,
1212
- "learning_rate": 3.276923076923077e-06,
1213
- "loss": 0.0806,
1214
- "step": 4875
1215
- },
1216
- {
1217
- "epoch": 1.23,
1218
- "learning_rate": 3.2384615384615385e-06,
1219
- "loss": 0.0816,
1220
- "step": 4900
1221
- },
1222
- {
1223
- "epoch": 1.23,
1224
- "learning_rate": 3.2000000000000003e-06,
1225
- "loss": 0.078,
1226
- "step": 4925
1227
- },
1228
- {
1229
- "epoch": 1.24,
1230
- "learning_rate": 3.161538461538462e-06,
1231
- "loss": 0.0858,
1232
- "step": 4950
1233
- },
1234
- {
1235
- "epoch": 1.24,
1236
- "learning_rate": 3.123076923076923e-06,
1237
- "loss": 0.0948,
1238
- "step": 4975
1239
- },
1240
- {
1241
- "epoch": 1.24,
1242
- "learning_rate": 3.0846153846153847e-06,
1243
- "loss": 0.0863,
1244
- "step": 5000
1245
- },
1246
- {
1247
- "epoch": 1.24,
1248
- "eval_loss": 0.23512449860572815,
1249
- "eval_runtime": 1865.1318,
1250
- "eval_samples_per_second": 3.534,
1251
- "eval_steps_per_second": 0.442,
1252
- "eval_wer": 14.787962250394912,
1253
- "step": 5000
1254
- }
1255
- ],
1256
- "max_steps": 7000,
1257
- "num_train_epochs": 9223372036854775807,
1258
- "total_flos": 2.041005694058496e+19,
1259
- "trial_name": null,
1260
- "trial_params": null
1261
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
checkpoint-5000/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ca27d9d1bef820c87cfd52d5adb15af5250e1ac4555cd1703a088ad0dad1a4
3
- size 3643
 
 
 
 
nohup.out DELETED
The diff for this file is too large to render. See raw diff
 
run.sh CHANGED
@@ -6,18 +6,18 @@ python run_speech_recognition_seq2seq_streaming.py \
6
  --train_split_name="train+validation" \
7
  --eval_split_name="test" \
8
  --model_index_name="Whisper Small Basque" \
9
- --max_steps="7000" \
10
  --output_dir="./" \
11
  --per_device_train_batch_size="4" \
12
  --per_device_eval_batch_size="8" \
13
  --gradient_accumulation_steps="1" \
14
  --logging_steps="25" \
15
- --learning_rate="1e-5" \
16
  --warmup_steps="500" \
17
  --evaluation_strategy="steps" \
18
  --eval_steps="1000" \
19
  --save_strategy="steps" \
20
- --save_steps="1000" \
21
  --generation_max_length="225" \
22
  --length_column_name="input_length" \
23
  --max_duration_in_seconds="30" \
 
6
  --train_split_name="train+validation" \
7
  --eval_split_name="test" \
8
  --model_index_name="Whisper Small Basque" \
9
+ --max_steps="8000" \
10
  --output_dir="./" \
11
  --per_device_train_batch_size="4" \
12
  --per_device_eval_batch_size="8" \
13
  --gradient_accumulation_steps="1" \
14
  --logging_steps="25" \
15
+ --learning_rate="6e-6" \
16
  --warmup_steps="500" \
17
  --evaluation_strategy="steps" \
18
  --eval_steps="1000" \
19
  --save_strategy="steps" \
20
+ --save_steps="500" \
21
  --generation_max_length="225" \
22
  --length_column_name="input_length" \
23
  --max_duration_in_seconds="30" \