Hugging Face's logo

ArtFair
/
openai-whisper-large-v2-finetune-jacob_filter-resplit

Safetensors
whisper
Model card Files Community
aimonbc24 commited on
Commit
a3bc393
·
verified ·
1 Parent(s): c4db3e9

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. checkpoint-11187/config.json +66 -0
  2. checkpoint-11187/generation_config.json +317 -0
  3. checkpoint-11187/model-00001-of-00002.safetensors +3 -0
  4. checkpoint-11187/model-00002-of-00002.safetensors +3 -0
  5. checkpoint-11187/model.safetensors.index.json +0 -0
  6. checkpoint-11187/optimizer.pt +3 -0
  7. checkpoint-11187/preprocessor_config.json +14 -0
  8. checkpoint-11187/rng_state.pth +3 -0
  9. checkpoint-11187/scheduler.pt +3 -0
  10. checkpoint-11187/trainer_state.json +2115 -0
  11. checkpoint-11187/training_args.bin +3 -0
  12. checkpoint-1243/config.json +66 -0
  13. checkpoint-1243/generation_config.json +317 -0
  14. checkpoint-1243/model-00001-of-00002.safetensors +3 -0
  15. checkpoint-1243/model-00002-of-00002.safetensors +3 -0
  16. checkpoint-1243/model.safetensors.index.json +0 -0
  17. checkpoint-1243/optimizer.pt +3 -0
  18. checkpoint-1243/preprocessor_config.json +14 -0
  19. checkpoint-1243/rng_state.pth +3 -0
  20. checkpoint-1243/scheduler.pt +3 -0
  21. checkpoint-1243/trainer_state.json +249 -0
  22. checkpoint-1243/training_args.bin +3 -0
  23. checkpoint-12430/config.json +66 -0
  24. checkpoint-12430/generation_config.json +317 -0
  25. checkpoint-12430/model-00001-of-00002.safetensors +3 -0
  26. checkpoint-12430/model-00002-of-00002.safetensors +3 -0
  27. checkpoint-12430/model.safetensors.index.json +0 -0
  28. checkpoint-12430/optimizer.pt +3 -0
  29. checkpoint-12430/preprocessor_config.json +14 -0
  30. checkpoint-12430/rng_state.pth +3 -0
  31. checkpoint-12430/scheduler.pt +3 -0
  32. checkpoint-12430/trainer_state.json +2349 -0
  33. checkpoint-12430/training_args.bin +3 -0
  34. checkpoint-2486/config.json +66 -0
  35. checkpoint-2486/generation_config.json +317 -0
  36. checkpoint-2486/model-00001-of-00002.safetensors +3 -0
  37. checkpoint-2486/model-00002-of-00002.safetensors +3 -0
  38. checkpoint-2486/model.safetensors.index.json +0 -0
  39. checkpoint-2486/optimizer.pt +3 -0
  40. checkpoint-2486/preprocessor_config.json +14 -0
  41. checkpoint-2486/rng_state.pth +3 -0
  42. checkpoint-2486/scheduler.pt +3 -0
  43. checkpoint-2486/trainer_state.json +483 -0
  44. checkpoint-2486/training_args.bin +3 -0
  45. checkpoint-3729/config.json +66 -0
  46. checkpoint-3729/generation_config.json +317 -0
  47. checkpoint-3729/model-00001-of-00002.safetensors +3 -0
  48. checkpoint-3729/model-00002-of-00002.safetensors +3 -0
  49. checkpoint-3729/model.safetensors.index.json +0 -0
  50. checkpoint-3729/optimizer.pt +3 -0
checkpoint-11187/config.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-large-v2",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "language": "English",
45
+ "mask_feature_length": 10,
46
+ "mask_feature_min_masks": 0,
47
+ "mask_feature_prob": 0.0,
48
+ "mask_time_length": 10,
49
+ "mask_time_min_masks": 2,
50
+ "mask_time_prob": 0.05,
51
+ "max_length": 448,
52
+ "max_source_positions": 1500,
53
+ "max_target_positions": 448,
54
+ "median_filter_width": 7,
55
+ "model_type": "whisper",
56
+ "num_hidden_layers": 32,
57
+ "num_mel_bins": 80,
58
+ "pad_token_id": 50257,
59
+ "scale_embedding": false,
60
+ "suppress_tokens": [],
61
+ "torch_dtype": "float32",
62
+ "transformers_version": "4.37.2",
63
+ "use_cache": true,
64
+ "use_weighted_layer_sum": false,
65
+ "vocab_size": 51865
66
+ }
checkpoint-11187/generation_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "begin_suppress_tokens": [
97
+ 220,
98
+ 50257
99
+ ],
100
+ "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
+ "forced_decoder_ids": [
104
+ [
105
+ 1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
+ ]
112
+ ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
+ "max_initial_timestamp_index": 50,
216
+ "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
+ "return_timestamps": false,
221
+ "suppress_tokens": [
222
+ 1,
223
+ 2,
224
+ 7,
225
+ 8,
226
+ 9,
227
+ 10,
228
+ 14,
229
+ 25,
230
+ 26,
231
+ 27,
232
+ 28,
233
+ 29,
234
+ 31,
235
+ 58,
236
+ 59,
237
+ 60,
238
+ 61,
239
+ 62,
240
+ 63,
241
+ 90,
242
+ 91,
243
+ 92,
244
+ 93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
305
+ 50358,
306
+ 50359,
307
+ 50360,
308
+ 50361,
309
+ 50362
310
+ ],
311
+ "task": "transcribe",
312
+ "task_to_id": {
313
+ "transcribe": 50359,
314
+ "translate": 50358
315
+ },
316
+ "transformers_version": "4.37.2"
317
+ }
checkpoint-11187/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a70e8357cbb3608475754c5d4356a63b69f760776340354f8406c72381c56af
3
+ size 4992706480
checkpoint-11187/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7af621993d6d1098dc2bc5fc088640cc33ee07cf43c8901a11360ba66b032480
3
+ size 1180663192
checkpoint-11187/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-11187/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1dae8d4b1de57b0280be379c2183fb590adc244ec78ba6d0499fc2cd80e013bd
3
+ size 3095074288
checkpoint-11187/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-11187/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47c1fe439e2ed06ec288514fea1b232061abe40004525cb256570d161743fcdb
3
+ size 14244
checkpoint-11187/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ddeaf5db1f24a634b0837bc7a1112c08ef16c1a47b9d0e6b9e76f836bec6e49a
3
+ size 1064
checkpoint-11187/trainer_state.json ADDED
@@ -0,0 +1,2115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.8992765273311897,
5
+ "eval_steps": 500,
6
+ "global_step": 11187,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 3.753351206434316e-08,
14
+ "loss": 4.9531,
15
+ "step": 32
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 8.042895442359249e-08,
20
+ "loss": 4.9011,
21
+ "step": 64
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 1.2198391420911528e-07,
26
+ "loss": 4.472,
27
+ "step": 96
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 1.648793565683646e-07,
32
+ "loss": 3.9678,
33
+ "step": 128
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 2.0777479892761392e-07,
38
+ "loss": 3.3412,
39
+ "step": 160
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 2.5067024128686325e-07,
44
+ "loss": 2.5657,
45
+ "step": 192
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 2.9356568364611256e-07,
50
+ "loss": 1.9677,
51
+ "step": 224
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 3.364611260053619e-07,
56
+ "loss": 1.6646,
57
+ "step": 256
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 3.7935656836461123e-07,
62
+ "loss": 1.5032,
63
+ "step": 288
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 4.222520107238606e-07,
68
+ "loss": 1.3443,
69
+ "step": 320
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 4.651474530831099e-07,
74
+ "loss": 1.2277,
75
+ "step": 352
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 5.080428954423593e-07,
80
+ "loss": 1.0104,
81
+ "step": 384
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 5.509383378016086e-07,
86
+ "loss": 0.7317,
87
+ "step": 416
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 5.938337801608579e-07,
92
+ "loss": 0.3644,
93
+ "step": 448
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 6.367292225201072e-07,
98
+ "loss": 0.388,
99
+ "step": 480
100
+ },
101
+ {
102
+ "epoch": 0.04,
103
+ "learning_rate": 6.796246648793566e-07,
104
+ "loss": 0.264,
105
+ "step": 512
106
+ },
107
+ {
108
+ "epoch": 0.04,
109
+ "learning_rate": 7.225201072386059e-07,
110
+ "loss": 0.2479,
111
+ "step": 544
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 7.654155495978551e-07,
116
+ "loss": 0.403,
117
+ "step": 576
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 8.083109919571045e-07,
122
+ "loss": 0.3062,
123
+ "step": 608
124
+ },
125
+ {
126
+ "epoch": 0.05,
127
+ "learning_rate": 8.512064343163538e-07,
128
+ "loss": 0.3016,
129
+ "step": 640
130
+ },
131
+ {
132
+ "epoch": 0.05,
133
+ "learning_rate": 8.941018766756032e-07,
134
+ "loss": 0.3445,
135
+ "step": 672
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.369973190348524e-07,
140
+ "loss": 0.2554,
141
+ "step": 704
142
+ },
143
+ {
144
+ "epoch": 0.06,
145
+ "learning_rate": 9.798927613941018e-07,
146
+ "loss": 0.2762,
147
+ "step": 736
148
+ },
149
+ {
150
+ "epoch": 0.06,
151
+ "learning_rate": 9.985462630408755e-07,
152
+ "loss": 0.2265,
153
+ "step": 768
154
+ },
155
+ {
156
+ "epoch": 0.06,
157
+ "learning_rate": 9.958098170001709e-07,
158
+ "loss": 0.2158,
159
+ "step": 800
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.930733709594662e-07,
164
+ "loss": 0.3106,
165
+ "step": 832
166
+ },
167
+ {
168
+ "epoch": 0.07,
169
+ "learning_rate": 9.903369249187618e-07,
170
+ "loss": 0.3085,
171
+ "step": 864
172
+ },
173
+ {
174
+ "epoch": 0.07,
175
+ "learning_rate": 9.876004788780571e-07,
176
+ "loss": 0.2633,
177
+ "step": 896
178
+ },
179
+ {
180
+ "epoch": 0.07,
181
+ "learning_rate": 9.848640328373525e-07,
182
+ "loss": 0.2145,
183
+ "step": 928
184
+ },
185
+ {
186
+ "epoch": 0.08,
187
+ "learning_rate": 9.821275867966478e-07,
188
+ "loss": 0.2594,
189
+ "step": 960
190
+ },
191
+ {
192
+ "epoch": 0.08,
193
+ "learning_rate": 9.793911407559432e-07,
194
+ "loss": 0.2264,
195
+ "step": 992
196
+ },
197
+ {
198
+ "epoch": 0.08,
199
+ "learning_rate": 9.766546947152385e-07,
200
+ "loss": 0.2512,
201
+ "step": 1024
202
+ },
203
+ {
204
+ "epoch": 0.08,
205
+ "learning_rate": 9.739182486745339e-07,
206
+ "loss": 0.3154,
207
+ "step": 1056
208
+ },
209
+ {
210
+ "epoch": 0.09,
211
+ "learning_rate": 9.711818026338292e-07,
212
+ "loss": 0.2672,
213
+ "step": 1088
214
+ },
215
+ {
216
+ "epoch": 0.09,
217
+ "learning_rate": 9.684453565931246e-07,
218
+ "loss": 0.2502,
219
+ "step": 1120
220
+ },
221
+ {
222
+ "epoch": 0.09,
223
+ "learning_rate": 9.6570891055242e-07,
224
+ "loss": 0.2702,
225
+ "step": 1152
226
+ },
227
+ {
228
+ "epoch": 0.1,
229
+ "learning_rate": 9.629724645117153e-07,
230
+ "loss": 0.2919,
231
+ "step": 1184
232
+ },
233
+ {
234
+ "epoch": 0.1,
235
+ "learning_rate": 9.602360184710108e-07,
236
+ "loss": 0.3925,
237
+ "step": 1216
238
+ },
239
+ {
240
+ "epoch": 0.1,
241
+ "learning_rate": 9.574995724303062e-07,
242
+ "loss": 0.285,
243
+ "step": 1248
244
+ },
245
+ {
246
+ "epoch": 0.1,
247
+ "learning_rate": 9.547631263896015e-07,
248
+ "loss": 0.3084,
249
+ "step": 1280
250
+ },
251
+ {
252
+ "epoch": 0.11,
253
+ "learning_rate": 9.520266803488969e-07,
254
+ "loss": 0.2275,
255
+ "step": 1312
256
+ },
257
+ {
258
+ "epoch": 0.11,
259
+ "learning_rate": 9.492902343081922e-07,
260
+ "loss": 0.245,
261
+ "step": 1344
262
+ },
263
+ {
264
+ "epoch": 0.11,
265
+ "learning_rate": 9.465537882674875e-07,
266
+ "loss": 0.233,
267
+ "step": 1376
268
+ },
269
+ {
270
+ "epoch": 0.11,
271
+ "learning_rate": 9.438173422267829e-07,
272
+ "loss": 0.2825,
273
+ "step": 1408
274
+ },
275
+ {
276
+ "epoch": 0.12,
277
+ "learning_rate": 9.410808961860783e-07,
278
+ "loss": 0.231,
279
+ "step": 1440
280
+ },
281
+ {
282
+ "epoch": 0.12,
283
+ "learning_rate": 9.383444501453737e-07,
284
+ "loss": 0.2449,
285
+ "step": 1472
286
+ },
287
+ {
288
+ "epoch": 0.12,
289
+ "learning_rate": 9.35608004104669e-07,
290
+ "loss": 0.2732,
291
+ "step": 1504
292
+ },
293
+ {
294
+ "epoch": 0.12,
295
+ "learning_rate": 9.328715580639644e-07,
296
+ "loss": 0.2031,
297
+ "step": 1536
298
+ },
299
+ {
300
+ "epoch": 0.13,
301
+ "learning_rate": 9.301351120232597e-07,
302
+ "loss": 0.1749,
303
+ "step": 1568
304
+ },
305
+ {
306
+ "epoch": 0.13,
307
+ "learning_rate": 9.273986659825551e-07,
308
+ "loss": 0.1722,
309
+ "step": 1600
310
+ },
311
+ {
312
+ "epoch": 0.13,
313
+ "learning_rate": 9.246622199418504e-07,
314
+ "loss": 0.2743,
315
+ "step": 1632
316
+ },
317
+ {
318
+ "epoch": 0.13,
319
+ "learning_rate": 9.219257739011459e-07,
320
+ "loss": 0.2907,
321
+ "step": 1664
322
+ },
323
+ {
324
+ "epoch": 0.14,
325
+ "learning_rate": 9.192748417992132e-07,
326
+ "loss": 0.2664,
327
+ "step": 1696
328
+ },
329
+ {
330
+ "epoch": 0.14,
331
+ "learning_rate": 9.165383957585086e-07,
332
+ "loss": 0.2085,
333
+ "step": 1728
334
+ },
335
+ {
336
+ "epoch": 0.14,
337
+ "learning_rate": 9.13801949717804e-07,
338
+ "loss": 0.1839,
339
+ "step": 1760
340
+ },
341
+ {
342
+ "epoch": 0.14,
343
+ "learning_rate": 9.110655036770994e-07,
344
+ "loss": 0.2667,
345
+ "step": 1792
346
+ },
347
+ {
348
+ "epoch": 0.15,
349
+ "learning_rate": 9.083290576363947e-07,
350
+ "loss": 0.1994,
351
+ "step": 1824
352
+ },
353
+ {
354
+ "epoch": 0.15,
355
+ "learning_rate": 9.0559261159569e-07,
356
+ "loss": 0.2568,
357
+ "step": 1856
358
+ },
359
+ {
360
+ "epoch": 0.15,
361
+ "learning_rate": 9.028561655549855e-07,
362
+ "loss": 0.2909,
363
+ "step": 1888
364
+ },
365
+ {
366
+ "epoch": 0.15,
367
+ "learning_rate": 9.001197195142807e-07,
368
+ "loss": 0.2697,
369
+ "step": 1920
370
+ },
371
+ {
372
+ "epoch": 0.16,
373
+ "learning_rate": 8.973832734735761e-07,
374
+ "loss": 0.3379,
375
+ "step": 1952
376
+ },
377
+ {
378
+ "epoch": 0.16,
379
+ "learning_rate": 8.946468274328715e-07,
380
+ "loss": 0.2866,
381
+ "step": 1984
382
+ },
383
+ {
384
+ "epoch": 0.16,
385
+ "learning_rate": 8.919103813921669e-07,
386
+ "loss": 0.2634,
387
+ "step": 2016
388
+ },
389
+ {
390
+ "epoch": 0.16,
391
+ "learning_rate": 8.891739353514622e-07,
392
+ "loss": 0.2234,
393
+ "step": 2048
394
+ },
395
+ {
396
+ "epoch": 0.17,
397
+ "learning_rate": 8.864374893107576e-07,
398
+ "loss": 0.2541,
399
+ "step": 2080
400
+ },
401
+ {
402
+ "epoch": 0.17,
403
+ "learning_rate": 8.83701043270053e-07,
404
+ "loss": 0.2341,
405
+ "step": 2112
406
+ },
407
+ {
408
+ "epoch": 0.17,
409
+ "learning_rate": 8.809645972293484e-07,
410
+ "loss": 0.2602,
411
+ "step": 2144
412
+ },
413
+ {
414
+ "epoch": 0.17,
415
+ "learning_rate": 8.782281511886437e-07,
416
+ "loss": 0.2602,
417
+ "step": 2176
418
+ },
419
+ {
420
+ "epoch": 0.18,
421
+ "learning_rate": 8.75491705147939e-07,
422
+ "loss": 0.2036,
423
+ "step": 2208
424
+ },
425
+ {
426
+ "epoch": 0.18,
427
+ "learning_rate": 8.727552591072344e-07,
428
+ "loss": 0.2342,
429
+ "step": 2240
430
+ },
431
+ {
432
+ "epoch": 0.18,
433
+ "learning_rate": 8.700188130665298e-07,
434
+ "loss": 0.2361,
435
+ "step": 2272
436
+ },
437
+ {
438
+ "epoch": 0.19,
439
+ "learning_rate": 8.672823670258251e-07,
440
+ "loss": 0.3299,
441
+ "step": 2304
442
+ },
443
+ {
444
+ "epoch": 0.19,
445
+ "learning_rate": 8.645459209851206e-07,
446
+ "loss": 0.3221,
447
+ "step": 2336
448
+ },
449
+ {
450
+ "epoch": 0.19,
451
+ "learning_rate": 8.618094749444159e-07,
452
+ "loss": 0.2119,
453
+ "step": 2368
454
+ },
455
+ {
456
+ "epoch": 0.19,
457
+ "learning_rate": 8.590730289037113e-07,
458
+ "loss": 0.1908,
459
+ "step": 2400
460
+ },
461
+ {
462
+ "epoch": 0.2,
463
+ "learning_rate": 8.563365828630066e-07,
464
+ "loss": 0.2736,
465
+ "step": 2432
466
+ },
467
+ {
468
+ "epoch": 0.2,
469
+ "learning_rate": 8.536001368223021e-07,
470
+ "loss": 0.1713,
471
+ "step": 2464
472
+ },
473
+ {
474
+ "epoch": 0.2,
475
+ "learning_rate": 8.508636907815974e-07,
476
+ "loss": 0.2658,
477
+ "step": 2496
478
+ },
479
+ {
480
+ "epoch": 0.2,
481
+ "learning_rate": 8.481272447408928e-07,
482
+ "loss": 0.2235,
483
+ "step": 2528
484
+ },
485
+ {
486
+ "epoch": 0.21,
487
+ "learning_rate": 8.453907987001881e-07,
488
+ "loss": 0.1858,
489
+ "step": 2560
490
+ },
491
+ {
492
+ "epoch": 0.21,
493
+ "learning_rate": 8.426543526594834e-07,
494
+ "loss": 0.2935,
495
+ "step": 2592
496
+ },
497
+ {
498
+ "epoch": 0.21,
499
+ "learning_rate": 8.399179066187788e-07,
500
+ "loss": 0.1996,
501
+ "step": 2624
502
+ },
503
+ {
504
+ "epoch": 0.21,
505
+ "learning_rate": 8.371814605780741e-07,
506
+ "loss": 0.2209,
507
+ "step": 2656
508
+ },
509
+ {
510
+ "epoch": 0.22,
511
+ "learning_rate": 8.344450145373696e-07,
512
+ "loss": 0.1611,
513
+ "step": 2688
514
+ },
515
+ {
516
+ "epoch": 0.22,
517
+ "learning_rate": 8.317085684966649e-07,
518
+ "loss": 0.28,
519
+ "step": 2720
520
+ },
521
+ {
522
+ "epoch": 0.22,
523
+ "learning_rate": 8.289721224559603e-07,
524
+ "loss": 0.2486,
525
+ "step": 2752
526
+ },
527
+ {
528
+ "epoch": 0.22,
529
+ "learning_rate": 8.262356764152556e-07,
530
+ "loss": 0.1978,
531
+ "step": 2784
532
+ },
533
+ {
534
+ "epoch": 0.23,
535
+ "learning_rate": 8.234992303745511e-07,
536
+ "loss": 0.2535,
537
+ "step": 2816
538
+ },
539
+ {
540
+ "epoch": 0.23,
541
+ "learning_rate": 8.207627843338464e-07,
542
+ "loss": 0.2666,
543
+ "step": 2848
544
+ },
545
+ {
546
+ "epoch": 0.23,
547
+ "learning_rate": 8.180263382931417e-07,
548
+ "loss": 0.1769,
549
+ "step": 2880
550
+ },
551
+ {
552
+ "epoch": 0.23,
553
+ "learning_rate": 8.152898922524371e-07,
554
+ "loss": 0.2803,
555
+ "step": 2912
556
+ },
557
+ {
558
+ "epoch": 0.24,
559
+ "learning_rate": 8.125534462117325e-07,
560
+ "loss": 0.2129,
561
+ "step": 2944
562
+ },
563
+ {
564
+ "epoch": 0.24,
565
+ "learning_rate": 8.098170001710278e-07,
566
+ "loss": 0.2255,
567
+ "step": 2976
568
+ },
569
+ {
570
+ "epoch": 0.24,
571
+ "learning_rate": 8.070805541303232e-07,
572
+ "loss": 0.1739,
573
+ "step": 3008
574
+ },
575
+ {
576
+ "epoch": 0.24,
577
+ "learning_rate": 8.043441080896186e-07,
578
+ "loss": 0.2321,
579
+ "step": 3040
580
+ },
581
+ {
582
+ "epoch": 0.25,
583
+ "learning_rate": 8.01607662048914e-07,
584
+ "loss": 0.2761,
585
+ "step": 3072
586
+ },
587
+ {
588
+ "epoch": 0.25,
589
+ "learning_rate": 7.988712160082093e-07,
590
+ "loss": 0.2867,
591
+ "step": 3104
592
+ },
593
+ {
594
+ "epoch": 0.25,
595
+ "learning_rate": 7.961347699675047e-07,
596
+ "loss": 0.1763,
597
+ "step": 3136
598
+ },
599
+ {
600
+ "epoch": 0.25,
601
+ "learning_rate": 7.933983239268001e-07,
602
+ "loss": 0.325,
603
+ "step": 3168
604
+ },
605
+ {
606
+ "epoch": 0.26,
607
+ "learning_rate": 7.906618778860953e-07,
608
+ "loss": 0.2515,
609
+ "step": 3200
610
+ },
611
+ {
612
+ "epoch": 0.26,
613
+ "learning_rate": 7.879254318453907e-07,
614
+ "loss": 0.1741,
615
+ "step": 3232
616
+ },
617
+ {
618
+ "epoch": 0.26,
619
+ "learning_rate": 7.851889858046861e-07,
620
+ "loss": 0.1999,
621
+ "step": 3264
622
+ },
623
+ {
624
+ "epoch": 0.26,
625
+ "learning_rate": 7.824525397639815e-07,
626
+ "loss": 0.2393,
627
+ "step": 3296
628
+ },
629
+ {
630
+ "epoch": 0.27,
631
+ "learning_rate": 7.797160937232768e-07,
632
+ "loss": 0.2242,
633
+ "step": 3328
634
+ },
635
+ {
636
+ "epoch": 0.27,
637
+ "learning_rate": 7.769796476825722e-07,
638
+ "loss": 0.1877,
639
+ "step": 3360
640
+ },
641
+ {
642
+ "epoch": 0.27,
643
+ "learning_rate": 7.742432016418676e-07,
644
+ "loss": 0.194,
645
+ "step": 3392
646
+ },
647
+ {
648
+ "epoch": 0.28,
649
+ "learning_rate": 7.71506755601163e-07,
650
+ "loss": 0.2499,
651
+ "step": 3424
652
+ },
653
+ {
654
+ "epoch": 0.28,
655
+ "learning_rate": 7.687703095604583e-07,
656
+ "loss": 0.2496,
657
+ "step": 3456
658
+ },
659
+ {
660
+ "epoch": 0.28,
661
+ "learning_rate": 7.660338635197537e-07,
662
+ "loss": 0.1899,
663
+ "step": 3488
664
+ },
665
+ {
666
+ "epoch": 0.28,
667
+ "learning_rate": 7.63297417479049e-07,
668
+ "loss": 0.1866,
669
+ "step": 3520
670
+ },
671
+ {
672
+ "epoch": 0.29,
673
+ "learning_rate": 7.605609714383444e-07,
674
+ "loss": 0.1843,
675
+ "step": 3552
676
+ },
677
+ {
678
+ "epoch": 0.29,
679
+ "learning_rate": 7.578245253976397e-07,
680
+ "loss": 0.1991,
681
+ "step": 3584
682
+ },
683
+ {
684
+ "epoch": 0.29,
685
+ "learning_rate": 7.550880793569352e-07,
686
+ "loss": 0.2122,
687
+ "step": 3616
688
+ },
689
+ {
690
+ "epoch": 0.29,
691
+ "learning_rate": 7.523516333162305e-07,
692
+ "loss": 0.2423,
693
+ "step": 3648
694
+ },
695
+ {
696
+ "epoch": 0.3,
697
+ "learning_rate": 7.496151872755259e-07,
698
+ "loss": 0.2568,
699
+ "step": 3680
700
+ },
701
+ {
702
+ "epoch": 0.3,
703
+ "learning_rate": 7.468787412348212e-07,
704
+ "loss": 0.2727,
705
+ "step": 3712
706
+ },
707
+ {
708
+ "epoch": 0.3,
709
+ "learning_rate": 7.441422951941167e-07,
710
+ "loss": 0.1825,
711
+ "step": 3744
712
+ },
713
+ {
714
+ "epoch": 0.3,
715
+ "learning_rate": 7.41405849153412e-07,
716
+ "loss": 0.1573,
717
+ "step": 3776
718
+ },
719
+ {
720
+ "epoch": 0.31,
721
+ "learning_rate": 7.386694031127074e-07,
722
+ "loss": 0.2034,
723
+ "step": 3808
724
+ },
725
+ {
726
+ "epoch": 0.31,
727
+ "learning_rate": 7.359329570720028e-07,
728
+ "loss": 0.1514,
729
+ "step": 3840
730
+ },
731
+ {
732
+ "epoch": 0.31,
733
+ "learning_rate": 7.33196511031298e-07,
734
+ "loss": 0.2618,
735
+ "step": 3872
736
+ },
737
+ {
738
+ "epoch": 0.31,
739
+ "learning_rate": 7.304600649905934e-07,
740
+ "loss": 0.244,
741
+ "step": 3904
742
+ },
743
+ {
744
+ "epoch": 0.32,
745
+ "learning_rate": 7.277236189498887e-07,
746
+ "loss": 0.1753,
747
+ "step": 3936
748
+ },
749
+ {
750
+ "epoch": 0.32,
751
+ "learning_rate": 7.249871729091842e-07,
752
+ "loss": 0.2044,
753
+ "step": 3968
754
+ },
755
+ {
756
+ "epoch": 0.32,
757
+ "learning_rate": 7.222507268684795e-07,
758
+ "loss": 0.1882,
759
+ "step": 4000
760
+ },
761
+ {
762
+ "epoch": 0.32,
763
+ "learning_rate": 7.195142808277749e-07,
764
+ "loss": 0.2397,
765
+ "step": 4032
766
+ },
767
+ {
768
+ "epoch": 0.33,
769
+ "learning_rate": 7.167778347870702e-07,
770
+ "loss": 0.2084,
771
+ "step": 4064
772
+ },
773
+ {
774
+ "epoch": 0.33,
775
+ "learning_rate": 7.140413887463657e-07,
776
+ "loss": 0.2635,
777
+ "step": 4096
778
+ },
779
+ {
780
+ "epoch": 0.33,
781
+ "learning_rate": 7.11304942705661e-07,
782
+ "loss": 0.2512,
783
+ "step": 4128
784
+ },
785
+ {
786
+ "epoch": 0.33,
787
+ "learning_rate": 7.085684966649563e-07,
788
+ "loss": 0.2411,
789
+ "step": 4160
790
+ },
791
+ {
792
+ "epoch": 0.34,
793
+ "learning_rate": 7.058320506242517e-07,
794
+ "loss": 0.1846,
795
+ "step": 4192
796
+ },
797
+ {
798
+ "epoch": 0.34,
799
+ "learning_rate": 7.030956045835471e-07,
800
+ "loss": 0.1447,
801
+ "step": 4224
802
+ },
803
+ {
804
+ "epoch": 0.34,
805
+ "learning_rate": 7.003591585428424e-07,
806
+ "loss": 0.2373,
807
+ "step": 4256
808
+ },
809
+ {
810
+ "epoch": 0.34,
811
+ "learning_rate": 6.976227125021378e-07,
812
+ "loss": 0.2097,
813
+ "step": 4288
814
+ },
815
+ {
816
+ "epoch": 0.35,
817
+ "learning_rate": 6.948862664614332e-07,
818
+ "loss": 0.2756,
819
+ "step": 4320
820
+ },
821
+ {
822
+ "epoch": 0.35,
823
+ "learning_rate": 6.922353343595006e-07,
824
+ "loss": 0.191,
825
+ "step": 4352
826
+ },
827
+ {
828
+ "epoch": 0.35,
829
+ "learning_rate": 6.894988883187959e-07,
830
+ "loss": 0.2076,
831
+ "step": 4384
832
+ },
833
+ {
834
+ "epoch": 0.35,
835
+ "learning_rate": 6.867624422780914e-07,
836
+ "loss": 0.2787,
837
+ "step": 4416
838
+ },
839
+ {
840
+ "epoch": 0.36,
841
+ "learning_rate": 6.840259962373867e-07,
842
+ "loss": 0.1894,
843
+ "step": 4448
844
+ },
845
+ {
846
+ "epoch": 0.36,
847
+ "learning_rate": 6.81289550196682e-07,
848
+ "loss": 0.1423,
849
+ "step": 4480
850
+ },
851
+ {
852
+ "epoch": 0.36,
853
+ "learning_rate": 6.785531041559774e-07,
854
+ "loss": 0.1738,
855
+ "step": 4512
856
+ },
857
+ {
858
+ "epoch": 0.37,
859
+ "learning_rate": 6.758166581152727e-07,
860
+ "loss": 0.2598,
861
+ "step": 4544
862
+ },
863
+ {
864
+ "epoch": 0.37,
865
+ "learning_rate": 6.730802120745681e-07,
866
+ "loss": 0.2753,
867
+ "step": 4576
868
+ },
869
+ {
870
+ "epoch": 0.37,
871
+ "learning_rate": 6.703437660338634e-07,
872
+ "loss": 0.2922,
873
+ "step": 4608
874
+ },
875
+ {
876
+ "epoch": 0.37,
877
+ "learning_rate": 6.676073199931589e-07,
878
+ "loss": 0.172,
879
+ "step": 4640
880
+ },
881
+ {
882
+ "epoch": 0.38,
883
+ "learning_rate": 6.648708739524542e-07,
884
+ "loss": 0.2269,
885
+ "step": 4672
886
+ },
887
+ {
888
+ "epoch": 0.38,
889
+ "learning_rate": 6.621344279117496e-07,
890
+ "loss": 0.2662,
891
+ "step": 4704
892
+ },
893
+ {
894
+ "epoch": 0.38,
895
+ "learning_rate": 6.59397981871045e-07,
896
+ "loss": 0.2674,
897
+ "step": 4736
898
+ },
899
+ {
900
+ "epoch": 0.38,
901
+ "learning_rate": 6.566615358303404e-07,
902
+ "loss": 0.2803,
903
+ "step": 4768
904
+ },
905
+ {
906
+ "epoch": 0.39,
907
+ "learning_rate": 6.539250897896357e-07,
908
+ "loss": 0.2253,
909
+ "step": 4800
910
+ },
911
+ {
912
+ "epoch": 0.39,
913
+ "learning_rate": 6.51188643748931e-07,
914
+ "loss": 0.2816,
915
+ "step": 4832
916
+ },
917
+ {
918
+ "epoch": 0.39,
919
+ "learning_rate": 6.484521977082264e-07,
920
+ "loss": 0.1596,
921
+ "step": 4864
922
+ },
923
+ {
924
+ "epoch": 0.39,
925
+ "learning_rate": 6.457157516675218e-07,
926
+ "loss": 0.2419,
927
+ "step": 4896
928
+ },
929
+ {
930
+ "epoch": 0.4,
931
+ "learning_rate": 6.429793056268171e-07,
932
+ "loss": 0.2344,
933
+ "step": 4928
934
+ },
935
+ {
936
+ "epoch": 0.4,
937
+ "learning_rate": 6.402428595861125e-07,
938
+ "loss": 0.2008,
939
+ "step": 4960
940
+ },
941
+ {
942
+ "epoch": 0.4,
943
+ "learning_rate": 6.375064135454079e-07,
944
+ "loss": 0.2291,
945
+ "step": 4992
946
+ },
947
+ {
948
+ "epoch": 0.4,
949
+ "learning_rate": 6.347699675047033e-07,
950
+ "loss": 0.1661,
951
+ "step": 5024
952
+ },
953
+ {
954
+ "epoch": 0.41,
955
+ "learning_rate": 6.320335214639986e-07,
956
+ "loss": 0.2022,
957
+ "step": 5056
958
+ },
959
+ {
960
+ "epoch": 0.41,
961
+ "learning_rate": 6.292970754232941e-07,
962
+ "loss": 0.2047,
963
+ "step": 5088
964
+ },
965
+ {
966
+ "epoch": 0.41,
967
+ "learning_rate": 6.265606293825894e-07,
968
+ "loss": 0.2004,
969
+ "step": 5120
970
+ },
971
+ {
972
+ "epoch": 0.41,
973
+ "learning_rate": 6.238241833418846e-07,
974
+ "loss": 0.1987,
975
+ "step": 5152
976
+ },
977
+ {
978
+ "epoch": 0.42,
979
+ "learning_rate": 6.2108773730118e-07,
980
+ "loss": 0.2315,
981
+ "step": 5184
982
+ },
983
+ {
984
+ "epoch": 0.42,
985
+ "learning_rate": 6.183512912604754e-07,
986
+ "loss": 0.2086,
987
+ "step": 5216
988
+ },
989
+ {
990
+ "epoch": 0.42,
991
+ "learning_rate": 6.156148452197708e-07,
992
+ "loss": 0.1717,
993
+ "step": 5248
994
+ },
995
+ {
996
+ "epoch": 0.42,
997
+ "learning_rate": 6.128783991790661e-07,
998
+ "loss": 0.2717,
999
+ "step": 5280
1000
+ },
1001
+ {
1002
+ "epoch": 0.43,
1003
+ "learning_rate": 6.101419531383615e-07,
1004
+ "loss": 0.2229,
1005
+ "step": 5312
1006
+ },
1007
+ {
1008
+ "epoch": 0.43,
1009
+ "learning_rate": 6.074055070976569e-07,
1010
+ "loss": 0.1996,
1011
+ "step": 5344
1012
+ },
1013
+ {
1014
+ "epoch": 0.43,
1015
+ "learning_rate": 6.046690610569523e-07,
1016
+ "loss": 0.2549,
1017
+ "step": 5376
1018
+ },
1019
+ {
1020
+ "epoch": 0.43,
1021
+ "learning_rate": 6.019326150162476e-07,
1022
+ "loss": 0.2378,
1023
+ "step": 5408
1024
+ },
1025
+ {
1026
+ "epoch": 0.44,
1027
+ "learning_rate": 5.991961689755431e-07,
1028
+ "loss": 0.2488,
1029
+ "step": 5440
1030
+ },
1031
+ {
1032
+ "epoch": 0.44,
1033
+ "learning_rate": 5.964597229348383e-07,
1034
+ "loss": 0.2459,
1035
+ "step": 5472
1036
+ },
1037
+ {
1038
+ "epoch": 0.44,
1039
+ "learning_rate": 5.937232768941337e-07,
1040
+ "loss": 0.1928,
1041
+ "step": 5504
1042
+ },
1043
+ {
1044
+ "epoch": 0.45,
1045
+ "learning_rate": 5.90986830853429e-07,
1046
+ "loss": 0.2617,
1047
+ "step": 5536
1048
+ },
1049
+ {
1050
+ "epoch": 0.45,
1051
+ "learning_rate": 5.882503848127245e-07,
1052
+ "loss": 0.1927,
1053
+ "step": 5568
1054
+ },
1055
+ {
1056
+ "epoch": 0.45,
1057
+ "learning_rate": 5.855139387720198e-07,
1058
+ "loss": 0.1943,
1059
+ "step": 5600
1060
+ },
1061
+ {
1062
+ "epoch": 0.45,
1063
+ "learning_rate": 5.827774927313152e-07,
1064
+ "loss": 0.241,
1065
+ "step": 5632
1066
+ },
1067
+ {
1068
+ "epoch": 0.46,
1069
+ "learning_rate": 5.800410466906105e-07,
1070
+ "loss": 0.2354,
1071
+ "step": 5664
1072
+ },
1073
+ {
1074
+ "epoch": 0.46,
1075
+ "learning_rate": 5.77304600649906e-07,
1076
+ "loss": 0.2227,
1077
+ "step": 5696
1078
+ },
1079
+ {
1080
+ "epoch": 0.46,
1081
+ "learning_rate": 5.745681546092013e-07,
1082
+ "loss": 0.2395,
1083
+ "step": 5728
1084
+ },
1085
+ {
1086
+ "epoch": 0.46,
1087
+ "learning_rate": 5.718317085684967e-07,
1088
+ "loss": 0.1895,
1089
+ "step": 5760
1090
+ },
1091
+ {
1092
+ "epoch": 0.47,
1093
+ "learning_rate": 5.690952625277921e-07,
1094
+ "loss": 0.1621,
1095
+ "step": 5792
1096
+ },
1097
+ {
1098
+ "epoch": 0.47,
1099
+ "learning_rate": 5.663588164870873e-07,
1100
+ "loss": 0.2169,
1101
+ "step": 5824
1102
+ },
1103
+ {
1104
+ "epoch": 0.47,
1105
+ "learning_rate": 5.636223704463827e-07,
1106
+ "loss": 0.1979,
1107
+ "step": 5856
1108
+ },
1109
+ {
1110
+ "epoch": 0.47,
1111
+ "learning_rate": 5.60885924405678e-07,
1112
+ "loss": 0.206,
1113
+ "step": 5888
1114
+ },
1115
+ {
1116
+ "epoch": 0.48,
1117
+ "learning_rate": 5.581494783649735e-07,
1118
+ "loss": 0.1625,
1119
+ "step": 5920
1120
+ },
1121
+ {
1122
+ "epoch": 0.48,
1123
+ "learning_rate": 5.554130323242688e-07,
1124
+ "loss": 0.2625,
1125
+ "step": 5952
1126
+ },
1127
+ {
1128
+ "epoch": 0.48,
1129
+ "learning_rate": 5.526765862835642e-07,
1130
+ "loss": 0.1797,
1131
+ "step": 5984
1132
+ },
1133
+ {
1134
+ "epoch": 0.48,
1135
+ "learning_rate": 5.499401402428595e-07,
1136
+ "loss": 0.2056,
1137
+ "step": 6016
1138
+ },
1139
+ {
1140
+ "epoch": 0.49,
1141
+ "learning_rate": 5.47203694202155e-07,
1142
+ "loss": 0.1659,
1143
+ "step": 6048
1144
+ },
1145
+ {
1146
+ "epoch": 0.49,
1147
+ "learning_rate": 5.444672481614503e-07,
1148
+ "loss": 0.1524,
1149
+ "step": 6080
1150
+ },
1151
+ {
1152
+ "epoch": 0.49,
1153
+ "learning_rate": 5.417308021207456e-07,
1154
+ "loss": 0.236,
1155
+ "step": 6112
1156
+ },
1157
+ {
1158
+ "epoch": 0.49,
1159
+ "learning_rate": 5.38994356080041e-07,
1160
+ "loss": 0.1921,
1161
+ "step": 6144
1162
+ },
1163
+ {
1164
+ "epoch": 0.5,
1165
+ "learning_rate": 5.362579100393364e-07,
1166
+ "loss": 0.217,
1167
+ "step": 6176
1168
+ },
1169
+ {
1170
+ "epoch": 0.5,
1171
+ "learning_rate": 5.335214639986317e-07,
1172
+ "loss": 0.1996,
1173
+ "step": 6208
1174
+ },
1175
+ {
1176
+ "epoch": 0.5,
1177
+ "learning_rate": 5.307850179579271e-07,
1178
+ "loss": 0.1467,
1179
+ "step": 6240
1180
+ },
1181
+ {
1182
+ "epoch": 0.5,
1183
+ "learning_rate": 5.280485719172225e-07,
1184
+ "loss": 0.2702,
1185
+ "step": 6272
1186
+ },
1187
+ {
1188
+ "epoch": 0.51,
1189
+ "learning_rate": 5.253121258765179e-07,
1190
+ "loss": 0.1736,
1191
+ "step": 6304
1192
+ },
1193
+ {
1194
+ "epoch": 0.51,
1195
+ "learning_rate": 5.225756798358132e-07,
1196
+ "loss": 0.2281,
1197
+ "step": 6336
1198
+ },
1199
+ {
1200
+ "epoch": 0.51,
1201
+ "learning_rate": 5.198392337951087e-07,
1202
+ "loss": 0.2114,
1203
+ "step": 6368
1204
+ },
1205
+ {
1206
+ "epoch": 0.51,
1207
+ "learning_rate": 5.17102787754404e-07,
1208
+ "loss": 0.2252,
1209
+ "step": 6400
1210
+ },
1211
+ {
1212
+ "epoch": 0.52,
1213
+ "learning_rate": 5.143663417136994e-07,
1214
+ "loss": 0.163,
1215
+ "step": 6432
1216
+ },
1217
+ {
1218
+ "epoch": 0.52,
1219
+ "learning_rate": 5.116298956729946e-07,
1220
+ "loss": 0.266,
1221
+ "step": 6464
1222
+ },
1223
+ {
1224
+ "epoch": 0.52,
1225
+ "learning_rate": 5.0889344963229e-07,
1226
+ "loss": 0.2154,
1227
+ "step": 6496
1228
+ },
1229
+ {
1230
+ "epoch": 0.52,
1231
+ "learning_rate": 5.061570035915854e-07,
1232
+ "loss": 0.2003,
1233
+ "step": 6528
1234
+ },
1235
+ {
1236
+ "epoch": 0.53,
1237
+ "learning_rate": 5.034205575508807e-07,
1238
+ "loss": 0.1969,
1239
+ "step": 6560
1240
+ },
1241
+ {
1242
+ "epoch": 0.53,
1243
+ "learning_rate": 5.006841115101761e-07,
1244
+ "loss": 0.2436,
1245
+ "step": 6592
1246
+ },
1247
+ {
1248
+ "epoch": 0.53,
1249
+ "learning_rate": 4.979476654694714e-07,
1250
+ "loss": 0.134,
1251
+ "step": 6624
1252
+ },
1253
+ {
1254
+ "epoch": 0.54,
1255
+ "learning_rate": 4.952112194287669e-07,
1256
+ "loss": 0.2246,
1257
+ "step": 6656
1258
+ },
1259
+ {
1260
+ "epoch": 0.54,
1261
+ "learning_rate": 4.924747733880622e-07,
1262
+ "loss": 0.1873,
1263
+ "step": 6688
1264
+ },
1265
+ {
1266
+ "epoch": 0.54,
1267
+ "learning_rate": 4.897383273473576e-07,
1268
+ "loss": 0.1549,
1269
+ "step": 6720
1270
+ },
1271
+ {
1272
+ "epoch": 0.54,
1273
+ "learning_rate": 4.87001881306653e-07,
1274
+ "loss": 0.2518,
1275
+ "step": 6752
1276
+ },
1277
+ {
1278
+ "epoch": 0.55,
1279
+ "learning_rate": 4.842654352659483e-07,
1280
+ "loss": 0.2676,
1281
+ "step": 6784
1282
+ },
1283
+ {
1284
+ "epoch": 0.55,
1285
+ "learning_rate": 4.815289892252437e-07,
1286
+ "loss": 0.1537,
1287
+ "step": 6816
1288
+ },
1289
+ {
1290
+ "epoch": 0.55,
1291
+ "learning_rate": 4.787925431845391e-07,
1292
+ "loss": 0.1925,
1293
+ "step": 6848
1294
+ },
1295
+ {
1296
+ "epoch": 0.55,
1297
+ "learning_rate": 4.760560971438344e-07,
1298
+ "loss": 0.2157,
1299
+ "step": 6880
1300
+ },
1301
+ {
1302
+ "epoch": 0.56,
1303
+ "learning_rate": 4.7331965110312977e-07,
1304
+ "loss": 0.1923,
1305
+ "step": 6912
1306
+ },
1307
+ {
1308
+ "epoch": 0.56,
1309
+ "learning_rate": 4.7058320506242517e-07,
1310
+ "loss": 0.1922,
1311
+ "step": 6944
1312
+ },
1313
+ {
1314
+ "epoch": 0.56,
1315
+ "learning_rate": 4.678467590217205e-07,
1316
+ "loss": 0.257,
1317
+ "step": 6976
1318
+ },
1319
+ {
1320
+ "epoch": 0.56,
1321
+ "learning_rate": 4.6511031298101586e-07,
1322
+ "loss": 0.1744,
1323
+ "step": 7008
1324
+ },
1325
+ {
1326
+ "epoch": 0.57,
1327
+ "learning_rate": 4.6237386694031126e-07,
1328
+ "loss": 0.2281,
1329
+ "step": 7040
1330
+ },
1331
+ {
1332
+ "epoch": 0.57,
1333
+ "learning_rate": 4.596374208996066e-07,
1334
+ "loss": 0.2135,
1335
+ "step": 7072
1336
+ },
1337
+ {
1338
+ "epoch": 0.57,
1339
+ "learning_rate": 4.56900974858902e-07,
1340
+ "loss": 0.1841,
1341
+ "step": 7104
1342
+ },
1343
+ {
1344
+ "epoch": 0.57,
1345
+ "learning_rate": 4.5416452881819735e-07,
1346
+ "loss": 0.2652,
1347
+ "step": 7136
1348
+ },
1349
+ {
1350
+ "epoch": 0.58,
1351
+ "learning_rate": 4.5142808277749275e-07,
1352
+ "loss": 0.225,
1353
+ "step": 7168
1354
+ },
1355
+ {
1356
+ "epoch": 0.58,
1357
+ "learning_rate": 4.4869163673678804e-07,
1358
+ "loss": 0.2377,
1359
+ "step": 7200
1360
+ },
1361
+ {
1362
+ "epoch": 0.58,
1363
+ "learning_rate": 4.4595519069608344e-07,
1364
+ "loss": 0.2323,
1365
+ "step": 7232
1366
+ },
1367
+ {
1368
+ "epoch": 0.58,
1369
+ "learning_rate": 4.432187446553788e-07,
1370
+ "loss": 0.2227,
1371
+ "step": 7264
1372
+ },
1373
+ {
1374
+ "epoch": 0.59,
1375
+ "learning_rate": 4.404822986146742e-07,
1376
+ "loss": 0.2968,
1377
+ "step": 7296
1378
+ },
1379
+ {
1380
+ "epoch": 0.59,
1381
+ "learning_rate": 4.377458525739695e-07,
1382
+ "loss": 0.2623,
1383
+ "step": 7328
1384
+ },
1385
+ {
1386
+ "epoch": 0.59,
1387
+ "learning_rate": 4.350094065332649e-07,
1388
+ "loss": 0.1943,
1389
+ "step": 7360
1390
+ },
1391
+ {
1392
+ "epoch": 0.59,
1393
+ "learning_rate": 4.322729604925603e-07,
1394
+ "loss": 0.158,
1395
+ "step": 7392
1396
+ },
1397
+ {
1398
+ "epoch": 0.6,
1399
+ "learning_rate": 4.2953651445185563e-07,
1400
+ "loss": 0.2413,
1401
+ "step": 7424
1402
+ },
1403
+ {
1404
+ "epoch": 0.6,
1405
+ "learning_rate": 4.2680006841115103e-07,
1406
+ "loss": 0.2413,
1407
+ "step": 7456
1408
+ },
1409
+ {
1410
+ "epoch": 0.6,
1411
+ "learning_rate": 4.240636223704464e-07,
1412
+ "loss": 0.1793,
1413
+ "step": 7488
1414
+ },
1415
+ {
1416
+ "epoch": 0.6,
1417
+ "learning_rate": 4.213271763297417e-07,
1418
+ "loss": 0.2505,
1419
+ "step": 7520
1420
+ },
1421
+ {
1422
+ "epoch": 0.61,
1423
+ "learning_rate": 4.1859073028903707e-07,
1424
+ "loss": 0.2078,
1425
+ "step": 7552
1426
+ },
1427
+ {
1428
+ "epoch": 0.61,
1429
+ "learning_rate": 4.1585428424833247e-07,
1430
+ "loss": 0.2603,
1431
+ "step": 7584
1432
+ },
1433
+ {
1434
+ "epoch": 0.61,
1435
+ "learning_rate": 4.131178382076278e-07,
1436
+ "loss": 0.1426,
1437
+ "step": 7616
1438
+ },
1439
+ {
1440
+ "epoch": 0.61,
1441
+ "learning_rate": 4.103813921669232e-07,
1442
+ "loss": 0.1823,
1443
+ "step": 7648
1444
+ },
1445
+ {
1446
+ "epoch": 0.62,
1447
+ "learning_rate": 4.0764494612621856e-07,
1448
+ "loss": 0.2215,
1449
+ "step": 7680
1450
+ },
1451
+ {
1452
+ "epoch": 0.62,
1453
+ "learning_rate": 4.049085000855139e-07,
1454
+ "loss": 0.2956,
1455
+ "step": 7712
1456
+ },
1457
+ {
1458
+ "epoch": 0.62,
1459
+ "learning_rate": 4.021720540448093e-07,
1460
+ "loss": 0.1766,
1461
+ "step": 7744
1462
+ },
1463
+ {
1464
+ "epoch": 0.63,
1465
+ "learning_rate": 3.9943560800410465e-07,
1466
+ "loss": 0.2124,
1467
+ "step": 7776
1468
+ },
1469
+ {
1470
+ "epoch": 0.63,
1471
+ "learning_rate": 3.9669916196340005e-07,
1472
+ "loss": 0.2123,
1473
+ "step": 7808
1474
+ },
1475
+ {
1476
+ "epoch": 0.63,
1477
+ "learning_rate": 3.9396271592269535e-07,
1478
+ "loss": 0.2172,
1479
+ "step": 7840
1480
+ },
1481
+ {
1482
+ "epoch": 0.63,
1483
+ "learning_rate": 3.9122626988199075e-07,
1484
+ "loss": 0.1737,
1485
+ "step": 7872
1486
+ },
1487
+ {
1488
+ "epoch": 0.64,
1489
+ "learning_rate": 3.884898238412861e-07,
1490
+ "loss": 0.2504,
1491
+ "step": 7904
1492
+ },
1493
+ {
1494
+ "epoch": 0.64,
1495
+ "learning_rate": 3.857533778005815e-07,
1496
+ "loss": 0.2242,
1497
+ "step": 7936
1498
+ },
1499
+ {
1500
+ "epoch": 0.64,
1501
+ "learning_rate": 3.8301693175987684e-07,
1502
+ "loss": 0.205,
1503
+ "step": 7968
1504
+ },
1505
+ {
1506
+ "epoch": 0.64,
1507
+ "learning_rate": 3.802804857191722e-07,
1508
+ "loss": 0.2286,
1509
+ "step": 8000
1510
+ },
1511
+ {
1512
+ "epoch": 0.65,
1513
+ "learning_rate": 3.775440396784676e-07,
1514
+ "loss": 0.1848,
1515
+ "step": 8032
1516
+ },
1517
+ {
1518
+ "epoch": 0.65,
1519
+ "learning_rate": 3.7480759363776293e-07,
1520
+ "loss": 0.2065,
1521
+ "step": 8064
1522
+ },
1523
+ {
1524
+ "epoch": 0.65,
1525
+ "learning_rate": 3.7207114759705833e-07,
1526
+ "loss": 0.1688,
1527
+ "step": 8096
1528
+ },
1529
+ {
1530
+ "epoch": 0.65,
1531
+ "learning_rate": 3.693347015563537e-07,
1532
+ "loss": 0.2489,
1533
+ "step": 8128
1534
+ },
1535
+ {
1536
+ "epoch": 0.66,
1537
+ "learning_rate": 3.66598255515649e-07,
1538
+ "loss": 0.2309,
1539
+ "step": 8160
1540
+ },
1541
+ {
1542
+ "epoch": 0.66,
1543
+ "learning_rate": 3.6386180947494437e-07,
1544
+ "loss": 0.1236,
1545
+ "step": 8192
1546
+ },
1547
+ {
1548
+ "epoch": 0.66,
1549
+ "learning_rate": 3.6112536343423977e-07,
1550
+ "loss": 0.2242,
1551
+ "step": 8224
1552
+ },
1553
+ {
1554
+ "epoch": 0.66,
1555
+ "learning_rate": 3.583889173935351e-07,
1556
+ "loss": 0.221,
1557
+ "step": 8256
1558
+ },
1559
+ {
1560
+ "epoch": 0.67,
1561
+ "learning_rate": 3.556524713528305e-07,
1562
+ "loss": 0.1663,
1563
+ "step": 8288
1564
+ },
1565
+ {
1566
+ "epoch": 0.67,
1567
+ "learning_rate": 3.5291602531212586e-07,
1568
+ "loss": 0.2063,
1569
+ "step": 8320
1570
+ },
1571
+ {
1572
+ "epoch": 0.67,
1573
+ "learning_rate": 3.501795792714212e-07,
1574
+ "loss": 0.2651,
1575
+ "step": 8352
1576
+ },
1577
+ {
1578
+ "epoch": 0.67,
1579
+ "learning_rate": 3.474431332307166e-07,
1580
+ "loss": 0.2451,
1581
+ "step": 8384
1582
+ },
1583
+ {
1584
+ "epoch": 0.68,
1585
+ "learning_rate": 3.4470668719001196e-07,
1586
+ "loss": 0.1645,
1587
+ "step": 8416
1588
+ },
1589
+ {
1590
+ "epoch": 0.68,
1591
+ "learning_rate": 3.4197024114930736e-07,
1592
+ "loss": 0.1446,
1593
+ "step": 8448
1594
+ },
1595
+ {
1596
+ "epoch": 0.68,
1597
+ "learning_rate": 3.3923379510860265e-07,
1598
+ "loss": 0.2006,
1599
+ "step": 8480
1600
+ },
1601
+ {
1602
+ "epoch": 0.68,
1603
+ "learning_rate": 3.3649734906789805e-07,
1604
+ "loss": 0.1702,
1605
+ "step": 8512
1606
+ },
1607
+ {
1608
+ "epoch": 0.69,
1609
+ "learning_rate": 3.337609030271934e-07,
1610
+ "loss": 0.2685,
1611
+ "step": 8544
1612
+ },
1613
+ {
1614
+ "epoch": 0.69,
1615
+ "learning_rate": 3.310244569864888e-07,
1616
+ "loss": 0.1944,
1617
+ "step": 8576
1618
+ },
1619
+ {
1620
+ "epoch": 0.69,
1621
+ "learning_rate": 3.2828801094578414e-07,
1622
+ "loss": 0.1461,
1623
+ "step": 8608
1624
+ },
1625
+ {
1626
+ "epoch": 0.69,
1627
+ "learning_rate": 3.255515649050795e-07,
1628
+ "loss": 0.2539,
1629
+ "step": 8640
1630
+ },
1631
+ {
1632
+ "epoch": 0.7,
1633
+ "learning_rate": 3.228151188643749e-07,
1634
+ "loss": 0.1783,
1635
+ "step": 8672
1636
+ },
1637
+ {
1638
+ "epoch": 0.7,
1639
+ "learning_rate": 3.2007867282367024e-07,
1640
+ "loss": 0.2084,
1641
+ "step": 8704
1642
+ },
1643
+ {
1644
+ "epoch": 0.7,
1645
+ "learning_rate": 3.1734222678296563e-07,
1646
+ "loss": 0.1596,
1647
+ "step": 8736
1648
+ },
1649
+ {
1650
+ "epoch": 0.7,
1651
+ "learning_rate": 3.14605780742261e-07,
1652
+ "loss": 0.1677,
1653
+ "step": 8768
1654
+ },
1655
+ {
1656
+ "epoch": 0.71,
1657
+ "learning_rate": 3.1186933470155633e-07,
1658
+ "loss": 0.1636,
1659
+ "step": 8800
1660
+ },
1661
+ {
1662
+ "epoch": 0.71,
1663
+ "learning_rate": 3.091328886608517e-07,
1664
+ "loss": 0.2336,
1665
+ "step": 8832
1666
+ },
1667
+ {
1668
+ "epoch": 0.71,
1669
+ "learning_rate": 3.063964426201471e-07,
1670
+ "loss": 0.1853,
1671
+ "step": 8864
1672
+ },
1673
+ {
1674
+ "epoch": 0.72,
1675
+ "learning_rate": 3.036599965794424e-07,
1676
+ "loss": 0.2304,
1677
+ "step": 8896
1678
+ },
1679
+ {
1680
+ "epoch": 0.72,
1681
+ "learning_rate": 3.009235505387378e-07,
1682
+ "loss": 0.2037,
1683
+ "step": 8928
1684
+ },
1685
+ {
1686
+ "epoch": 0.72,
1687
+ "learning_rate": 2.9818710449803317e-07,
1688
+ "loss": 0.2159,
1689
+ "step": 8960
1690
+ },
1691
+ {
1692
+ "epoch": 0.72,
1693
+ "learning_rate": 2.954506584573285e-07,
1694
+ "loss": 0.2512,
1695
+ "step": 8992
1696
+ },
1697
+ {
1698
+ "epoch": 0.73,
1699
+ "learning_rate": 2.927142124166239e-07,
1700
+ "loss": 0.2011,
1701
+ "step": 9024
1702
+ },
1703
+ {
1704
+ "epoch": 0.73,
1705
+ "learning_rate": 2.8997776637591926e-07,
1706
+ "loss": 0.1677,
1707
+ "step": 9056
1708
+ },
1709
+ {
1710
+ "epoch": 0.73,
1711
+ "learning_rate": 2.8724132033521466e-07,
1712
+ "loss": 0.192,
1713
+ "step": 9088
1714
+ },
1715
+ {
1716
+ "epoch": 0.73,
1717
+ "learning_rate": 2.8450487429450995e-07,
1718
+ "loss": 0.1615,
1719
+ "step": 9120
1720
+ },
1721
+ {
1722
+ "epoch": 0.74,
1723
+ "learning_rate": 2.8176842825380535e-07,
1724
+ "loss": 0.1646,
1725
+ "step": 9152
1726
+ },
1727
+ {
1728
+ "epoch": 0.74,
1729
+ "learning_rate": 2.790319822131007e-07,
1730
+ "loss": 0.2301,
1731
+ "step": 9184
1732
+ },
1733
+ {
1734
+ "epoch": 0.74,
1735
+ "learning_rate": 2.762955361723961e-07,
1736
+ "loss": 0.1663,
1737
+ "step": 9216
1738
+ },
1739
+ {
1740
+ "epoch": 0.74,
1741
+ "learning_rate": 2.7355909013169145e-07,
1742
+ "loss": 0.251,
1743
+ "step": 9248
1744
+ },
1745
+ {
1746
+ "epoch": 0.75,
1747
+ "learning_rate": 2.708226440909868e-07,
1748
+ "loss": 0.2301,
1749
+ "step": 9280
1750
+ },
1751
+ {
1752
+ "epoch": 0.75,
1753
+ "learning_rate": 2.680861980502822e-07,
1754
+ "loss": 0.2222,
1755
+ "step": 9312
1756
+ },
1757
+ {
1758
+ "epoch": 0.75,
1759
+ "learning_rate": 2.6534975200957754e-07,
1760
+ "loss": 0.1784,
1761
+ "step": 9344
1762
+ },
1763
+ {
1764
+ "epoch": 0.75,
1765
+ "learning_rate": 2.6261330596887294e-07,
1766
+ "loss": 0.1714,
1767
+ "step": 9376
1768
+ },
1769
+ {
1770
+ "epoch": 0.76,
1771
+ "learning_rate": 2.598768599281683e-07,
1772
+ "loss": 0.2258,
1773
+ "step": 9408
1774
+ },
1775
+ {
1776
+ "epoch": 0.76,
1777
+ "learning_rate": 2.571404138874637e-07,
1778
+ "loss": 0.1907,
1779
+ "step": 9440
1780
+ },
1781
+ {
1782
+ "epoch": 0.76,
1783
+ "learning_rate": 2.54403967846759e-07,
1784
+ "loss": 0.1938,
1785
+ "step": 9472
1786
+ },
1787
+ {
1788
+ "epoch": 0.76,
1789
+ "learning_rate": 2.516675218060544e-07,
1790
+ "loss": 0.1831,
1791
+ "step": 9504
1792
+ },
1793
+ {
1794
+ "epoch": 0.77,
1795
+ "learning_rate": 2.489310757653497e-07,
1796
+ "loss": 0.1833,
1797
+ "step": 9536
1798
+ },
1799
+ {
1800
+ "epoch": 0.77,
1801
+ "learning_rate": 2.461946297246451e-07,
1802
+ "loss": 0.2551,
1803
+ "step": 9568
1804
+ },
1805
+ {
1806
+ "epoch": 0.77,
1807
+ "learning_rate": 2.4345818368394047e-07,
1808
+ "loss": 0.1553,
1809
+ "step": 9600
1810
+ },
1811
+ {
1812
+ "epoch": 0.77,
1813
+ "learning_rate": 2.407217376432358e-07,
1814
+ "loss": 0.1785,
1815
+ "step": 9632
1816
+ },
1817
+ {
1818
+ "epoch": 0.78,
1819
+ "learning_rate": 2.379852916025312e-07,
1820
+ "loss": 0.2903,
1821
+ "step": 9664
1822
+ },
1823
+ {
1824
+ "epoch": 0.78,
1825
+ "learning_rate": 2.3524884556182656e-07,
1826
+ "loss": 0.2093,
1827
+ "step": 9696
1828
+ },
1829
+ {
1830
+ "epoch": 0.78,
1831
+ "learning_rate": 2.325123995211219e-07,
1832
+ "loss": 0.2865,
1833
+ "step": 9728
1834
+ },
1835
+ {
1836
+ "epoch": 0.78,
1837
+ "learning_rate": 2.2977595348041728e-07,
1838
+ "loss": 0.2033,
1839
+ "step": 9760
1840
+ },
1841
+ {
1842
+ "epoch": 0.79,
1843
+ "learning_rate": 2.2703950743971268e-07,
1844
+ "loss": 0.1929,
1845
+ "step": 9792
1846
+ },
1847
+ {
1848
+ "epoch": 0.79,
1849
+ "learning_rate": 2.2430306139900803e-07,
1850
+ "loss": 0.1968,
1851
+ "step": 9824
1852
+ },
1853
+ {
1854
+ "epoch": 0.79,
1855
+ "learning_rate": 2.215666153583034e-07,
1856
+ "loss": 0.1985,
1857
+ "step": 9856
1858
+ },
1859
+ {
1860
+ "epoch": 0.79,
1861
+ "learning_rate": 2.1883016931759877e-07,
1862
+ "loss": 0.1509,
1863
+ "step": 9888
1864
+ },
1865
+ {
1866
+ "epoch": 0.8,
1867
+ "learning_rate": 2.1609372327689412e-07,
1868
+ "loss": 0.1988,
1869
+ "step": 9920
1870
+ },
1871
+ {
1872
+ "epoch": 0.8,
1873
+ "learning_rate": 2.133572772361895e-07,
1874
+ "loss": 0.2059,
1875
+ "step": 9952
1876
+ },
1877
+ {
1878
+ "epoch": 0.8,
1879
+ "learning_rate": 2.1062083119548484e-07,
1880
+ "loss": 0.1701,
1881
+ "step": 9984
1882
+ },
1883
+ {
1884
+ "epoch": 0.81,
1885
+ "learning_rate": 2.0788438515478021e-07,
1886
+ "loss": 0.1524,
1887
+ "step": 10016
1888
+ },
1889
+ {
1890
+ "epoch": 0.81,
1891
+ "learning_rate": 2.051479391140756e-07,
1892
+ "loss": 0.2236,
1893
+ "step": 10048
1894
+ },
1895
+ {
1896
+ "epoch": 0.81,
1897
+ "learning_rate": 2.0241149307337093e-07,
1898
+ "loss": 0.189,
1899
+ "step": 10080
1900
+ },
1901
+ {
1902
+ "epoch": 0.81,
1903
+ "learning_rate": 1.9967504703266633e-07,
1904
+ "loss": 0.1464,
1905
+ "step": 10112
1906
+ },
1907
+ {
1908
+ "epoch": 0.82,
1909
+ "learning_rate": 1.9693860099196168e-07,
1910
+ "loss": 0.2245,
1911
+ "step": 10144
1912
+ },
1913
+ {
1914
+ "epoch": 0.82,
1915
+ "learning_rate": 1.9420215495125705e-07,
1916
+ "loss": 0.2619,
1917
+ "step": 10176
1918
+ },
1919
+ {
1920
+ "epoch": 0.82,
1921
+ "learning_rate": 1.9146570891055243e-07,
1922
+ "loss": 0.2331,
1923
+ "step": 10208
1924
+ },
1925
+ {
1926
+ "epoch": 0.82,
1927
+ "learning_rate": 1.8872926286984777e-07,
1928
+ "loss": 0.1925,
1929
+ "step": 10240
1930
+ },
1931
+ {
1932
+ "epoch": 0.83,
1933
+ "learning_rate": 1.8599281682914315e-07,
1934
+ "loss": 0.2707,
1935
+ "step": 10272
1936
+ },
1937
+ {
1938
+ "epoch": 0.83,
1939
+ "learning_rate": 1.832563707884385e-07,
1940
+ "loss": 0.1792,
1941
+ "step": 10304
1942
+ },
1943
+ {
1944
+ "epoch": 0.83,
1945
+ "learning_rate": 1.8051992474773387e-07,
1946
+ "loss": 0.1663,
1947
+ "step": 10336
1948
+ },
1949
+ {
1950
+ "epoch": 0.83,
1951
+ "learning_rate": 1.7778347870702924e-07,
1952
+ "loss": 0.1783,
1953
+ "step": 10368
1954
+ },
1955
+ {
1956
+ "epoch": 0.84,
1957
+ "learning_rate": 1.7504703266632459e-07,
1958
+ "loss": 0.2028,
1959
+ "step": 10400
1960
+ },
1961
+ {
1962
+ "epoch": 0.84,
1963
+ "learning_rate": 1.7231058662561998e-07,
1964
+ "loss": 0.2036,
1965
+ "step": 10432
1966
+ },
1967
+ {
1968
+ "epoch": 0.84,
1969
+ "learning_rate": 1.6957414058491533e-07,
1970
+ "loss": 0.1832,
1971
+ "step": 10464
1972
+ },
1973
+ {
1974
+ "epoch": 0.84,
1975
+ "learning_rate": 1.668376945442107e-07,
1976
+ "loss": 0.1795,
1977
+ "step": 10496
1978
+ },
1979
+ {
1980
+ "epoch": 0.85,
1981
+ "learning_rate": 1.6410124850350608e-07,
1982
+ "loss": 0.1507,
1983
+ "step": 10528
1984
+ },
1985
+ {
1986
+ "epoch": 0.85,
1987
+ "learning_rate": 1.6136480246280142e-07,
1988
+ "loss": 0.2078,
1989
+ "step": 10560
1990
+ },
1991
+ {
1992
+ "epoch": 0.85,
1993
+ "learning_rate": 1.586283564220968e-07,
1994
+ "loss": 0.2228,
1995
+ "step": 10592
1996
+ },
1997
+ {
1998
+ "epoch": 0.85,
1999
+ "learning_rate": 1.5589191038139214e-07,
2000
+ "loss": 0.2,
2001
+ "step": 10624
2002
+ },
2003
+ {
2004
+ "epoch": 0.86,
2005
+ "learning_rate": 1.5315546434068752e-07,
2006
+ "loss": 0.1615,
2007
+ "step": 10656
2008
+ },
2009
+ {
2010
+ "epoch": 0.86,
2011
+ "learning_rate": 1.504190182999829e-07,
2012
+ "loss": 0.1746,
2013
+ "step": 10688
2014
+ },
2015
+ {
2016
+ "epoch": 0.86,
2017
+ "learning_rate": 1.4768257225927824e-07,
2018
+ "loss": 0.1857,
2019
+ "step": 10720
2020
+ },
2021
+ {
2022
+ "epoch": 0.86,
2023
+ "learning_rate": 1.4494612621857364e-07,
2024
+ "loss": 0.2143,
2025
+ "step": 10752
2026
+ },
2027
+ {
2028
+ "epoch": 0.87,
2029
+ "learning_rate": 1.4220968017786898e-07,
2030
+ "loss": 0.1919,
2031
+ "step": 10784
2032
+ },
2033
+ {
2034
+ "epoch": 0.87,
2035
+ "learning_rate": 1.3947323413716436e-07,
2036
+ "loss": 0.1885,
2037
+ "step": 10816
2038
+ },
2039
+ {
2040
+ "epoch": 0.87,
2041
+ "learning_rate": 1.3673678809645973e-07,
2042
+ "loss": 0.1617,
2043
+ "step": 10848
2044
+ },
2045
+ {
2046
+ "epoch": 0.87,
2047
+ "learning_rate": 1.3400034205575508e-07,
2048
+ "loss": 0.1512,
2049
+ "step": 10880
2050
+ },
2051
+ {
2052
+ "epoch": 0.88,
2053
+ "learning_rate": 1.3126389601505045e-07,
2054
+ "loss": 0.1666,
2055
+ "step": 10912
2056
+ },
2057
+ {
2058
+ "epoch": 0.88,
2059
+ "learning_rate": 1.2852744997434582e-07,
2060
+ "loss": 0.1519,
2061
+ "step": 10944
2062
+ },
2063
+ {
2064
+ "epoch": 0.88,
2065
+ "learning_rate": 1.2579100393364117e-07,
2066
+ "loss": 0.1474,
2067
+ "step": 10976
2068
+ },
2069
+ {
2070
+ "epoch": 0.88,
2071
+ "learning_rate": 1.2305455789293654e-07,
2072
+ "loss": 0.3074,
2073
+ "step": 11008
2074
+ },
2075
+ {
2076
+ "epoch": 0.89,
2077
+ "learning_rate": 1.2031811185223191e-07,
2078
+ "loss": 0.1517,
2079
+ "step": 11040
2080
+ },
2081
+ {
2082
+ "epoch": 0.89,
2083
+ "learning_rate": 1.1758166581152727e-07,
2084
+ "loss": 0.3134,
2085
+ "step": 11072
2086
+ },
2087
+ {
2088
+ "epoch": 0.89,
2089
+ "learning_rate": 1.1484521977082263e-07,
2090
+ "loss": 0.1734,
2091
+ "step": 11104
2092
+ },
2093
+ {
2094
+ "epoch": 0.9,
2095
+ "learning_rate": 1.12108773730118e-07,
2096
+ "loss": 0.24,
2097
+ "step": 11136
2098
+ },
2099
+ {
2100
+ "epoch": 0.9,
2101
+ "learning_rate": 1.0937232768941338e-07,
2102
+ "loss": 0.2515,
2103
+ "step": 11168
2104
+ }
2105
+ ],
2106
+ "logging_steps": 32,
2107
+ "max_steps": 12440,
2108
+ "num_input_tokens_seen": 0,
2109
+ "num_train_epochs": 1,
2110
+ "save_steps": 1243,
2111
+ "total_flos": 4.75038284193792e+19,
2112
+ "train_batch_size": 2,
2113
+ "trial_name": null,
2114
+ "trial_params": null
2115
+ }
checkpoint-11187/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2343c53f43ecba9eda3fbab0cfda8b6794aaa661c23626f520c968ab283189ab
3
+ size 5048
checkpoint-1243/config.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-large-v2",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "language": "English",
45
+ "mask_feature_length": 10,
46
+ "mask_feature_min_masks": 0,
47
+ "mask_feature_prob": 0.0,
48
+ "mask_time_length": 10,
49
+ "mask_time_min_masks": 2,
50
+ "mask_time_prob": 0.05,
51
+ "max_length": 448,
52
+ "max_source_positions": 1500,
53
+ "max_target_positions": 448,
54
+ "median_filter_width": 7,
55
+ "model_type": "whisper",
56
+ "num_hidden_layers": 32,
57
+ "num_mel_bins": 80,
58
+ "pad_token_id": 50257,
59
+ "scale_embedding": false,
60
+ "suppress_tokens": [],
61
+ "torch_dtype": "float32",
62
+ "transformers_version": "4.37.2",
63
+ "use_cache": true,
64
+ "use_weighted_layer_sum": false,
65
+ "vocab_size": 51865
66
+ }
checkpoint-1243/generation_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "begin_suppress_tokens": [
97
+ 220,
98
+ 50257
99
+ ],
100
+ "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
+ "forced_decoder_ids": [
104
+ [
105
+ 1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
+ ]
112
+ ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
+ "max_initial_timestamp_index": 50,
216
+ "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
+ "return_timestamps": false,
221
+ "suppress_tokens": [
222
+ 1,
223
+ 2,
224
+ 7,
225
+ 8,
226
+ 9,
227
+ 10,
228
+ 14,
229
+ 25,
230
+ 26,
231
+ 27,
232
+ 28,
233
+ 29,
234
+ 31,
235
+ 58,
236
+ 59,
237
+ 60,
238
+ 61,
239
+ 62,
240
+ 63,
241
+ 90,
242
+ 91,
243
+ 92,
244
+ 93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
305
+ 50358,
306
+ 50359,
307
+ 50360,
308
+ 50361,
309
+ 50362
310
+ ],
311
+ "task": "transcribe",
312
+ "task_to_id": {
313
+ "transcribe": 50359,
314
+ "translate": 50358
315
+ },
316
+ "transformers_version": "4.37.2"
317
+ }
checkpoint-1243/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:53dd4fe1b7540e808539f0bb9c5b944929f00c7f7900095a0db225b8d088bfdd
3
+ size 4992706480
checkpoint-1243/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3517c5ea86516a1c757abf66466a305606ea6f39c196daf6daf7818556aad9c3
3
+ size 1180663192
checkpoint-1243/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1243/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6e1aa5b65d67fc0981e7ad567dfb4d23be63817ba29537dca76aca65af892f2
3
+ size 3095074288
checkpoint-1243/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-1243/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3903d2c24f5d909c3588231ccdda3e1ce965628ad1d6bbc00bcf6420dbfb5a6
3
+ size 14244
checkpoint-1243/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca00b3236d817653aaa251d7bfffcc2d627276188e63cc50758b7f5fdc1187d1
3
+ size 1064
checkpoint-1243/trainer_state.json ADDED
@@ -0,0 +1,249 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.09991961414790997,
5
+ "eval_steps": 500,
6
+ "global_step": 1243,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 3.753351206434316e-08,
14
+ "loss": 4.9531,
15
+ "step": 32
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 8.042895442359249e-08,
20
+ "loss": 4.9011,
21
+ "step": 64
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 1.2198391420911528e-07,
26
+ "loss": 4.472,
27
+ "step": 96
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 1.648793565683646e-07,
32
+ "loss": 3.9678,
33
+ "step": 128
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 2.0777479892761392e-07,
38
+ "loss": 3.3412,
39
+ "step": 160
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 2.5067024128686325e-07,
44
+ "loss": 2.5657,
45
+ "step": 192
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 2.9356568364611256e-07,
50
+ "loss": 1.9677,
51
+ "step": 224
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 3.364611260053619e-07,
56
+ "loss": 1.6646,
57
+ "step": 256
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 3.7935656836461123e-07,
62
+ "loss": 1.5032,
63
+ "step": 288
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 4.222520107238606e-07,
68
+ "loss": 1.3443,
69
+ "step": 320
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 4.651474530831099e-07,
74
+ "loss": 1.2277,
75
+ "step": 352
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 5.080428954423593e-07,
80
+ "loss": 1.0104,
81
+ "step": 384
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 5.509383378016086e-07,
86
+ "loss": 0.7317,
87
+ "step": 416
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 5.938337801608579e-07,
92
+ "loss": 0.3644,
93
+ "step": 448
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 6.367292225201072e-07,
98
+ "loss": 0.388,
99
+ "step": 480
100
+ },
101
+ {
102
+ "epoch": 0.04,
103
+ "learning_rate": 6.796246648793566e-07,
104
+ "loss": 0.264,
105
+ "step": 512
106
+ },
107
+ {
108
+ "epoch": 0.04,
109
+ "learning_rate": 7.225201072386059e-07,
110
+ "loss": 0.2479,
111
+ "step": 544
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 7.654155495978551e-07,
116
+ "loss": 0.403,
117
+ "step": 576
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 8.083109919571045e-07,
122
+ "loss": 0.3062,
123
+ "step": 608
124
+ },
125
+ {
126
+ "epoch": 0.05,
127
+ "learning_rate": 8.512064343163538e-07,
128
+ "loss": 0.3016,
129
+ "step": 640
130
+ },
131
+ {
132
+ "epoch": 0.05,
133
+ "learning_rate": 8.941018766756032e-07,
134
+ "loss": 0.3445,
135
+ "step": 672
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.369973190348524e-07,
140
+ "loss": 0.2554,
141
+ "step": 704
142
+ },
143
+ {
144
+ "epoch": 0.06,
145
+ "learning_rate": 9.798927613941018e-07,
146
+ "loss": 0.2762,
147
+ "step": 736
148
+ },
149
+ {
150
+ "epoch": 0.06,
151
+ "learning_rate": 9.985462630408755e-07,
152
+ "loss": 0.2265,
153
+ "step": 768
154
+ },
155
+ {
156
+ "epoch": 0.06,
157
+ "learning_rate": 9.958098170001709e-07,
158
+ "loss": 0.2158,
159
+ "step": 800
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.930733709594662e-07,
164
+ "loss": 0.3106,
165
+ "step": 832
166
+ },
167
+ {
168
+ "epoch": 0.07,
169
+ "learning_rate": 9.903369249187618e-07,
170
+ "loss": 0.3085,
171
+ "step": 864
172
+ },
173
+ {
174
+ "epoch": 0.07,
175
+ "learning_rate": 9.876004788780571e-07,
176
+ "loss": 0.2633,
177
+ "step": 896
178
+ },
179
+ {
180
+ "epoch": 0.07,
181
+ "learning_rate": 9.848640328373525e-07,
182
+ "loss": 0.2145,
183
+ "step": 928
184
+ },
185
+ {
186
+ "epoch": 0.08,
187
+ "learning_rate": 9.821275867966478e-07,
188
+ "loss": 0.2594,
189
+ "step": 960
190
+ },
191
+ {
192
+ "epoch": 0.08,
193
+ "learning_rate": 9.793911407559432e-07,
194
+ "loss": 0.2264,
195
+ "step": 992
196
+ },
197
+ {
198
+ "epoch": 0.08,
199
+ "learning_rate": 9.766546947152385e-07,
200
+ "loss": 0.2512,
201
+ "step": 1024
202
+ },
203
+ {
204
+ "epoch": 0.08,
205
+ "learning_rate": 9.739182486745339e-07,
206
+ "loss": 0.3154,
207
+ "step": 1056
208
+ },
209
+ {
210
+ "epoch": 0.09,
211
+ "learning_rate": 9.711818026338292e-07,
212
+ "loss": 0.2672,
213
+ "step": 1088
214
+ },
215
+ {
216
+ "epoch": 0.09,
217
+ "learning_rate": 9.684453565931246e-07,
218
+ "loss": 0.2502,
219
+ "step": 1120
220
+ },
221
+ {
222
+ "epoch": 0.09,
223
+ "learning_rate": 9.6570891055242e-07,
224
+ "loss": 0.2702,
225
+ "step": 1152
226
+ },
227
+ {
228
+ "epoch": 0.1,
229
+ "learning_rate": 9.629724645117153e-07,
230
+ "loss": 0.2919,
231
+ "step": 1184
232
+ },
233
+ {
234
+ "epoch": 0.1,
235
+ "learning_rate": 9.602360184710108e-07,
236
+ "loss": 0.3925,
237
+ "step": 1216
238
+ }
239
+ ],
240
+ "logging_steps": 32,
241
+ "max_steps": 12440,
242
+ "num_input_tokens_seen": 0,
243
+ "num_train_epochs": 1,
244
+ "save_steps": 1243,
245
+ "total_flos": 5.2782031577088e+18,
246
+ "train_batch_size": 2,
247
+ "trial_name": null,
248
+ "trial_params": null
249
+ }
checkpoint-1243/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2343c53f43ecba9eda3fbab0cfda8b6794aaa661c23626f520c968ab283189ab
3
+ size 5048
checkpoint-12430/config.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-large-v2",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "language": "English",
45
+ "mask_feature_length": 10,
46
+ "mask_feature_min_masks": 0,
47
+ "mask_feature_prob": 0.0,
48
+ "mask_time_length": 10,
49
+ "mask_time_min_masks": 2,
50
+ "mask_time_prob": 0.05,
51
+ "max_length": 448,
52
+ "max_source_positions": 1500,
53
+ "max_target_positions": 448,
54
+ "median_filter_width": 7,
55
+ "model_type": "whisper",
56
+ "num_hidden_layers": 32,
57
+ "num_mel_bins": 80,
58
+ "pad_token_id": 50257,
59
+ "scale_embedding": false,
60
+ "suppress_tokens": [],
61
+ "torch_dtype": "float32",
62
+ "transformers_version": "4.37.2",
63
+ "use_cache": true,
64
+ "use_weighted_layer_sum": false,
65
+ "vocab_size": 51865
66
+ }
checkpoint-12430/generation_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "begin_suppress_tokens": [
97
+ 220,
98
+ 50257
99
+ ],
100
+ "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
+ "forced_decoder_ids": [
104
+ [
105
+ 1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
+ ]
112
+ ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
+ "max_initial_timestamp_index": 50,
216
+ "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
+ "return_timestamps": false,
221
+ "suppress_tokens": [
222
+ 1,
223
+ 2,
224
+ 7,
225
+ 8,
226
+ 9,
227
+ 10,
228
+ 14,
229
+ 25,
230
+ 26,
231
+ 27,
232
+ 28,
233
+ 29,
234
+ 31,
235
+ 58,
236
+ 59,
237
+ 60,
238
+ 61,
239
+ 62,
240
+ 63,
241
+ 90,
242
+ 91,
243
+ 92,
244
+ 93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
305
+ 50358,
306
+ 50359,
307
+ 50360,
308
+ 50361,
309
+ 50362
310
+ ],
311
+ "task": "transcribe",
312
+ "task_to_id": {
313
+ "transcribe": 50359,
314
+ "translate": 50358
315
+ },
316
+ "transformers_version": "4.37.2"
317
+ }
checkpoint-12430/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abcb38ffd84cb0cdd12bdbdcd2e051ed89beb5816bfa127790e24dbbc4760bb9
3
+ size 4992706480
checkpoint-12430/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aed090fddcb2645613adf66061b9db912fca6dba963bef57f78daba1b931f1a
3
+ size 1180663192
checkpoint-12430/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-12430/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e37796041013f1d55d29eb17136b55bdca42f392762bf7a0c91ff990105736d
3
+ size 3095074288
checkpoint-12430/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-12430/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:068026cd5cf8e98946c482d433bf9914b7464f4949e9495be438b56bdc0b35f0
3
+ size 14244
checkpoint-12430/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a93ce86da525a84bdadccb70efa15322ea47e5a4e5d6b55309383d7e3340d1d
3
+ size 1064
checkpoint-12430/trainer_state.json ADDED
@@ -0,0 +1,2349 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9991961414790996,
5
+ "eval_steps": 500,
6
+ "global_step": 12430,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 3.753351206434316e-08,
14
+ "loss": 4.9531,
15
+ "step": 32
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 8.042895442359249e-08,
20
+ "loss": 4.9011,
21
+ "step": 64
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 1.2198391420911528e-07,
26
+ "loss": 4.472,
27
+ "step": 96
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 1.648793565683646e-07,
32
+ "loss": 3.9678,
33
+ "step": 128
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 2.0777479892761392e-07,
38
+ "loss": 3.3412,
39
+ "step": 160
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 2.5067024128686325e-07,
44
+ "loss": 2.5657,
45
+ "step": 192
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 2.9356568364611256e-07,
50
+ "loss": 1.9677,
51
+ "step": 224
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 3.364611260053619e-07,
56
+ "loss": 1.6646,
57
+ "step": 256
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 3.7935656836461123e-07,
62
+ "loss": 1.5032,
63
+ "step": 288
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 4.222520107238606e-07,
68
+ "loss": 1.3443,
69
+ "step": 320
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 4.651474530831099e-07,
74
+ "loss": 1.2277,
75
+ "step": 352
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 5.080428954423593e-07,
80
+ "loss": 1.0104,
81
+ "step": 384
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 5.509383378016086e-07,
86
+ "loss": 0.7317,
87
+ "step": 416
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 5.938337801608579e-07,
92
+ "loss": 0.3644,
93
+ "step": 448
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 6.367292225201072e-07,
98
+ "loss": 0.388,
99
+ "step": 480
100
+ },
101
+ {
102
+ "epoch": 0.04,
103
+ "learning_rate": 6.796246648793566e-07,
104
+ "loss": 0.264,
105
+ "step": 512
106
+ },
107
+ {
108
+ "epoch": 0.04,
109
+ "learning_rate": 7.225201072386059e-07,
110
+ "loss": 0.2479,
111
+ "step": 544
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 7.654155495978551e-07,
116
+ "loss": 0.403,
117
+ "step": 576
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 8.083109919571045e-07,
122
+ "loss": 0.3062,
123
+ "step": 608
124
+ },
125
+ {
126
+ "epoch": 0.05,
127
+ "learning_rate": 8.512064343163538e-07,
128
+ "loss": 0.3016,
129
+ "step": 640
130
+ },
131
+ {
132
+ "epoch": 0.05,
133
+ "learning_rate": 8.941018766756032e-07,
134
+ "loss": 0.3445,
135
+ "step": 672
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.369973190348524e-07,
140
+ "loss": 0.2554,
141
+ "step": 704
142
+ },
143
+ {
144
+ "epoch": 0.06,
145
+ "learning_rate": 9.798927613941018e-07,
146
+ "loss": 0.2762,
147
+ "step": 736
148
+ },
149
+ {
150
+ "epoch": 0.06,
151
+ "learning_rate": 9.985462630408755e-07,
152
+ "loss": 0.2265,
153
+ "step": 768
154
+ },
155
+ {
156
+ "epoch": 0.06,
157
+ "learning_rate": 9.958098170001709e-07,
158
+ "loss": 0.2158,
159
+ "step": 800
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.930733709594662e-07,
164
+ "loss": 0.3106,
165
+ "step": 832
166
+ },
167
+ {
168
+ "epoch": 0.07,
169
+ "learning_rate": 9.903369249187618e-07,
170
+ "loss": 0.3085,
171
+ "step": 864
172
+ },
173
+ {
174
+ "epoch": 0.07,
175
+ "learning_rate": 9.876004788780571e-07,
176
+ "loss": 0.2633,
177
+ "step": 896
178
+ },
179
+ {
180
+ "epoch": 0.07,
181
+ "learning_rate": 9.848640328373525e-07,
182
+ "loss": 0.2145,
183
+ "step": 928
184
+ },
185
+ {
186
+ "epoch": 0.08,
187
+ "learning_rate": 9.821275867966478e-07,
188
+ "loss": 0.2594,
189
+ "step": 960
190
+ },
191
+ {
192
+ "epoch": 0.08,
193
+ "learning_rate": 9.793911407559432e-07,
194
+ "loss": 0.2264,
195
+ "step": 992
196
+ },
197
+ {
198
+ "epoch": 0.08,
199
+ "learning_rate": 9.766546947152385e-07,
200
+ "loss": 0.2512,
201
+ "step": 1024
202
+ },
203
+ {
204
+ "epoch": 0.08,
205
+ "learning_rate": 9.739182486745339e-07,
206
+ "loss": 0.3154,
207
+ "step": 1056
208
+ },
209
+ {
210
+ "epoch": 0.09,
211
+ "learning_rate": 9.711818026338292e-07,
212
+ "loss": 0.2672,
213
+ "step": 1088
214
+ },
215
+ {
216
+ "epoch": 0.09,
217
+ "learning_rate": 9.684453565931246e-07,
218
+ "loss": 0.2502,
219
+ "step": 1120
220
+ },
221
+ {
222
+ "epoch": 0.09,
223
+ "learning_rate": 9.6570891055242e-07,
224
+ "loss": 0.2702,
225
+ "step": 1152
226
+ },
227
+ {
228
+ "epoch": 0.1,
229
+ "learning_rate": 9.629724645117153e-07,
230
+ "loss": 0.2919,
231
+ "step": 1184
232
+ },
233
+ {
234
+ "epoch": 0.1,
235
+ "learning_rate": 9.602360184710108e-07,
236
+ "loss": 0.3925,
237
+ "step": 1216
238
+ },
239
+ {
240
+ "epoch": 0.1,
241
+ "learning_rate": 9.574995724303062e-07,
242
+ "loss": 0.285,
243
+ "step": 1248
244
+ },
245
+ {
246
+ "epoch": 0.1,
247
+ "learning_rate": 9.547631263896015e-07,
248
+ "loss": 0.3084,
249
+ "step": 1280
250
+ },
251
+ {
252
+ "epoch": 0.11,
253
+ "learning_rate": 9.520266803488969e-07,
254
+ "loss": 0.2275,
255
+ "step": 1312
256
+ },
257
+ {
258
+ "epoch": 0.11,
259
+ "learning_rate": 9.492902343081922e-07,
260
+ "loss": 0.245,
261
+ "step": 1344
262
+ },
263
+ {
264
+ "epoch": 0.11,
265
+ "learning_rate": 9.465537882674875e-07,
266
+ "loss": 0.233,
267
+ "step": 1376
268
+ },
269
+ {
270
+ "epoch": 0.11,
271
+ "learning_rate": 9.438173422267829e-07,
272
+ "loss": 0.2825,
273
+ "step": 1408
274
+ },
275
+ {
276
+ "epoch": 0.12,
277
+ "learning_rate": 9.410808961860783e-07,
278
+ "loss": 0.231,
279
+ "step": 1440
280
+ },
281
+ {
282
+ "epoch": 0.12,
283
+ "learning_rate": 9.383444501453737e-07,
284
+ "loss": 0.2449,
285
+ "step": 1472
286
+ },
287
+ {
288
+ "epoch": 0.12,
289
+ "learning_rate": 9.35608004104669e-07,
290
+ "loss": 0.2732,
291
+ "step": 1504
292
+ },
293
+ {
294
+ "epoch": 0.12,
295
+ "learning_rate": 9.328715580639644e-07,
296
+ "loss": 0.2031,
297
+ "step": 1536
298
+ },
299
+ {
300
+ "epoch": 0.13,
301
+ "learning_rate": 9.301351120232597e-07,
302
+ "loss": 0.1749,
303
+ "step": 1568
304
+ },
305
+ {
306
+ "epoch": 0.13,
307
+ "learning_rate": 9.273986659825551e-07,
308
+ "loss": 0.1722,
309
+ "step": 1600
310
+ },
311
+ {
312
+ "epoch": 0.13,
313
+ "learning_rate": 9.246622199418504e-07,
314
+ "loss": 0.2743,
315
+ "step": 1632
316
+ },
317
+ {
318
+ "epoch": 0.13,
319
+ "learning_rate": 9.219257739011459e-07,
320
+ "loss": 0.2907,
321
+ "step": 1664
322
+ },
323
+ {
324
+ "epoch": 0.14,
325
+ "learning_rate": 9.192748417992132e-07,
326
+ "loss": 0.2664,
327
+ "step": 1696
328
+ },
329
+ {
330
+ "epoch": 0.14,
331
+ "learning_rate": 9.165383957585086e-07,
332
+ "loss": 0.2085,
333
+ "step": 1728
334
+ },
335
+ {
336
+ "epoch": 0.14,
337
+ "learning_rate": 9.13801949717804e-07,
338
+ "loss": 0.1839,
339
+ "step": 1760
340
+ },
341
+ {
342
+ "epoch": 0.14,
343
+ "learning_rate": 9.110655036770994e-07,
344
+ "loss": 0.2667,
345
+ "step": 1792
346
+ },
347
+ {
348
+ "epoch": 0.15,
349
+ "learning_rate": 9.083290576363947e-07,
350
+ "loss": 0.1994,
351
+ "step": 1824
352
+ },
353
+ {
354
+ "epoch": 0.15,
355
+ "learning_rate": 9.0559261159569e-07,
356
+ "loss": 0.2568,
357
+ "step": 1856
358
+ },
359
+ {
360
+ "epoch": 0.15,
361
+ "learning_rate": 9.028561655549855e-07,
362
+ "loss": 0.2909,
363
+ "step": 1888
364
+ },
365
+ {
366
+ "epoch": 0.15,
367
+ "learning_rate": 9.001197195142807e-07,
368
+ "loss": 0.2697,
369
+ "step": 1920
370
+ },
371
+ {
372
+ "epoch": 0.16,
373
+ "learning_rate": 8.973832734735761e-07,
374
+ "loss": 0.3379,
375
+ "step": 1952
376
+ },
377
+ {
378
+ "epoch": 0.16,
379
+ "learning_rate": 8.946468274328715e-07,
380
+ "loss": 0.2866,
381
+ "step": 1984
382
+ },
383
+ {
384
+ "epoch": 0.16,
385
+ "learning_rate": 8.919103813921669e-07,
386
+ "loss": 0.2634,
387
+ "step": 2016
388
+ },
389
+ {
390
+ "epoch": 0.16,
391
+ "learning_rate": 8.891739353514622e-07,
392
+ "loss": 0.2234,
393
+ "step": 2048
394
+ },
395
+ {
396
+ "epoch": 0.17,
397
+ "learning_rate": 8.864374893107576e-07,
398
+ "loss": 0.2541,
399
+ "step": 2080
400
+ },
401
+ {
402
+ "epoch": 0.17,
403
+ "learning_rate": 8.83701043270053e-07,
404
+ "loss": 0.2341,
405
+ "step": 2112
406
+ },
407
+ {
408
+ "epoch": 0.17,
409
+ "learning_rate": 8.809645972293484e-07,
410
+ "loss": 0.2602,
411
+ "step": 2144
412
+ },
413
+ {
414
+ "epoch": 0.17,
415
+ "learning_rate": 8.782281511886437e-07,
416
+ "loss": 0.2602,
417
+ "step": 2176
418
+ },
419
+ {
420
+ "epoch": 0.18,
421
+ "learning_rate": 8.75491705147939e-07,
422
+ "loss": 0.2036,
423
+ "step": 2208
424
+ },
425
+ {
426
+ "epoch": 0.18,
427
+ "learning_rate": 8.727552591072344e-07,
428
+ "loss": 0.2342,
429
+ "step": 2240
430
+ },
431
+ {
432
+ "epoch": 0.18,
433
+ "learning_rate": 8.700188130665298e-07,
434
+ "loss": 0.2361,
435
+ "step": 2272
436
+ },
437
+ {
438
+ "epoch": 0.19,
439
+ "learning_rate": 8.672823670258251e-07,
440
+ "loss": 0.3299,
441
+ "step": 2304
442
+ },
443
+ {
444
+ "epoch": 0.19,
445
+ "learning_rate": 8.645459209851206e-07,
446
+ "loss": 0.3221,
447
+ "step": 2336
448
+ },
449
+ {
450
+ "epoch": 0.19,
451
+ "learning_rate": 8.618094749444159e-07,
452
+ "loss": 0.2119,
453
+ "step": 2368
454
+ },
455
+ {
456
+ "epoch": 0.19,
457
+ "learning_rate": 8.590730289037113e-07,
458
+ "loss": 0.1908,
459
+ "step": 2400
460
+ },
461
+ {
462
+ "epoch": 0.2,
463
+ "learning_rate": 8.563365828630066e-07,
464
+ "loss": 0.2736,
465
+ "step": 2432
466
+ },
467
+ {
468
+ "epoch": 0.2,
469
+ "learning_rate": 8.536001368223021e-07,
470
+ "loss": 0.1713,
471
+ "step": 2464
472
+ },
473
+ {
474
+ "epoch": 0.2,
475
+ "learning_rate": 8.508636907815974e-07,
476
+ "loss": 0.2658,
477
+ "step": 2496
478
+ },
479
+ {
480
+ "epoch": 0.2,
481
+ "learning_rate": 8.481272447408928e-07,
482
+ "loss": 0.2235,
483
+ "step": 2528
484
+ },
485
+ {
486
+ "epoch": 0.21,
487
+ "learning_rate": 8.453907987001881e-07,
488
+ "loss": 0.1858,
489
+ "step": 2560
490
+ },
491
+ {
492
+ "epoch": 0.21,
493
+ "learning_rate": 8.426543526594834e-07,
494
+ "loss": 0.2935,
495
+ "step": 2592
496
+ },
497
+ {
498
+ "epoch": 0.21,
499
+ "learning_rate": 8.399179066187788e-07,
500
+ "loss": 0.1996,
501
+ "step": 2624
502
+ },
503
+ {
504
+ "epoch": 0.21,
505
+ "learning_rate": 8.371814605780741e-07,
506
+ "loss": 0.2209,
507
+ "step": 2656
508
+ },
509
+ {
510
+ "epoch": 0.22,
511
+ "learning_rate": 8.344450145373696e-07,
512
+ "loss": 0.1611,
513
+ "step": 2688
514
+ },
515
+ {
516
+ "epoch": 0.22,
517
+ "learning_rate": 8.317085684966649e-07,
518
+ "loss": 0.28,
519
+ "step": 2720
520
+ },
521
+ {
522
+ "epoch": 0.22,
523
+ "learning_rate": 8.289721224559603e-07,
524
+ "loss": 0.2486,
525
+ "step": 2752
526
+ },
527
+ {
528
+ "epoch": 0.22,
529
+ "learning_rate": 8.262356764152556e-07,
530
+ "loss": 0.1978,
531
+ "step": 2784
532
+ },
533
+ {
534
+ "epoch": 0.23,
535
+ "learning_rate": 8.234992303745511e-07,
536
+ "loss": 0.2535,
537
+ "step": 2816
538
+ },
539
+ {
540
+ "epoch": 0.23,
541
+ "learning_rate": 8.207627843338464e-07,
542
+ "loss": 0.2666,
543
+ "step": 2848
544
+ },
545
+ {
546
+ "epoch": 0.23,
547
+ "learning_rate": 8.180263382931417e-07,
548
+ "loss": 0.1769,
549
+ "step": 2880
550
+ },
551
+ {
552
+ "epoch": 0.23,
553
+ "learning_rate": 8.152898922524371e-07,
554
+ "loss": 0.2803,
555
+ "step": 2912
556
+ },
557
+ {
558
+ "epoch": 0.24,
559
+ "learning_rate": 8.125534462117325e-07,
560
+ "loss": 0.2129,
561
+ "step": 2944
562
+ },
563
+ {
564
+ "epoch": 0.24,
565
+ "learning_rate": 8.098170001710278e-07,
566
+ "loss": 0.2255,
567
+ "step": 2976
568
+ },
569
+ {
570
+ "epoch": 0.24,
571
+ "learning_rate": 8.070805541303232e-07,
572
+ "loss": 0.1739,
573
+ "step": 3008
574
+ },
575
+ {
576
+ "epoch": 0.24,
577
+ "learning_rate": 8.043441080896186e-07,
578
+ "loss": 0.2321,
579
+ "step": 3040
580
+ },
581
+ {
582
+ "epoch": 0.25,
583
+ "learning_rate": 8.01607662048914e-07,
584
+ "loss": 0.2761,
585
+ "step": 3072
586
+ },
587
+ {
588
+ "epoch": 0.25,
589
+ "learning_rate": 7.988712160082093e-07,
590
+ "loss": 0.2867,
591
+ "step": 3104
592
+ },
593
+ {
594
+ "epoch": 0.25,
595
+ "learning_rate": 7.961347699675047e-07,
596
+ "loss": 0.1763,
597
+ "step": 3136
598
+ },
599
+ {
600
+ "epoch": 0.25,
601
+ "learning_rate": 7.933983239268001e-07,
602
+ "loss": 0.325,
603
+ "step": 3168
604
+ },
605
+ {
606
+ "epoch": 0.26,
607
+ "learning_rate": 7.906618778860953e-07,
608
+ "loss": 0.2515,
609
+ "step": 3200
610
+ },
611
+ {
612
+ "epoch": 0.26,
613
+ "learning_rate": 7.879254318453907e-07,
614
+ "loss": 0.1741,
615
+ "step": 3232
616
+ },
617
+ {
618
+ "epoch": 0.26,
619
+ "learning_rate": 7.851889858046861e-07,
620
+ "loss": 0.1999,
621
+ "step": 3264
622
+ },
623
+ {
624
+ "epoch": 0.26,
625
+ "learning_rate": 7.824525397639815e-07,
626
+ "loss": 0.2393,
627
+ "step": 3296
628
+ },
629
+ {
630
+ "epoch": 0.27,
631
+ "learning_rate": 7.797160937232768e-07,
632
+ "loss": 0.2242,
633
+ "step": 3328
634
+ },
635
+ {
636
+ "epoch": 0.27,
637
+ "learning_rate": 7.769796476825722e-07,
638
+ "loss": 0.1877,
639
+ "step": 3360
640
+ },
641
+ {
642
+ "epoch": 0.27,
643
+ "learning_rate": 7.742432016418676e-07,
644
+ "loss": 0.194,
645
+ "step": 3392
646
+ },
647
+ {
648
+ "epoch": 0.28,
649
+ "learning_rate": 7.71506755601163e-07,
650
+ "loss": 0.2499,
651
+ "step": 3424
652
+ },
653
+ {
654
+ "epoch": 0.28,
655
+ "learning_rate": 7.687703095604583e-07,
656
+ "loss": 0.2496,
657
+ "step": 3456
658
+ },
659
+ {
660
+ "epoch": 0.28,
661
+ "learning_rate": 7.660338635197537e-07,
662
+ "loss": 0.1899,
663
+ "step": 3488
664
+ },
665
+ {
666
+ "epoch": 0.28,
667
+ "learning_rate": 7.63297417479049e-07,
668
+ "loss": 0.1866,
669
+ "step": 3520
670
+ },
671
+ {
672
+ "epoch": 0.29,
673
+ "learning_rate": 7.605609714383444e-07,
674
+ "loss": 0.1843,
675
+ "step": 3552
676
+ },
677
+ {
678
+ "epoch": 0.29,
679
+ "learning_rate": 7.578245253976397e-07,
680
+ "loss": 0.1991,
681
+ "step": 3584
682
+ },
683
+ {
684
+ "epoch": 0.29,
685
+ "learning_rate": 7.550880793569352e-07,
686
+ "loss": 0.2122,
687
+ "step": 3616
688
+ },
689
+ {
690
+ "epoch": 0.29,
691
+ "learning_rate": 7.523516333162305e-07,
692
+ "loss": 0.2423,
693
+ "step": 3648
694
+ },
695
+ {
696
+ "epoch": 0.3,
697
+ "learning_rate": 7.496151872755259e-07,
698
+ "loss": 0.2568,
699
+ "step": 3680
700
+ },
701
+ {
702
+ "epoch": 0.3,
703
+ "learning_rate": 7.468787412348212e-07,
704
+ "loss": 0.2727,
705
+ "step": 3712
706
+ },
707
+ {
708
+ "epoch": 0.3,
709
+ "learning_rate": 7.441422951941167e-07,
710
+ "loss": 0.1825,
711
+ "step": 3744
712
+ },
713
+ {
714
+ "epoch": 0.3,
715
+ "learning_rate": 7.41405849153412e-07,
716
+ "loss": 0.1573,
717
+ "step": 3776
718
+ },
719
+ {
720
+ "epoch": 0.31,
721
+ "learning_rate": 7.386694031127074e-07,
722
+ "loss": 0.2034,
723
+ "step": 3808
724
+ },
725
+ {
726
+ "epoch": 0.31,
727
+ "learning_rate": 7.359329570720028e-07,
728
+ "loss": 0.1514,
729
+ "step": 3840
730
+ },
731
+ {
732
+ "epoch": 0.31,
733
+ "learning_rate": 7.33196511031298e-07,
734
+ "loss": 0.2618,
735
+ "step": 3872
736
+ },
737
+ {
738
+ "epoch": 0.31,
739
+ "learning_rate": 7.304600649905934e-07,
740
+ "loss": 0.244,
741
+ "step": 3904
742
+ },
743
+ {
744
+ "epoch": 0.32,
745
+ "learning_rate": 7.277236189498887e-07,
746
+ "loss": 0.1753,
747
+ "step": 3936
748
+ },
749
+ {
750
+ "epoch": 0.32,
751
+ "learning_rate": 7.249871729091842e-07,
752
+ "loss": 0.2044,
753
+ "step": 3968
754
+ },
755
+ {
756
+ "epoch": 0.32,
757
+ "learning_rate": 7.222507268684795e-07,
758
+ "loss": 0.1882,
759
+ "step": 4000
760
+ },
761
+ {
762
+ "epoch": 0.32,
763
+ "learning_rate": 7.195142808277749e-07,
764
+ "loss": 0.2397,
765
+ "step": 4032
766
+ },
767
+ {
768
+ "epoch": 0.33,
769
+ "learning_rate": 7.167778347870702e-07,
770
+ "loss": 0.2084,
771
+ "step": 4064
772
+ },
773
+ {
774
+ "epoch": 0.33,
775
+ "learning_rate": 7.140413887463657e-07,
776
+ "loss": 0.2635,
777
+ "step": 4096
778
+ },
779
+ {
780
+ "epoch": 0.33,
781
+ "learning_rate": 7.11304942705661e-07,
782
+ "loss": 0.2512,
783
+ "step": 4128
784
+ },
785
+ {
786
+ "epoch": 0.33,
787
+ "learning_rate": 7.085684966649563e-07,
788
+ "loss": 0.2411,
789
+ "step": 4160
790
+ },
791
+ {
792
+ "epoch": 0.34,
793
+ "learning_rate": 7.058320506242517e-07,
794
+ "loss": 0.1846,
795
+ "step": 4192
796
+ },
797
+ {
798
+ "epoch": 0.34,
799
+ "learning_rate": 7.030956045835471e-07,
800
+ "loss": 0.1447,
801
+ "step": 4224
802
+ },
803
+ {
804
+ "epoch": 0.34,
805
+ "learning_rate": 7.003591585428424e-07,
806
+ "loss": 0.2373,
807
+ "step": 4256
808
+ },
809
+ {
810
+ "epoch": 0.34,
811
+ "learning_rate": 6.976227125021378e-07,
812
+ "loss": 0.2097,
813
+ "step": 4288
814
+ },
815
+ {
816
+ "epoch": 0.35,
817
+ "learning_rate": 6.948862664614332e-07,
818
+ "loss": 0.2756,
819
+ "step": 4320
820
+ },
821
+ {
822
+ "epoch": 0.35,
823
+ "learning_rate": 6.922353343595006e-07,
824
+ "loss": 0.191,
825
+ "step": 4352
826
+ },
827
+ {
828
+ "epoch": 0.35,
829
+ "learning_rate": 6.894988883187959e-07,
830
+ "loss": 0.2076,
831
+ "step": 4384
832
+ },
833
+ {
834
+ "epoch": 0.35,
835
+ "learning_rate": 6.867624422780914e-07,
836
+ "loss": 0.2787,
837
+ "step": 4416
838
+ },
839
+ {
840
+ "epoch": 0.36,
841
+ "learning_rate": 6.840259962373867e-07,
842
+ "loss": 0.1894,
843
+ "step": 4448
844
+ },
845
+ {
846
+ "epoch": 0.36,
847
+ "learning_rate": 6.81289550196682e-07,
848
+ "loss": 0.1423,
849
+ "step": 4480
850
+ },
851
+ {
852
+ "epoch": 0.36,
853
+ "learning_rate": 6.785531041559774e-07,
854
+ "loss": 0.1738,
855
+ "step": 4512
856
+ },
857
+ {
858
+ "epoch": 0.37,
859
+ "learning_rate": 6.758166581152727e-07,
860
+ "loss": 0.2598,
861
+ "step": 4544
862
+ },
863
+ {
864
+ "epoch": 0.37,
865
+ "learning_rate": 6.730802120745681e-07,
866
+ "loss": 0.2753,
867
+ "step": 4576
868
+ },
869
+ {
870
+ "epoch": 0.37,
871
+ "learning_rate": 6.703437660338634e-07,
872
+ "loss": 0.2922,
873
+ "step": 4608
874
+ },
875
+ {
876
+ "epoch": 0.37,
877
+ "learning_rate": 6.676073199931589e-07,
878
+ "loss": 0.172,
879
+ "step": 4640
880
+ },
881
+ {
882
+ "epoch": 0.38,
883
+ "learning_rate": 6.648708739524542e-07,
884
+ "loss": 0.2269,
885
+ "step": 4672
886
+ },
887
+ {
888
+ "epoch": 0.38,
889
+ "learning_rate": 6.621344279117496e-07,
890
+ "loss": 0.2662,
891
+ "step": 4704
892
+ },
893
+ {
894
+ "epoch": 0.38,
895
+ "learning_rate": 6.59397981871045e-07,
896
+ "loss": 0.2674,
897
+ "step": 4736
898
+ },
899
+ {
900
+ "epoch": 0.38,
901
+ "learning_rate": 6.566615358303404e-07,
902
+ "loss": 0.2803,
903
+ "step": 4768
904
+ },
905
+ {
906
+ "epoch": 0.39,
907
+ "learning_rate": 6.539250897896357e-07,
908
+ "loss": 0.2253,
909
+ "step": 4800
910
+ },
911
+ {
912
+ "epoch": 0.39,
913
+ "learning_rate": 6.51188643748931e-07,
914
+ "loss": 0.2816,
915
+ "step": 4832
916
+ },
917
+ {
918
+ "epoch": 0.39,
919
+ "learning_rate": 6.484521977082264e-07,
920
+ "loss": 0.1596,
921
+ "step": 4864
922
+ },
923
+ {
924
+ "epoch": 0.39,
925
+ "learning_rate": 6.457157516675218e-07,
926
+ "loss": 0.2419,
927
+ "step": 4896
928
+ },
929
+ {
930
+ "epoch": 0.4,
931
+ "learning_rate": 6.429793056268171e-07,
932
+ "loss": 0.2344,
933
+ "step": 4928
934
+ },
935
+ {
936
+ "epoch": 0.4,
937
+ "learning_rate": 6.402428595861125e-07,
938
+ "loss": 0.2008,
939
+ "step": 4960
940
+ },
941
+ {
942
+ "epoch": 0.4,
943
+ "learning_rate": 6.375064135454079e-07,
944
+ "loss": 0.2291,
945
+ "step": 4992
946
+ },
947
+ {
948
+ "epoch": 0.4,
949
+ "learning_rate": 6.347699675047033e-07,
950
+ "loss": 0.1661,
951
+ "step": 5024
952
+ },
953
+ {
954
+ "epoch": 0.41,
955
+ "learning_rate": 6.320335214639986e-07,
956
+ "loss": 0.2022,
957
+ "step": 5056
958
+ },
959
+ {
960
+ "epoch": 0.41,
961
+ "learning_rate": 6.292970754232941e-07,
962
+ "loss": 0.2047,
963
+ "step": 5088
964
+ },
965
+ {
966
+ "epoch": 0.41,
967
+ "learning_rate": 6.265606293825894e-07,
968
+ "loss": 0.2004,
969
+ "step": 5120
970
+ },
971
+ {
972
+ "epoch": 0.41,
973
+ "learning_rate": 6.238241833418846e-07,
974
+ "loss": 0.1987,
975
+ "step": 5152
976
+ },
977
+ {
978
+ "epoch": 0.42,
979
+ "learning_rate": 6.2108773730118e-07,
980
+ "loss": 0.2315,
981
+ "step": 5184
982
+ },
983
+ {
984
+ "epoch": 0.42,
985
+ "learning_rate": 6.183512912604754e-07,
986
+ "loss": 0.2086,
987
+ "step": 5216
988
+ },
989
+ {
990
+ "epoch": 0.42,
991
+ "learning_rate": 6.156148452197708e-07,
992
+ "loss": 0.1717,
993
+ "step": 5248
994
+ },
995
+ {
996
+ "epoch": 0.42,
997
+ "learning_rate": 6.128783991790661e-07,
998
+ "loss": 0.2717,
999
+ "step": 5280
1000
+ },
1001
+ {
1002
+ "epoch": 0.43,
1003
+ "learning_rate": 6.101419531383615e-07,
1004
+ "loss": 0.2229,
1005
+ "step": 5312
1006
+ },
1007
+ {
1008
+ "epoch": 0.43,
1009
+ "learning_rate": 6.074055070976569e-07,
1010
+ "loss": 0.1996,
1011
+ "step": 5344
1012
+ },
1013
+ {
1014
+ "epoch": 0.43,
1015
+ "learning_rate": 6.046690610569523e-07,
1016
+ "loss": 0.2549,
1017
+ "step": 5376
1018
+ },
1019
+ {
1020
+ "epoch": 0.43,
1021
+ "learning_rate": 6.019326150162476e-07,
1022
+ "loss": 0.2378,
1023
+ "step": 5408
1024
+ },
1025
+ {
1026
+ "epoch": 0.44,
1027
+ "learning_rate": 5.991961689755431e-07,
1028
+ "loss": 0.2488,
1029
+ "step": 5440
1030
+ },
1031
+ {
1032
+ "epoch": 0.44,
1033
+ "learning_rate": 5.964597229348383e-07,
1034
+ "loss": 0.2459,
1035
+ "step": 5472
1036
+ },
1037
+ {
1038
+ "epoch": 0.44,
1039
+ "learning_rate": 5.937232768941337e-07,
1040
+ "loss": 0.1928,
1041
+ "step": 5504
1042
+ },
1043
+ {
1044
+ "epoch": 0.45,
1045
+ "learning_rate": 5.90986830853429e-07,
1046
+ "loss": 0.2617,
1047
+ "step": 5536
1048
+ },
1049
+ {
1050
+ "epoch": 0.45,
1051
+ "learning_rate": 5.882503848127245e-07,
1052
+ "loss": 0.1927,
1053
+ "step": 5568
1054
+ },
1055
+ {
1056
+ "epoch": 0.45,
1057
+ "learning_rate": 5.855139387720198e-07,
1058
+ "loss": 0.1943,
1059
+ "step": 5600
1060
+ },
1061
+ {
1062
+ "epoch": 0.45,
1063
+ "learning_rate": 5.827774927313152e-07,
1064
+ "loss": 0.241,
1065
+ "step": 5632
1066
+ },
1067
+ {
1068
+ "epoch": 0.46,
1069
+ "learning_rate": 5.800410466906105e-07,
1070
+ "loss": 0.2354,
1071
+ "step": 5664
1072
+ },
1073
+ {
1074
+ "epoch": 0.46,
1075
+ "learning_rate": 5.77304600649906e-07,
1076
+ "loss": 0.2227,
1077
+ "step": 5696
1078
+ },
1079
+ {
1080
+ "epoch": 0.46,
1081
+ "learning_rate": 5.745681546092013e-07,
1082
+ "loss": 0.2395,
1083
+ "step": 5728
1084
+ },
1085
+ {
1086
+ "epoch": 0.46,
1087
+ "learning_rate": 5.718317085684967e-07,
1088
+ "loss": 0.1895,
1089
+ "step": 5760
1090
+ },
1091
+ {
1092
+ "epoch": 0.47,
1093
+ "learning_rate": 5.690952625277921e-07,
1094
+ "loss": 0.1621,
1095
+ "step": 5792
1096
+ },
1097
+ {
1098
+ "epoch": 0.47,
1099
+ "learning_rate": 5.663588164870873e-07,
1100
+ "loss": 0.2169,
1101
+ "step": 5824
1102
+ },
1103
+ {
1104
+ "epoch": 0.47,
1105
+ "learning_rate": 5.636223704463827e-07,
1106
+ "loss": 0.1979,
1107
+ "step": 5856
1108
+ },
1109
+ {
1110
+ "epoch": 0.47,
1111
+ "learning_rate": 5.60885924405678e-07,
1112
+ "loss": 0.206,
1113
+ "step": 5888
1114
+ },
1115
+ {
1116
+ "epoch": 0.48,
1117
+ "learning_rate": 5.581494783649735e-07,
1118
+ "loss": 0.1625,
1119
+ "step": 5920
1120
+ },
1121
+ {
1122
+ "epoch": 0.48,
1123
+ "learning_rate": 5.554130323242688e-07,
1124
+ "loss": 0.2625,
1125
+ "step": 5952
1126
+ },
1127
+ {
1128
+ "epoch": 0.48,
1129
+ "learning_rate": 5.526765862835642e-07,
1130
+ "loss": 0.1797,
1131
+ "step": 5984
1132
+ },
1133
+ {
1134
+ "epoch": 0.48,
1135
+ "learning_rate": 5.499401402428595e-07,
1136
+ "loss": 0.2056,
1137
+ "step": 6016
1138
+ },
1139
+ {
1140
+ "epoch": 0.49,
1141
+ "learning_rate": 5.47203694202155e-07,
1142
+ "loss": 0.1659,
1143
+ "step": 6048
1144
+ },
1145
+ {
1146
+ "epoch": 0.49,
1147
+ "learning_rate": 5.444672481614503e-07,
1148
+ "loss": 0.1524,
1149
+ "step": 6080
1150
+ },
1151
+ {
1152
+ "epoch": 0.49,
1153
+ "learning_rate": 5.417308021207456e-07,
1154
+ "loss": 0.236,
1155
+ "step": 6112
1156
+ },
1157
+ {
1158
+ "epoch": 0.49,
1159
+ "learning_rate": 5.38994356080041e-07,
1160
+ "loss": 0.1921,
1161
+ "step": 6144
1162
+ },
1163
+ {
1164
+ "epoch": 0.5,
1165
+ "learning_rate": 5.362579100393364e-07,
1166
+ "loss": 0.217,
1167
+ "step": 6176
1168
+ },
1169
+ {
1170
+ "epoch": 0.5,
1171
+ "learning_rate": 5.335214639986317e-07,
1172
+ "loss": 0.1996,
1173
+ "step": 6208
1174
+ },
1175
+ {
1176
+ "epoch": 0.5,
1177
+ "learning_rate": 5.307850179579271e-07,
1178
+ "loss": 0.1467,
1179
+ "step": 6240
1180
+ },
1181
+ {
1182
+ "epoch": 0.5,
1183
+ "learning_rate": 5.280485719172225e-07,
1184
+ "loss": 0.2702,
1185
+ "step": 6272
1186
+ },
1187
+ {
1188
+ "epoch": 0.51,
1189
+ "learning_rate": 5.253121258765179e-07,
1190
+ "loss": 0.1736,
1191
+ "step": 6304
1192
+ },
1193
+ {
1194
+ "epoch": 0.51,
1195
+ "learning_rate": 5.225756798358132e-07,
1196
+ "loss": 0.2281,
1197
+ "step": 6336
1198
+ },
1199
+ {
1200
+ "epoch": 0.51,
1201
+ "learning_rate": 5.198392337951087e-07,
1202
+ "loss": 0.2114,
1203
+ "step": 6368
1204
+ },
1205
+ {
1206
+ "epoch": 0.51,
1207
+ "learning_rate": 5.17102787754404e-07,
1208
+ "loss": 0.2252,
1209
+ "step": 6400
1210
+ },
1211
+ {
1212
+ "epoch": 0.52,
1213
+ "learning_rate": 5.143663417136994e-07,
1214
+ "loss": 0.163,
1215
+ "step": 6432
1216
+ },
1217
+ {
1218
+ "epoch": 0.52,
1219
+ "learning_rate": 5.116298956729946e-07,
1220
+ "loss": 0.266,
1221
+ "step": 6464
1222
+ },
1223
+ {
1224
+ "epoch": 0.52,
1225
+ "learning_rate": 5.0889344963229e-07,
1226
+ "loss": 0.2154,
1227
+ "step": 6496
1228
+ },
1229
+ {
1230
+ "epoch": 0.52,
1231
+ "learning_rate": 5.061570035915854e-07,
1232
+ "loss": 0.2003,
1233
+ "step": 6528
1234
+ },
1235
+ {
1236
+ "epoch": 0.53,
1237
+ "learning_rate": 5.034205575508807e-07,
1238
+ "loss": 0.1969,
1239
+ "step": 6560
1240
+ },
1241
+ {
1242
+ "epoch": 0.53,
1243
+ "learning_rate": 5.006841115101761e-07,
1244
+ "loss": 0.2436,
1245
+ "step": 6592
1246
+ },
1247
+ {
1248
+ "epoch": 0.53,
1249
+ "learning_rate": 4.979476654694714e-07,
1250
+ "loss": 0.134,
1251
+ "step": 6624
1252
+ },
1253
+ {
1254
+ "epoch": 0.54,
1255
+ "learning_rate": 4.952112194287669e-07,
1256
+ "loss": 0.2246,
1257
+ "step": 6656
1258
+ },
1259
+ {
1260
+ "epoch": 0.54,
1261
+ "learning_rate": 4.924747733880622e-07,
1262
+ "loss": 0.1873,
1263
+ "step": 6688
1264
+ },
1265
+ {
1266
+ "epoch": 0.54,
1267
+ "learning_rate": 4.897383273473576e-07,
1268
+ "loss": 0.1549,
1269
+ "step": 6720
1270
+ },
1271
+ {
1272
+ "epoch": 0.54,
1273
+ "learning_rate": 4.87001881306653e-07,
1274
+ "loss": 0.2518,
1275
+ "step": 6752
1276
+ },
1277
+ {
1278
+ "epoch": 0.55,
1279
+ "learning_rate": 4.842654352659483e-07,
1280
+ "loss": 0.2676,
1281
+ "step": 6784
1282
+ },
1283
+ {
1284
+ "epoch": 0.55,
1285
+ "learning_rate": 4.815289892252437e-07,
1286
+ "loss": 0.1537,
1287
+ "step": 6816
1288
+ },
1289
+ {
1290
+ "epoch": 0.55,
1291
+ "learning_rate": 4.787925431845391e-07,
1292
+ "loss": 0.1925,
1293
+ "step": 6848
1294
+ },
1295
+ {
1296
+ "epoch": 0.55,
1297
+ "learning_rate": 4.760560971438344e-07,
1298
+ "loss": 0.2157,
1299
+ "step": 6880
1300
+ },
1301
+ {
1302
+ "epoch": 0.56,
1303
+ "learning_rate": 4.7331965110312977e-07,
1304
+ "loss": 0.1923,
1305
+ "step": 6912
1306
+ },
1307
+ {
1308
+ "epoch": 0.56,
1309
+ "learning_rate": 4.7058320506242517e-07,
1310
+ "loss": 0.1922,
1311
+ "step": 6944
1312
+ },
1313
+ {
1314
+ "epoch": 0.56,
1315
+ "learning_rate": 4.678467590217205e-07,
1316
+ "loss": 0.257,
1317
+ "step": 6976
1318
+ },
1319
+ {
1320
+ "epoch": 0.56,
1321
+ "learning_rate": 4.6511031298101586e-07,
1322
+ "loss": 0.1744,
1323
+ "step": 7008
1324
+ },
1325
+ {
1326
+ "epoch": 0.57,
1327
+ "learning_rate": 4.6237386694031126e-07,
1328
+ "loss": 0.2281,
1329
+ "step": 7040
1330
+ },
1331
+ {
1332
+ "epoch": 0.57,
1333
+ "learning_rate": 4.596374208996066e-07,
1334
+ "loss": 0.2135,
1335
+ "step": 7072
1336
+ },
1337
+ {
1338
+ "epoch": 0.57,
1339
+ "learning_rate": 4.56900974858902e-07,
1340
+ "loss": 0.1841,
1341
+ "step": 7104
1342
+ },
1343
+ {
1344
+ "epoch": 0.57,
1345
+ "learning_rate": 4.5416452881819735e-07,
1346
+ "loss": 0.2652,
1347
+ "step": 7136
1348
+ },
1349
+ {
1350
+ "epoch": 0.58,
1351
+ "learning_rate": 4.5142808277749275e-07,
1352
+ "loss": 0.225,
1353
+ "step": 7168
1354
+ },
1355
+ {
1356
+ "epoch": 0.58,
1357
+ "learning_rate": 4.4869163673678804e-07,
1358
+ "loss": 0.2377,
1359
+ "step": 7200
1360
+ },
1361
+ {
1362
+ "epoch": 0.58,
1363
+ "learning_rate": 4.4595519069608344e-07,
1364
+ "loss": 0.2323,
1365
+ "step": 7232
1366
+ },
1367
+ {
1368
+ "epoch": 0.58,
1369
+ "learning_rate": 4.432187446553788e-07,
1370
+ "loss": 0.2227,
1371
+ "step": 7264
1372
+ },
1373
+ {
1374
+ "epoch": 0.59,
1375
+ "learning_rate": 4.404822986146742e-07,
1376
+ "loss": 0.2968,
1377
+ "step": 7296
1378
+ },
1379
+ {
1380
+ "epoch": 0.59,
1381
+ "learning_rate": 4.377458525739695e-07,
1382
+ "loss": 0.2623,
1383
+ "step": 7328
1384
+ },
1385
+ {
1386
+ "epoch": 0.59,
1387
+ "learning_rate": 4.350094065332649e-07,
1388
+ "loss": 0.1943,
1389
+ "step": 7360
1390
+ },
1391
+ {
1392
+ "epoch": 0.59,
1393
+ "learning_rate": 4.322729604925603e-07,
1394
+ "loss": 0.158,
1395
+ "step": 7392
1396
+ },
1397
+ {
1398
+ "epoch": 0.6,
1399
+ "learning_rate": 4.2953651445185563e-07,
1400
+ "loss": 0.2413,
1401
+ "step": 7424
1402
+ },
1403
+ {
1404
+ "epoch": 0.6,
1405
+ "learning_rate": 4.2680006841115103e-07,
1406
+ "loss": 0.2413,
1407
+ "step": 7456
1408
+ },
1409
+ {
1410
+ "epoch": 0.6,
1411
+ "learning_rate": 4.240636223704464e-07,
1412
+ "loss": 0.1793,
1413
+ "step": 7488
1414
+ },
1415
+ {
1416
+ "epoch": 0.6,
1417
+ "learning_rate": 4.213271763297417e-07,
1418
+ "loss": 0.2505,
1419
+ "step": 7520
1420
+ },
1421
+ {
1422
+ "epoch": 0.61,
1423
+ "learning_rate": 4.1859073028903707e-07,
1424
+ "loss": 0.2078,
1425
+ "step": 7552
1426
+ },
1427
+ {
1428
+ "epoch": 0.61,
1429
+ "learning_rate": 4.1585428424833247e-07,
1430
+ "loss": 0.2603,
1431
+ "step": 7584
1432
+ },
1433
+ {
1434
+ "epoch": 0.61,
1435
+ "learning_rate": 4.131178382076278e-07,
1436
+ "loss": 0.1426,
1437
+ "step": 7616
1438
+ },
1439
+ {
1440
+ "epoch": 0.61,
1441
+ "learning_rate": 4.103813921669232e-07,
1442
+ "loss": 0.1823,
1443
+ "step": 7648
1444
+ },
1445
+ {
1446
+ "epoch": 0.62,
1447
+ "learning_rate": 4.0764494612621856e-07,
1448
+ "loss": 0.2215,
1449
+ "step": 7680
1450
+ },
1451
+ {
1452
+ "epoch": 0.62,
1453
+ "learning_rate": 4.049085000855139e-07,
1454
+ "loss": 0.2956,
1455
+ "step": 7712
1456
+ },
1457
+ {
1458
+ "epoch": 0.62,
1459
+ "learning_rate": 4.021720540448093e-07,
1460
+ "loss": 0.1766,
1461
+ "step": 7744
1462
+ },
1463
+ {
1464
+ "epoch": 0.63,
1465
+ "learning_rate": 3.9943560800410465e-07,
1466
+ "loss": 0.2124,
1467
+ "step": 7776
1468
+ },
1469
+ {
1470
+ "epoch": 0.63,
1471
+ "learning_rate": 3.9669916196340005e-07,
1472
+ "loss": 0.2123,
1473
+ "step": 7808
1474
+ },
1475
+ {
1476
+ "epoch": 0.63,
1477
+ "learning_rate": 3.9396271592269535e-07,
1478
+ "loss": 0.2172,
1479
+ "step": 7840
1480
+ },
1481
+ {
1482
+ "epoch": 0.63,
1483
+ "learning_rate": 3.9122626988199075e-07,
1484
+ "loss": 0.1737,
1485
+ "step": 7872
1486
+ },
1487
+ {
1488
+ "epoch": 0.64,
1489
+ "learning_rate": 3.884898238412861e-07,
1490
+ "loss": 0.2504,
1491
+ "step": 7904
1492
+ },
1493
+ {
1494
+ "epoch": 0.64,
1495
+ "learning_rate": 3.857533778005815e-07,
1496
+ "loss": 0.2242,
1497
+ "step": 7936
1498
+ },
1499
+ {
1500
+ "epoch": 0.64,
1501
+ "learning_rate": 3.8301693175987684e-07,
1502
+ "loss": 0.205,
1503
+ "step": 7968
1504
+ },
1505
+ {
1506
+ "epoch": 0.64,
1507
+ "learning_rate": 3.802804857191722e-07,
1508
+ "loss": 0.2286,
1509
+ "step": 8000
1510
+ },
1511
+ {
1512
+ "epoch": 0.65,
1513
+ "learning_rate": 3.775440396784676e-07,
1514
+ "loss": 0.1848,
1515
+ "step": 8032
1516
+ },
1517
+ {
1518
+ "epoch": 0.65,
1519
+ "learning_rate": 3.7480759363776293e-07,
1520
+ "loss": 0.2065,
1521
+ "step": 8064
1522
+ },
1523
+ {
1524
+ "epoch": 0.65,
1525
+ "learning_rate": 3.7207114759705833e-07,
1526
+ "loss": 0.1688,
1527
+ "step": 8096
1528
+ },
1529
+ {
1530
+ "epoch": 0.65,
1531
+ "learning_rate": 3.693347015563537e-07,
1532
+ "loss": 0.2489,
1533
+ "step": 8128
1534
+ },
1535
+ {
1536
+ "epoch": 0.66,
1537
+ "learning_rate": 3.66598255515649e-07,
1538
+ "loss": 0.2309,
1539
+ "step": 8160
1540
+ },
1541
+ {
1542
+ "epoch": 0.66,
1543
+ "learning_rate": 3.6386180947494437e-07,
1544
+ "loss": 0.1236,
1545
+ "step": 8192
1546
+ },
1547
+ {
1548
+ "epoch": 0.66,
1549
+ "learning_rate": 3.6112536343423977e-07,
1550
+ "loss": 0.2242,
1551
+ "step": 8224
1552
+ },
1553
+ {
1554
+ "epoch": 0.66,
1555
+ "learning_rate": 3.583889173935351e-07,
1556
+ "loss": 0.221,
1557
+ "step": 8256
1558
+ },
1559
+ {
1560
+ "epoch": 0.67,
1561
+ "learning_rate": 3.556524713528305e-07,
1562
+ "loss": 0.1663,
1563
+ "step": 8288
1564
+ },
1565
+ {
1566
+ "epoch": 0.67,
1567
+ "learning_rate": 3.5291602531212586e-07,
1568
+ "loss": 0.2063,
1569
+ "step": 8320
1570
+ },
1571
+ {
1572
+ "epoch": 0.67,
1573
+ "learning_rate": 3.501795792714212e-07,
1574
+ "loss": 0.2651,
1575
+ "step": 8352
1576
+ },
1577
+ {
1578
+ "epoch": 0.67,
1579
+ "learning_rate": 3.474431332307166e-07,
1580
+ "loss": 0.2451,
1581
+ "step": 8384
1582
+ },
1583
+ {
1584
+ "epoch": 0.68,
1585
+ "learning_rate": 3.4470668719001196e-07,
1586
+ "loss": 0.1645,
1587
+ "step": 8416
1588
+ },
1589
+ {
1590
+ "epoch": 0.68,
1591
+ "learning_rate": 3.4197024114930736e-07,
1592
+ "loss": 0.1446,
1593
+ "step": 8448
1594
+ },
1595
+ {
1596
+ "epoch": 0.68,
1597
+ "learning_rate": 3.3923379510860265e-07,
1598
+ "loss": 0.2006,
1599
+ "step": 8480
1600
+ },
1601
+ {
1602
+ "epoch": 0.68,
1603
+ "learning_rate": 3.3649734906789805e-07,
1604
+ "loss": 0.1702,
1605
+ "step": 8512
1606
+ },
1607
+ {
1608
+ "epoch": 0.69,
1609
+ "learning_rate": 3.337609030271934e-07,
1610
+ "loss": 0.2685,
1611
+ "step": 8544
1612
+ },
1613
+ {
1614
+ "epoch": 0.69,
1615
+ "learning_rate": 3.310244569864888e-07,
1616
+ "loss": 0.1944,
1617
+ "step": 8576
1618
+ },
1619
+ {
1620
+ "epoch": 0.69,
1621
+ "learning_rate": 3.2828801094578414e-07,
1622
+ "loss": 0.1461,
1623
+ "step": 8608
1624
+ },
1625
+ {
1626
+ "epoch": 0.69,
1627
+ "learning_rate": 3.255515649050795e-07,
1628
+ "loss": 0.2539,
1629
+ "step": 8640
1630
+ },
1631
+ {
1632
+ "epoch": 0.7,
1633
+ "learning_rate": 3.228151188643749e-07,
1634
+ "loss": 0.1783,
1635
+ "step": 8672
1636
+ },
1637
+ {
1638
+ "epoch": 0.7,
1639
+ "learning_rate": 3.2007867282367024e-07,
1640
+ "loss": 0.2084,
1641
+ "step": 8704
1642
+ },
1643
+ {
1644
+ "epoch": 0.7,
1645
+ "learning_rate": 3.1734222678296563e-07,
1646
+ "loss": 0.1596,
1647
+ "step": 8736
1648
+ },
1649
+ {
1650
+ "epoch": 0.7,
1651
+ "learning_rate": 3.14605780742261e-07,
1652
+ "loss": 0.1677,
1653
+ "step": 8768
1654
+ },
1655
+ {
1656
+ "epoch": 0.71,
1657
+ "learning_rate": 3.1186933470155633e-07,
1658
+ "loss": 0.1636,
1659
+ "step": 8800
1660
+ },
1661
+ {
1662
+ "epoch": 0.71,
1663
+ "learning_rate": 3.091328886608517e-07,
1664
+ "loss": 0.2336,
1665
+ "step": 8832
1666
+ },
1667
+ {
1668
+ "epoch": 0.71,
1669
+ "learning_rate": 3.063964426201471e-07,
1670
+ "loss": 0.1853,
1671
+ "step": 8864
1672
+ },
1673
+ {
1674
+ "epoch": 0.72,
1675
+ "learning_rate": 3.036599965794424e-07,
1676
+ "loss": 0.2304,
1677
+ "step": 8896
1678
+ },
1679
+ {
1680
+ "epoch": 0.72,
1681
+ "learning_rate": 3.009235505387378e-07,
1682
+ "loss": 0.2037,
1683
+ "step": 8928
1684
+ },
1685
+ {
1686
+ "epoch": 0.72,
1687
+ "learning_rate": 2.9818710449803317e-07,
1688
+ "loss": 0.2159,
1689
+ "step": 8960
1690
+ },
1691
+ {
1692
+ "epoch": 0.72,
1693
+ "learning_rate": 2.954506584573285e-07,
1694
+ "loss": 0.2512,
1695
+ "step": 8992
1696
+ },
1697
+ {
1698
+ "epoch": 0.73,
1699
+ "learning_rate": 2.927142124166239e-07,
1700
+ "loss": 0.2011,
1701
+ "step": 9024
1702
+ },
1703
+ {
1704
+ "epoch": 0.73,
1705
+ "learning_rate": 2.8997776637591926e-07,
1706
+ "loss": 0.1677,
1707
+ "step": 9056
1708
+ },
1709
+ {
1710
+ "epoch": 0.73,
1711
+ "learning_rate": 2.8724132033521466e-07,
1712
+ "loss": 0.192,
1713
+ "step": 9088
1714
+ },
1715
+ {
1716
+ "epoch": 0.73,
1717
+ "learning_rate": 2.8450487429450995e-07,
1718
+ "loss": 0.1615,
1719
+ "step": 9120
1720
+ },
1721
+ {
1722
+ "epoch": 0.74,
1723
+ "learning_rate": 2.8176842825380535e-07,
1724
+ "loss": 0.1646,
1725
+ "step": 9152
1726
+ },
1727
+ {
1728
+ "epoch": 0.74,
1729
+ "learning_rate": 2.790319822131007e-07,
1730
+ "loss": 0.2301,
1731
+ "step": 9184
1732
+ },
1733
+ {
1734
+ "epoch": 0.74,
1735
+ "learning_rate": 2.762955361723961e-07,
1736
+ "loss": 0.1663,
1737
+ "step": 9216
1738
+ },
1739
+ {
1740
+ "epoch": 0.74,
1741
+ "learning_rate": 2.7355909013169145e-07,
1742
+ "loss": 0.251,
1743
+ "step": 9248
1744
+ },
1745
+ {
1746
+ "epoch": 0.75,
1747
+ "learning_rate": 2.708226440909868e-07,
1748
+ "loss": 0.2301,
1749
+ "step": 9280
1750
+ },
1751
+ {
1752
+ "epoch": 0.75,
1753
+ "learning_rate": 2.680861980502822e-07,
1754
+ "loss": 0.2222,
1755
+ "step": 9312
1756
+ },
1757
+ {
1758
+ "epoch": 0.75,
1759
+ "learning_rate": 2.6534975200957754e-07,
1760
+ "loss": 0.1784,
1761
+ "step": 9344
1762
+ },
1763
+ {
1764
+ "epoch": 0.75,
1765
+ "learning_rate": 2.6261330596887294e-07,
1766
+ "loss": 0.1714,
1767
+ "step": 9376
1768
+ },
1769
+ {
1770
+ "epoch": 0.76,
1771
+ "learning_rate": 2.598768599281683e-07,
1772
+ "loss": 0.2258,
1773
+ "step": 9408
1774
+ },
1775
+ {
1776
+ "epoch": 0.76,
1777
+ "learning_rate": 2.571404138874637e-07,
1778
+ "loss": 0.1907,
1779
+ "step": 9440
1780
+ },
1781
+ {
1782
+ "epoch": 0.76,
1783
+ "learning_rate": 2.54403967846759e-07,
1784
+ "loss": 0.1938,
1785
+ "step": 9472
1786
+ },
1787
+ {
1788
+ "epoch": 0.76,
1789
+ "learning_rate": 2.516675218060544e-07,
1790
+ "loss": 0.1831,
1791
+ "step": 9504
1792
+ },
1793
+ {
1794
+ "epoch": 0.77,
1795
+ "learning_rate": 2.489310757653497e-07,
1796
+ "loss": 0.1833,
1797
+ "step": 9536
1798
+ },
1799
+ {
1800
+ "epoch": 0.77,
1801
+ "learning_rate": 2.461946297246451e-07,
1802
+ "loss": 0.2551,
1803
+ "step": 9568
1804
+ },
1805
+ {
1806
+ "epoch": 0.77,
1807
+ "learning_rate": 2.4345818368394047e-07,
1808
+ "loss": 0.1553,
1809
+ "step": 9600
1810
+ },
1811
+ {
1812
+ "epoch": 0.77,
1813
+ "learning_rate": 2.407217376432358e-07,
1814
+ "loss": 0.1785,
1815
+ "step": 9632
1816
+ },
1817
+ {
1818
+ "epoch": 0.78,
1819
+ "learning_rate": 2.379852916025312e-07,
1820
+ "loss": 0.2903,
1821
+ "step": 9664
1822
+ },
1823
+ {
1824
+ "epoch": 0.78,
1825
+ "learning_rate": 2.3524884556182656e-07,
1826
+ "loss": 0.2093,
1827
+ "step": 9696
1828
+ },
1829
+ {
1830
+ "epoch": 0.78,
1831
+ "learning_rate": 2.325123995211219e-07,
1832
+ "loss": 0.2865,
1833
+ "step": 9728
1834
+ },
1835
+ {
1836
+ "epoch": 0.78,
1837
+ "learning_rate": 2.2977595348041728e-07,
1838
+ "loss": 0.2033,
1839
+ "step": 9760
1840
+ },
1841
+ {
1842
+ "epoch": 0.79,
1843
+ "learning_rate": 2.2703950743971268e-07,
1844
+ "loss": 0.1929,
1845
+ "step": 9792
1846
+ },
1847
+ {
1848
+ "epoch": 0.79,
1849
+ "learning_rate": 2.2430306139900803e-07,
1850
+ "loss": 0.1968,
1851
+ "step": 9824
1852
+ },
1853
+ {
1854
+ "epoch": 0.79,
1855
+ "learning_rate": 2.215666153583034e-07,
1856
+ "loss": 0.1985,
1857
+ "step": 9856
1858
+ },
1859
+ {
1860
+ "epoch": 0.79,
1861
+ "learning_rate": 2.1883016931759877e-07,
1862
+ "loss": 0.1509,
1863
+ "step": 9888
1864
+ },
1865
+ {
1866
+ "epoch": 0.8,
1867
+ "learning_rate": 2.1609372327689412e-07,
1868
+ "loss": 0.1988,
1869
+ "step": 9920
1870
+ },
1871
+ {
1872
+ "epoch": 0.8,
1873
+ "learning_rate": 2.133572772361895e-07,
1874
+ "loss": 0.2059,
1875
+ "step": 9952
1876
+ },
1877
+ {
1878
+ "epoch": 0.8,
1879
+ "learning_rate": 2.1062083119548484e-07,
1880
+ "loss": 0.1701,
1881
+ "step": 9984
1882
+ },
1883
+ {
1884
+ "epoch": 0.81,
1885
+ "learning_rate": 2.0788438515478021e-07,
1886
+ "loss": 0.1524,
1887
+ "step": 10016
1888
+ },
1889
+ {
1890
+ "epoch": 0.81,
1891
+ "learning_rate": 2.051479391140756e-07,
1892
+ "loss": 0.2236,
1893
+ "step": 10048
1894
+ },
1895
+ {
1896
+ "epoch": 0.81,
1897
+ "learning_rate": 2.0241149307337093e-07,
1898
+ "loss": 0.189,
1899
+ "step": 10080
1900
+ },
1901
+ {
1902
+ "epoch": 0.81,
1903
+ "learning_rate": 1.9967504703266633e-07,
1904
+ "loss": 0.1464,
1905
+ "step": 10112
1906
+ },
1907
+ {
1908
+ "epoch": 0.82,
1909
+ "learning_rate": 1.9693860099196168e-07,
1910
+ "loss": 0.2245,
1911
+ "step": 10144
1912
+ },
1913
+ {
1914
+ "epoch": 0.82,
1915
+ "learning_rate": 1.9420215495125705e-07,
1916
+ "loss": 0.2619,
1917
+ "step": 10176
1918
+ },
1919
+ {
1920
+ "epoch": 0.82,
1921
+ "learning_rate": 1.9146570891055243e-07,
1922
+ "loss": 0.2331,
1923
+ "step": 10208
1924
+ },
1925
+ {
1926
+ "epoch": 0.82,
1927
+ "learning_rate": 1.8872926286984777e-07,
1928
+ "loss": 0.1925,
1929
+ "step": 10240
1930
+ },
1931
+ {
1932
+ "epoch": 0.83,
1933
+ "learning_rate": 1.8599281682914315e-07,
1934
+ "loss": 0.2707,
1935
+ "step": 10272
1936
+ },
1937
+ {
1938
+ "epoch": 0.83,
1939
+ "learning_rate": 1.832563707884385e-07,
1940
+ "loss": 0.1792,
1941
+ "step": 10304
1942
+ },
1943
+ {
1944
+ "epoch": 0.83,
1945
+ "learning_rate": 1.8051992474773387e-07,
1946
+ "loss": 0.1663,
1947
+ "step": 10336
1948
+ },
1949
+ {
1950
+ "epoch": 0.83,
1951
+ "learning_rate": 1.7778347870702924e-07,
1952
+ "loss": 0.1783,
1953
+ "step": 10368
1954
+ },
1955
+ {
1956
+ "epoch": 0.84,
1957
+ "learning_rate": 1.7504703266632459e-07,
1958
+ "loss": 0.2028,
1959
+ "step": 10400
1960
+ },
1961
+ {
1962
+ "epoch": 0.84,
1963
+ "learning_rate": 1.7231058662561998e-07,
1964
+ "loss": 0.2036,
1965
+ "step": 10432
1966
+ },
1967
+ {
1968
+ "epoch": 0.84,
1969
+ "learning_rate": 1.6957414058491533e-07,
1970
+ "loss": 0.1832,
1971
+ "step": 10464
1972
+ },
1973
+ {
1974
+ "epoch": 0.84,
1975
+ "learning_rate": 1.668376945442107e-07,
1976
+ "loss": 0.1795,
1977
+ "step": 10496
1978
+ },
1979
+ {
1980
+ "epoch": 0.85,
1981
+ "learning_rate": 1.6410124850350608e-07,
1982
+ "loss": 0.1507,
1983
+ "step": 10528
1984
+ },
1985
+ {
1986
+ "epoch": 0.85,
1987
+ "learning_rate": 1.6136480246280142e-07,
1988
+ "loss": 0.2078,
1989
+ "step": 10560
1990
+ },
1991
+ {
1992
+ "epoch": 0.85,
1993
+ "learning_rate": 1.586283564220968e-07,
1994
+ "loss": 0.2228,
1995
+ "step": 10592
1996
+ },
1997
+ {
1998
+ "epoch": 0.85,
1999
+ "learning_rate": 1.5589191038139214e-07,
2000
+ "loss": 0.2,
2001
+ "step": 10624
2002
+ },
2003
+ {
2004
+ "epoch": 0.86,
2005
+ "learning_rate": 1.5315546434068752e-07,
2006
+ "loss": 0.1615,
2007
+ "step": 10656
2008
+ },
2009
+ {
2010
+ "epoch": 0.86,
2011
+ "learning_rate": 1.504190182999829e-07,
2012
+ "loss": 0.1746,
2013
+ "step": 10688
2014
+ },
2015
+ {
2016
+ "epoch": 0.86,
2017
+ "learning_rate": 1.4768257225927824e-07,
2018
+ "loss": 0.1857,
2019
+ "step": 10720
2020
+ },
2021
+ {
2022
+ "epoch": 0.86,
2023
+ "learning_rate": 1.4494612621857364e-07,
2024
+ "loss": 0.2143,
2025
+ "step": 10752
2026
+ },
2027
+ {
2028
+ "epoch": 0.87,
2029
+ "learning_rate": 1.4220968017786898e-07,
2030
+ "loss": 0.1919,
2031
+ "step": 10784
2032
+ },
2033
+ {
2034
+ "epoch": 0.87,
2035
+ "learning_rate": 1.3947323413716436e-07,
2036
+ "loss": 0.1885,
2037
+ "step": 10816
2038
+ },
2039
+ {
2040
+ "epoch": 0.87,
2041
+ "learning_rate": 1.3673678809645973e-07,
2042
+ "loss": 0.1617,
2043
+ "step": 10848
2044
+ },
2045
+ {
2046
+ "epoch": 0.87,
2047
+ "learning_rate": 1.3400034205575508e-07,
2048
+ "loss": 0.1512,
2049
+ "step": 10880
2050
+ },
2051
+ {
2052
+ "epoch": 0.88,
2053
+ "learning_rate": 1.3126389601505045e-07,
2054
+ "loss": 0.1666,
2055
+ "step": 10912
2056
+ },
2057
+ {
2058
+ "epoch": 0.88,
2059
+ "learning_rate": 1.2852744997434582e-07,
2060
+ "loss": 0.1519,
2061
+ "step": 10944
2062
+ },
2063
+ {
2064
+ "epoch": 0.88,
2065
+ "learning_rate": 1.2579100393364117e-07,
2066
+ "loss": 0.1474,
2067
+ "step": 10976
2068
+ },
2069
+ {
2070
+ "epoch": 0.88,
2071
+ "learning_rate": 1.2305455789293654e-07,
2072
+ "loss": 0.3074,
2073
+ "step": 11008
2074
+ },
2075
+ {
2076
+ "epoch": 0.89,
2077
+ "learning_rate": 1.2031811185223191e-07,
2078
+ "loss": 0.1517,
2079
+ "step": 11040
2080
+ },
2081
+ {
2082
+ "epoch": 0.89,
2083
+ "learning_rate": 1.1758166581152727e-07,
2084
+ "loss": 0.3134,
2085
+ "step": 11072
2086
+ },
2087
+ {
2088
+ "epoch": 0.89,
2089
+ "learning_rate": 1.1484521977082263e-07,
2090
+ "loss": 0.1734,
2091
+ "step": 11104
2092
+ },
2093
+ {
2094
+ "epoch": 0.9,
2095
+ "learning_rate": 1.12108773730118e-07,
2096
+ "loss": 0.24,
2097
+ "step": 11136
2098
+ },
2099
+ {
2100
+ "epoch": 0.9,
2101
+ "learning_rate": 1.0937232768941338e-07,
2102
+ "loss": 0.2515,
2103
+ "step": 11168
2104
+ },
2105
+ {
2106
+ "epoch": 0.9,
2107
+ "learning_rate": 1.0663588164870874e-07,
2108
+ "loss": 0.2518,
2109
+ "step": 11200
2110
+ },
2111
+ {
2112
+ "epoch": 0.9,
2113
+ "learning_rate": 1.038994356080041e-07,
2114
+ "loss": 0.2273,
2115
+ "step": 11232
2116
+ },
2117
+ {
2118
+ "epoch": 0.91,
2119
+ "learning_rate": 1.0116298956729946e-07,
2120
+ "loss": 0.2285,
2121
+ "step": 11264
2122
+ },
2123
+ {
2124
+ "epoch": 0.91,
2125
+ "learning_rate": 9.842654352659482e-08,
2126
+ "loss": 0.1872,
2127
+ "step": 11296
2128
+ },
2129
+ {
2130
+ "epoch": 0.91,
2131
+ "learning_rate": 9.56900974858902e-08,
2132
+ "loss": 0.1507,
2133
+ "step": 11328
2134
+ },
2135
+ {
2136
+ "epoch": 0.91,
2137
+ "learning_rate": 9.295365144518557e-08,
2138
+ "loss": 0.2274,
2139
+ "step": 11360
2140
+ },
2141
+ {
2142
+ "epoch": 0.92,
2143
+ "learning_rate": 9.021720540448093e-08,
2144
+ "loss": 0.2448,
2145
+ "step": 11392
2146
+ },
2147
+ {
2148
+ "epoch": 0.92,
2149
+ "learning_rate": 8.748075936377629e-08,
2150
+ "loss": 0.2097,
2151
+ "step": 11424
2152
+ },
2153
+ {
2154
+ "epoch": 0.92,
2155
+ "learning_rate": 8.474431332307165e-08,
2156
+ "loss": 0.1815,
2157
+ "step": 11456
2158
+ },
2159
+ {
2160
+ "epoch": 0.92,
2161
+ "learning_rate": 8.200786728236703e-08,
2162
+ "loss": 0.1796,
2163
+ "step": 11488
2164
+ },
2165
+ {
2166
+ "epoch": 0.93,
2167
+ "learning_rate": 7.927142124166239e-08,
2168
+ "loss": 0.1981,
2169
+ "step": 11520
2170
+ },
2171
+ {
2172
+ "epoch": 0.93,
2173
+ "learning_rate": 7.653497520095775e-08,
2174
+ "loss": 0.168,
2175
+ "step": 11552
2176
+ },
2177
+ {
2178
+ "epoch": 0.93,
2179
+ "learning_rate": 7.379852916025311e-08,
2180
+ "loss": 0.152,
2181
+ "step": 11584
2182
+ },
2183
+ {
2184
+ "epoch": 0.93,
2185
+ "learning_rate": 7.106208311954847e-08,
2186
+ "loss": 0.1627,
2187
+ "step": 11616
2188
+ },
2189
+ {
2190
+ "epoch": 0.94,
2191
+ "learning_rate": 6.832563707884386e-08,
2192
+ "loss": 0.1114,
2193
+ "step": 11648
2194
+ },
2195
+ {
2196
+ "epoch": 0.94,
2197
+ "learning_rate": 6.558919103813922e-08,
2198
+ "loss": 0.2017,
2199
+ "step": 11680
2200
+ },
2201
+ {
2202
+ "epoch": 0.94,
2203
+ "learning_rate": 6.285274499743458e-08,
2204
+ "loss": 0.3228,
2205
+ "step": 11712
2206
+ },
2207
+ {
2208
+ "epoch": 0.94,
2209
+ "learning_rate": 6.011629895672994e-08,
2210
+ "loss": 0.1934,
2211
+ "step": 11744
2212
+ },
2213
+ {
2214
+ "epoch": 0.95,
2215
+ "learning_rate": 5.737985291602531e-08,
2216
+ "loss": 0.2293,
2217
+ "step": 11776
2218
+ },
2219
+ {
2220
+ "epoch": 0.95,
2221
+ "learning_rate": 5.464340687532067e-08,
2222
+ "loss": 0.1689,
2223
+ "step": 11808
2224
+ },
2225
+ {
2226
+ "epoch": 0.95,
2227
+ "learning_rate": 5.190696083461604e-08,
2228
+ "loss": 0.1984,
2229
+ "step": 11840
2230
+ },
2231
+ {
2232
+ "epoch": 0.95,
2233
+ "learning_rate": 4.91705147939114e-08,
2234
+ "loss": 0.2479,
2235
+ "step": 11872
2236
+ },
2237
+ {
2238
+ "epoch": 0.96,
2239
+ "learning_rate": 4.643406875320677e-08,
2240
+ "loss": 0.1582,
2241
+ "step": 11904
2242
+ },
2243
+ {
2244
+ "epoch": 0.96,
2245
+ "learning_rate": 4.3697622712502136e-08,
2246
+ "loss": 0.2271,
2247
+ "step": 11936
2248
+ },
2249
+ {
2250
+ "epoch": 0.96,
2251
+ "learning_rate": 4.09611766717975e-08,
2252
+ "loss": 0.1918,
2253
+ "step": 11968
2254
+ },
2255
+ {
2256
+ "epoch": 0.96,
2257
+ "learning_rate": 3.831024456986488e-08,
2258
+ "loss": 0.1229,
2259
+ "step": 12000
2260
+ },
2261
+ {
2262
+ "epoch": 0.97,
2263
+ "learning_rate": 3.557379852916025e-08,
2264
+ "loss": 0.1692,
2265
+ "step": 12032
2266
+ },
2267
+ {
2268
+ "epoch": 0.97,
2269
+ "learning_rate": 3.2837352488455616e-08,
2270
+ "loss": 0.2048,
2271
+ "step": 12064
2272
+ },
2273
+ {
2274
+ "epoch": 0.97,
2275
+ "learning_rate": 3.010090644775098e-08,
2276
+ "loss": 0.2802,
2277
+ "step": 12096
2278
+ },
2279
+ {
2280
+ "epoch": 0.97,
2281
+ "learning_rate": 2.736446040704635e-08,
2282
+ "loss": 0.1542,
2283
+ "step": 12128
2284
+ },
2285
+ {
2286
+ "epoch": 0.98,
2287
+ "learning_rate": 2.4628014366341712e-08,
2288
+ "loss": 0.1748,
2289
+ "step": 12160
2290
+ },
2291
+ {
2292
+ "epoch": 0.98,
2293
+ "learning_rate": 2.1891568325637078e-08,
2294
+ "loss": 0.1391,
2295
+ "step": 12192
2296
+ },
2297
+ {
2298
+ "epoch": 0.98,
2299
+ "learning_rate": 1.915512228493244e-08,
2300
+ "loss": 0.21,
2301
+ "step": 12224
2302
+ },
2303
+ {
2304
+ "epoch": 0.99,
2305
+ "learning_rate": 1.6418676244227808e-08,
2306
+ "loss": 0.1632,
2307
+ "step": 12256
2308
+ },
2309
+ {
2310
+ "epoch": 0.99,
2311
+ "learning_rate": 1.3682230203523174e-08,
2312
+ "loss": 0.2148,
2313
+ "step": 12288
2314
+ },
2315
+ {
2316
+ "epoch": 0.99,
2317
+ "learning_rate": 1.0945784162818539e-08,
2318
+ "loss": 0.1609,
2319
+ "step": 12320
2320
+ },
2321
+ {
2322
+ "epoch": 0.99,
2323
+ "learning_rate": 8.209338122113904e-09,
2324
+ "loss": 0.2158,
2325
+ "step": 12352
2326
+ },
2327
+ {
2328
+ "epoch": 1.0,
2329
+ "learning_rate": 5.4728920814092695e-09,
2330
+ "loss": 0.2854,
2331
+ "step": 12384
2332
+ },
2333
+ {
2334
+ "epoch": 1.0,
2335
+ "learning_rate": 2.7364460407046348e-09,
2336
+ "loss": 0.2332,
2337
+ "step": 12416
2338
+ }
2339
+ ],
2340
+ "logging_steps": 32,
2341
+ "max_steps": 12440,
2342
+ "num_input_tokens_seen": 0,
2343
+ "num_train_epochs": 1,
2344
+ "save_steps": 1243,
2345
+ "total_flos": 5.2782031577088e+19,
2346
+ "train_batch_size": 2,
2347
+ "trial_name": null,
2348
+ "trial_params": null
2349
+ }
checkpoint-12430/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2343c53f43ecba9eda3fbab0cfda8b6794aaa661c23626f520c968ab283189ab
3
+ size 5048
checkpoint-2486/config.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-large-v2",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "language": "English",
45
+ "mask_feature_length": 10,
46
+ "mask_feature_min_masks": 0,
47
+ "mask_feature_prob": 0.0,
48
+ "mask_time_length": 10,
49
+ "mask_time_min_masks": 2,
50
+ "mask_time_prob": 0.05,
51
+ "max_length": 448,
52
+ "max_source_positions": 1500,
53
+ "max_target_positions": 448,
54
+ "median_filter_width": 7,
55
+ "model_type": "whisper",
56
+ "num_hidden_layers": 32,
57
+ "num_mel_bins": 80,
58
+ "pad_token_id": 50257,
59
+ "scale_embedding": false,
60
+ "suppress_tokens": [],
61
+ "torch_dtype": "float32",
62
+ "transformers_version": "4.37.2",
63
+ "use_cache": true,
64
+ "use_weighted_layer_sum": false,
65
+ "vocab_size": 51865
66
+ }
checkpoint-2486/generation_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "begin_suppress_tokens": [
97
+ 220,
98
+ 50257
99
+ ],
100
+ "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
+ "forced_decoder_ids": [
104
+ [
105
+ 1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
+ ]
112
+ ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
+ "max_initial_timestamp_index": 50,
216
+ "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
+ "return_timestamps": false,
221
+ "suppress_tokens": [
222
+ 1,
223
+ 2,
224
+ 7,
225
+ 8,
226
+ 9,
227
+ 10,
228
+ 14,
229
+ 25,
230
+ 26,
231
+ 27,
232
+ 28,
233
+ 29,
234
+ 31,
235
+ 58,
236
+ 59,
237
+ 60,
238
+ 61,
239
+ 62,
240
+ 63,
241
+ 90,
242
+ 91,
243
+ 92,
244
+ 93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
305
+ 50358,
306
+ 50359,
307
+ 50360,
308
+ 50361,
309
+ 50362
310
+ ],
311
+ "task": "transcribe",
312
+ "task_to_id": {
313
+ "transcribe": 50359,
314
+ "translate": 50358
315
+ },
316
+ "transformers_version": "4.37.2"
317
+ }
checkpoint-2486/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af5b5cd4e92c4ca2fbc9baa1a4945af97031f91130431638ad40a1714ec3bb21
3
+ size 4992706480
checkpoint-2486/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5d16b87cd20c1d44823b7898e9bfdd283fb03c515b83651be38123fc9b80465
3
+ size 1180663192
checkpoint-2486/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2486/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8efbd0884bd23b1cde370826dc6e14f56f54bcabe35739950b6664d802d64c54
3
+ size 3095074288
checkpoint-2486/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-2486/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc80761296b6bb99c8df5ed01b6fd5d4143d86e84f2efeb88107b2eee1981469
3
+ size 14244
checkpoint-2486/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a40a7a944db2c20c81a7aea659ca6dd544ac9a20ffe9620a05ff3249027674fd
3
+ size 1064
checkpoint-2486/trainer_state.json ADDED
@@ -0,0 +1,483 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.19983922829581993,
5
+ "eval_steps": 500,
6
+ "global_step": 2486,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 3.753351206434316e-08,
14
+ "loss": 4.9531,
15
+ "step": 32
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 8.042895442359249e-08,
20
+ "loss": 4.9011,
21
+ "step": 64
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 1.2198391420911528e-07,
26
+ "loss": 4.472,
27
+ "step": 96
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 1.648793565683646e-07,
32
+ "loss": 3.9678,
33
+ "step": 128
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 2.0777479892761392e-07,
38
+ "loss": 3.3412,
39
+ "step": 160
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 2.5067024128686325e-07,
44
+ "loss": 2.5657,
45
+ "step": 192
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 2.9356568364611256e-07,
50
+ "loss": 1.9677,
51
+ "step": 224
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 3.364611260053619e-07,
56
+ "loss": 1.6646,
57
+ "step": 256
58
+ },
59
+ {
60
+ "epoch": 0.02,
61
+ "learning_rate": 3.7935656836461123e-07,
62
+ "loss": 1.5032,
63
+ "step": 288
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 4.222520107238606e-07,
68
+ "loss": 1.3443,
69
+ "step": 320
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 4.651474530831099e-07,
74
+ "loss": 1.2277,
75
+ "step": 352
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 5.080428954423593e-07,
80
+ "loss": 1.0104,
81
+ "step": 384
82
+ },
83
+ {
84
+ "epoch": 0.03,
85
+ "learning_rate": 5.509383378016086e-07,
86
+ "loss": 0.7317,
87
+ "step": 416
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 5.938337801608579e-07,
92
+ "loss": 0.3644,
93
+ "step": 448
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 6.367292225201072e-07,
98
+ "loss": 0.388,
99
+ "step": 480
100
+ },
101
+ {
102
+ "epoch": 0.04,
103
+ "learning_rate": 6.796246648793566e-07,
104
+ "loss": 0.264,
105
+ "step": 512
106
+ },
107
+ {
108
+ "epoch": 0.04,
109
+ "learning_rate": 7.225201072386059e-07,
110
+ "loss": 0.2479,
111
+ "step": 544
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 7.654155495978551e-07,
116
+ "loss": 0.403,
117
+ "step": 576
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 8.083109919571045e-07,
122
+ "loss": 0.3062,
123
+ "step": 608
124
+ },
125
+ {
126
+ "epoch": 0.05,
127
+ "learning_rate": 8.512064343163538e-07,
128
+ "loss": 0.3016,
129
+ "step": 640
130
+ },
131
+ {
132
+ "epoch": 0.05,
133
+ "learning_rate": 8.941018766756032e-07,
134
+ "loss": 0.3445,
135
+ "step": 672
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.369973190348524e-07,
140
+ "loss": 0.2554,
141
+ "step": 704
142
+ },
143
+ {
144
+ "epoch": 0.06,
145
+ "learning_rate": 9.798927613941018e-07,
146
+ "loss": 0.2762,
147
+ "step": 736
148
+ },
149
+ {
150
+ "epoch": 0.06,
151
+ "learning_rate": 9.985462630408755e-07,
152
+ "loss": 0.2265,
153
+ "step": 768
154
+ },
155
+ {
156
+ "epoch": 0.06,
157
+ "learning_rate": 9.958098170001709e-07,
158
+ "loss": 0.2158,
159
+ "step": 800
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.930733709594662e-07,
164
+ "loss": 0.3106,
165
+ "step": 832
166
+ },
167
+ {
168
+ "epoch": 0.07,
169
+ "learning_rate": 9.903369249187618e-07,
170
+ "loss": 0.3085,
171
+ "step": 864
172
+ },
173
+ {
174
+ "epoch": 0.07,
175
+ "learning_rate": 9.876004788780571e-07,
176
+ "loss": 0.2633,
177
+ "step": 896
178
+ },
179
+ {
180
+ "epoch": 0.07,
181
+ "learning_rate": 9.848640328373525e-07,
182
+ "loss": 0.2145,
183
+ "step": 928
184
+ },
185
+ {
186
+ "epoch": 0.08,
187
+ "learning_rate": 9.821275867966478e-07,
188
+ "loss": 0.2594,
189
+ "step": 960
190
+ },
191
+ {
192
+ "epoch": 0.08,
193
+ "learning_rate": 9.793911407559432e-07,
194
+ "loss": 0.2264,
195
+ "step": 992
196
+ },
197
+ {
198
+ "epoch": 0.08,
199
+ "learning_rate": 9.766546947152385e-07,
200
+ "loss": 0.2512,
201
+ "step": 1024
202
+ },
203
+ {
204
+ "epoch": 0.08,
205
+ "learning_rate": 9.739182486745339e-07,
206
+ "loss": 0.3154,
207
+ "step": 1056
208
+ },
209
+ {
210
+ "epoch": 0.09,
211
+ "learning_rate": 9.711818026338292e-07,
212
+ "loss": 0.2672,
213
+ "step": 1088
214
+ },
215
+ {
216
+ "epoch": 0.09,
217
+ "learning_rate": 9.684453565931246e-07,
218
+ "loss": 0.2502,
219
+ "step": 1120
220
+ },
221
+ {
222
+ "epoch": 0.09,
223
+ "learning_rate": 9.6570891055242e-07,
224
+ "loss": 0.2702,
225
+ "step": 1152
226
+ },
227
+ {
228
+ "epoch": 0.1,
229
+ "learning_rate": 9.629724645117153e-07,
230
+ "loss": 0.2919,
231
+ "step": 1184
232
+ },
233
+ {
234
+ "epoch": 0.1,
235
+ "learning_rate": 9.602360184710108e-07,
236
+ "loss": 0.3925,
237
+ "step": 1216
238
+ },
239
+ {
240
+ "epoch": 0.1,
241
+ "learning_rate": 9.574995724303062e-07,
242
+ "loss": 0.285,
243
+ "step": 1248
244
+ },
245
+ {
246
+ "epoch": 0.1,
247
+ "learning_rate": 9.547631263896015e-07,
248
+ "loss": 0.3084,
249
+ "step": 1280
250
+ },
251
+ {
252
+ "epoch": 0.11,
253
+ "learning_rate": 9.520266803488969e-07,
254
+ "loss": 0.2275,
255
+ "step": 1312
256
+ },
257
+ {
258
+ "epoch": 0.11,
259
+ "learning_rate": 9.492902343081922e-07,
260
+ "loss": 0.245,
261
+ "step": 1344
262
+ },
263
+ {
264
+ "epoch": 0.11,
265
+ "learning_rate": 9.465537882674875e-07,
266
+ "loss": 0.233,
267
+ "step": 1376
268
+ },
269
+ {
270
+ "epoch": 0.11,
271
+ "learning_rate": 9.438173422267829e-07,
272
+ "loss": 0.2825,
273
+ "step": 1408
274
+ },
275
+ {
276
+ "epoch": 0.12,
277
+ "learning_rate": 9.410808961860783e-07,
278
+ "loss": 0.231,
279
+ "step": 1440
280
+ },
281
+ {
282
+ "epoch": 0.12,
283
+ "learning_rate": 9.383444501453737e-07,
284
+ "loss": 0.2449,
285
+ "step": 1472
286
+ },
287
+ {
288
+ "epoch": 0.12,
289
+ "learning_rate": 9.35608004104669e-07,
290
+ "loss": 0.2732,
291
+ "step": 1504
292
+ },
293
+ {
294
+ "epoch": 0.12,
295
+ "learning_rate": 9.328715580639644e-07,
296
+ "loss": 0.2031,
297
+ "step": 1536
298
+ },
299
+ {
300
+ "epoch": 0.13,
301
+ "learning_rate": 9.301351120232597e-07,
302
+ "loss": 0.1749,
303
+ "step": 1568
304
+ },
305
+ {
306
+ "epoch": 0.13,
307
+ "learning_rate": 9.273986659825551e-07,
308
+ "loss": 0.1722,
309
+ "step": 1600
310
+ },
311
+ {
312
+ "epoch": 0.13,
313
+ "learning_rate": 9.246622199418504e-07,
314
+ "loss": 0.2743,
315
+ "step": 1632
316
+ },
317
+ {
318
+ "epoch": 0.13,
319
+ "learning_rate": 9.219257739011459e-07,
320
+ "loss": 0.2907,
321
+ "step": 1664
322
+ },
323
+ {
324
+ "epoch": 0.14,
325
+ "learning_rate": 9.192748417992132e-07,
326
+ "loss": 0.2664,
327
+ "step": 1696
328
+ },
329
+ {
330
+ "epoch": 0.14,
331
+ "learning_rate": 9.165383957585086e-07,
332
+ "loss": 0.2085,
333
+ "step": 1728
334
+ },
335
+ {
336
+ "epoch": 0.14,
337
+ "learning_rate": 9.13801949717804e-07,
338
+ "loss": 0.1839,
339
+ "step": 1760
340
+ },
341
+ {
342
+ "epoch": 0.14,
343
+ "learning_rate": 9.110655036770994e-07,
344
+ "loss": 0.2667,
345
+ "step": 1792
346
+ },
347
+ {
348
+ "epoch": 0.15,
349
+ "learning_rate": 9.083290576363947e-07,
350
+ "loss": 0.1994,
351
+ "step": 1824
352
+ },
353
+ {
354
+ "epoch": 0.15,
355
+ "learning_rate": 9.0559261159569e-07,
356
+ "loss": 0.2568,
357
+ "step": 1856
358
+ },
359
+ {
360
+ "epoch": 0.15,
361
+ "learning_rate": 9.028561655549855e-07,
362
+ "loss": 0.2909,
363
+ "step": 1888
364
+ },
365
+ {
366
+ "epoch": 0.15,
367
+ "learning_rate": 9.001197195142807e-07,
368
+ "loss": 0.2697,
369
+ "step": 1920
370
+ },
371
+ {
372
+ "epoch": 0.16,
373
+ "learning_rate": 8.973832734735761e-07,
374
+ "loss": 0.3379,
375
+ "step": 1952
376
+ },
377
+ {
378
+ "epoch": 0.16,
379
+ "learning_rate": 8.946468274328715e-07,
380
+ "loss": 0.2866,
381
+ "step": 1984
382
+ },
383
+ {
384
+ "epoch": 0.16,
385
+ "learning_rate": 8.919103813921669e-07,
386
+ "loss": 0.2634,
387
+ "step": 2016
388
+ },
389
+ {
390
+ "epoch": 0.16,
391
+ "learning_rate": 8.891739353514622e-07,
392
+ "loss": 0.2234,
393
+ "step": 2048
394
+ },
395
+ {
396
+ "epoch": 0.17,
397
+ "learning_rate": 8.864374893107576e-07,
398
+ "loss": 0.2541,
399
+ "step": 2080
400
+ },
401
+ {
402
+ "epoch": 0.17,
403
+ "learning_rate": 8.83701043270053e-07,
404
+ "loss": 0.2341,
405
+ "step": 2112
406
+ },
407
+ {
408
+ "epoch": 0.17,
409
+ "learning_rate": 8.809645972293484e-07,
410
+ "loss": 0.2602,
411
+ "step": 2144
412
+ },
413
+ {
414
+ "epoch": 0.17,
415
+ "learning_rate": 8.782281511886437e-07,
416
+ "loss": 0.2602,
417
+ "step": 2176
418
+ },
419
+ {
420
+ "epoch": 0.18,
421
+ "learning_rate": 8.75491705147939e-07,
422
+ "loss": 0.2036,
423
+ "step": 2208
424
+ },
425
+ {
426
+ "epoch": 0.18,
427
+ "learning_rate": 8.727552591072344e-07,
428
+ "loss": 0.2342,
429
+ "step": 2240
430
+ },
431
+ {
432
+ "epoch": 0.18,
433
+ "learning_rate": 8.700188130665298e-07,
434
+ "loss": 0.2361,
435
+ "step": 2272
436
+ },
437
+ {
438
+ "epoch": 0.19,
439
+ "learning_rate": 8.672823670258251e-07,
440
+ "loss": 0.3299,
441
+ "step": 2304
442
+ },
443
+ {
444
+ "epoch": 0.19,
445
+ "learning_rate": 8.645459209851206e-07,
446
+ "loss": 0.3221,
447
+ "step": 2336
448
+ },
449
+ {
450
+ "epoch": 0.19,
451
+ "learning_rate": 8.618094749444159e-07,
452
+ "loss": 0.2119,
453
+ "step": 2368
454
+ },
455
+ {
456
+ "epoch": 0.19,
457
+ "learning_rate": 8.590730289037113e-07,
458
+ "loss": 0.1908,
459
+ "step": 2400
460
+ },
461
+ {
462
+ "epoch": 0.2,
463
+ "learning_rate": 8.563365828630066e-07,
464
+ "loss": 0.2736,
465
+ "step": 2432
466
+ },
467
+ {
468
+ "epoch": 0.2,
469
+ "learning_rate": 8.536001368223021e-07,
470
+ "loss": 0.1713,
471
+ "step": 2464
472
+ }
473
+ ],
474
+ "logging_steps": 32,
475
+ "max_steps": 12440,
476
+ "num_input_tokens_seen": 0,
477
+ "num_train_epochs": 1,
478
+ "save_steps": 1243,
479
+ "total_flos": 1.05564063154176e+19,
480
+ "train_batch_size": 2,
481
+ "trial_name": null,
482
+ "trial_params": null
483
+ }
checkpoint-2486/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2343c53f43ecba9eda3fbab0cfda8b6794aaa661c23626f520c968ab283189ab
3
+ size 5048
checkpoint-3729/config.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-large-v2",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "language": "English",
45
+ "mask_feature_length": 10,
46
+ "mask_feature_min_masks": 0,
47
+ "mask_feature_prob": 0.0,
48
+ "mask_time_length": 10,
49
+ "mask_time_min_masks": 2,
50
+ "mask_time_prob": 0.05,
51
+ "max_length": 448,
52
+ "max_source_positions": 1500,
53
+ "max_target_positions": 448,
54
+ "median_filter_width": 7,
55
+ "model_type": "whisper",
56
+ "num_hidden_layers": 32,
57
+ "num_mel_bins": 80,
58
+ "pad_token_id": 50257,
59
+ "scale_embedding": false,
60
+ "suppress_tokens": [],
61
+ "torch_dtype": "float32",
62
+ "transformers_version": "4.37.2",
63
+ "use_cache": true,
64
+ "use_weighted_layer_sum": false,
65
+ "vocab_size": 51865
66
+ }
checkpoint-3729/generation_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "begin_suppress_tokens": [
97
+ 220,
98
+ 50257
99
+ ],
100
+ "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
+ "forced_decoder_ids": [
104
+ [
105
+ 1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
+ ]
112
+ ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
+ "max_initial_timestamp_index": 50,
216
+ "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
+ "return_timestamps": false,
221
+ "suppress_tokens": [
222
+ 1,
223
+ 2,
224
+ 7,
225
+ 8,
226
+ 9,
227
+ 10,
228
+ 14,
229
+ 25,
230
+ 26,
231
+ 27,
232
+ 28,
233
+ 29,
234
+ 31,
235
+ 58,
236
+ 59,
237
+ 60,
238
+ 61,
239
+ 62,
240
+ 63,
241
+ 90,
242
+ 91,
243
+ 92,
244
+ 93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
305
+ 50358,
306
+ 50359,
307
+ 50360,
308
+ 50361,
309
+ 50362
310
+ ],
311
+ "task": "transcribe",
312
+ "task_to_id": {
313
+ "transcribe": 50359,
314
+ "translate": 50358
315
+ },
316
+ "transformers_version": "4.37.2"
317
+ }
checkpoint-3729/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e46ed3553eacb1b9780a5a6af5d2ea15d38821b724163f1df8d7116dd9ce1c6c
3
+ size 4992706480
checkpoint-3729/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0adf238584a0a8b67790f24f04dcbb2861816a19bb410166458d2a3fbbcccf65
3
+ size 1180663192
checkpoint-3729/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3729/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68d9d2f6b00536d584bc6d9a89541bedb9489fa757129122b98d9935478d40fa
3
+ size 3095074288