hsethu commited on
Commit
4dc4e16
·
verified ·
1 Parent(s): 28a84a0

Upload folder using huggingface_hub

Browse files
Files changed (40) hide show
  1. checkpoint-14990/config.json +66 -0
  2. checkpoint-14990/generation_config.json +317 -0
  3. checkpoint-14990/model-00001-of-00002.safetensors +3 -0
  4. checkpoint-14990/model-00002-of-00002.safetensors +3 -0
  5. checkpoint-14990/model.safetensors.index.json +0 -0
  6. checkpoint-14990/optimizer.pt +3 -0
  7. checkpoint-14990/preprocessor_config.json +14 -0
  8. checkpoint-14990/rng_state.pth +3 -0
  9. checkpoint-14990/scheduler.pt +3 -0
  10. checkpoint-14990/trainer_state.json +1425 -0
  11. checkpoint-14990/training_args.bin +3 -0
  12. checkpoint-22485/config.json +66 -0
  13. checkpoint-22485/generation_config.json +317 -0
  14. checkpoint-22485/model-00001-of-00002.safetensors +3 -0
  15. checkpoint-22485/model-00002-of-00002.safetensors +3 -0
  16. checkpoint-22485/model.safetensors.index.json +0 -0
  17. checkpoint-22485/optimizer.pt +3 -0
  18. checkpoint-22485/preprocessor_config.json +14 -0
  19. checkpoint-22485/rng_state.pth +3 -0
  20. checkpoint-22485/scheduler.pt +3 -0
  21. checkpoint-22485/trainer_state.json +2127 -0
  22. checkpoint-22485/training_args.bin +3 -0
  23. checkpoint-7495/config.json +66 -0
  24. checkpoint-7495/generation_config.json +317 -0
  25. checkpoint-7495/model-00001-of-00002.safetensors +3 -0
  26. checkpoint-7495/model-00002-of-00002.safetensors +3 -0
  27. checkpoint-7495/model.safetensors.index.json +0 -0
  28. checkpoint-7495/optimizer.pt +3 -0
  29. checkpoint-7495/preprocessor_config.json +14 -0
  30. checkpoint-7495/rng_state.pth +3 -0
  31. checkpoint-7495/scheduler.pt +3 -0
  32. checkpoint-7495/trainer_state.json +723 -0
  33. checkpoint-7495/training_args.bin +3 -0
  34. config.json +66 -0
  35. generation_config.json +317 -0
  36. model-00001-of-00002.safetensors +3 -0
  37. model-00002-of-00002.safetensors +3 -0
  38. model.safetensors.index.json +0 -0
  39. preprocessor_config.json +14 -0
  40. training_args.bin +3 -0
checkpoint-14990/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-14990/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-14990/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbc623fadff2cf53e5d8cf0c0da3f584676ddac82933872aba781061ccd890be
3
+ size 4992706480
checkpoint-14990/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:277e5a04728c11e47864dff711b92ad46c7c7aee9caffd3185f49159861be475
3
+ size 1180663192
checkpoint-14990/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-14990/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:961339fd8a3479d8763d426936c91313563cf976d5b13bd039ba303c9ed2f513
3
+ size 3095074288
checkpoint-14990/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-14990/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19c443289acfa60805307d69881b0ec1c387b572434a6fa73e290bf4e81910fa
3
+ size 14244
checkpoint-14990/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14cc4f9e7c1f179fe34a511480a210c48f33768f8bb7769891b393550f91de02
3
+ size 1064
checkpoint-14990/trainer_state.json ADDED
@@ -0,0 +1,1425 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.6666370185893444,
5
+ "eval_steps": 500,
6
+ "global_step": 14990,
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": 9.973761451569865e-07,
14
+ "loss": 3.933,
15
+ "step": 64
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 9.945299297340568e-07,
20
+ "loss": 1.6429,
21
+ "step": 128
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 9.917726585430935e-07,
26
+ "loss": 1.1012,
27
+ "step": 192
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 9.889264431201636e-07,
32
+ "loss": 0.3376,
33
+ "step": 256
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 9.860802276972338e-07,
38
+ "loss": 0.136,
39
+ "step": 320
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 9.832340122743039e-07,
44
+ "loss": 0.2753,
45
+ "step": 384
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 9.803877968513742e-07,
50
+ "loss": 0.25,
51
+ "step": 448
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 9.775415814284444e-07,
56
+ "loss": 0.3117,
57
+ "step": 512
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 9.746953660055145e-07,
62
+ "loss": 0.2853,
63
+ "step": 576
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 9.718491505825848e-07,
68
+ "loss": 0.2489,
69
+ "step": 640
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 9.690029351596548e-07,
74
+ "loss": 0.217,
75
+ "step": 704
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 9.66156719736725e-07,
80
+ "loss": 0.1804,
81
+ "step": 768
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 9.633105043137952e-07,
86
+ "loss": 0.2842,
87
+ "step": 832
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 9.604642888908654e-07,
92
+ "loss": 0.2621,
93
+ "step": 896
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 9.576180734679355e-07,
98
+ "loss": 0.3316,
99
+ "step": 960
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 9.547718580450058e-07,
104
+ "loss": 0.1856,
105
+ "step": 1024
106
+ },
107
+ {
108
+ "epoch": 0.05,
109
+ "learning_rate": 9.519256426220758e-07,
110
+ "loss": 0.2112,
111
+ "step": 1088
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 9.490794271991461e-07,
116
+ "loss": 0.1778,
117
+ "step": 1152
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 9.462332117762163e-07,
122
+ "loss": 0.319,
123
+ "step": 1216
124
+ },
125
+ {
126
+ "epoch": 0.06,
127
+ "learning_rate": 9.433869963532864e-07,
128
+ "loss": 0.1678,
129
+ "step": 1280
130
+ },
131
+ {
132
+ "epoch": 0.06,
133
+ "learning_rate": 9.405407809303566e-07,
134
+ "loss": 0.2549,
135
+ "step": 1344
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.376945655074268e-07,
140
+ "loss": 0.1723,
141
+ "step": 1408
142
+ },
143
+ {
144
+ "epoch": 0.07,
145
+ "learning_rate": 9.34848350084497e-07,
146
+ "loss": 0.149,
147
+ "step": 1472
148
+ },
149
+ {
150
+ "epoch": 0.07,
151
+ "learning_rate": 9.320021346615672e-07,
152
+ "loss": 0.2097,
153
+ "step": 1536
154
+ },
155
+ {
156
+ "epoch": 0.07,
157
+ "learning_rate": 9.291559192386374e-07,
158
+ "loss": 0.3716,
159
+ "step": 1600
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.263097038157075e-07,
164
+ "loss": 0.1303,
165
+ "step": 1664
166
+ },
167
+ {
168
+ "epoch": 0.08,
169
+ "learning_rate": 9.234634883927777e-07,
170
+ "loss": 0.2041,
171
+ "step": 1728
172
+ },
173
+ {
174
+ "epoch": 0.08,
175
+ "learning_rate": 9.206172729698479e-07,
176
+ "loss": 0.2529,
177
+ "step": 1792
178
+ },
179
+ {
180
+ "epoch": 0.08,
181
+ "learning_rate": 9.177710575469181e-07,
182
+ "loss": 0.2241,
183
+ "step": 1856
184
+ },
185
+ {
186
+ "epoch": 0.09,
187
+ "learning_rate": 9.149248421239883e-07,
188
+ "loss": 0.2056,
189
+ "step": 1920
190
+ },
191
+ {
192
+ "epoch": 0.09,
193
+ "learning_rate": 9.120786267010584e-07,
194
+ "loss": 0.2956,
195
+ "step": 1984
196
+ },
197
+ {
198
+ "epoch": 0.09,
199
+ "learning_rate": 9.092324112781285e-07,
200
+ "loss": 0.2177,
201
+ "step": 2048
202
+ },
203
+ {
204
+ "epoch": 0.09,
205
+ "learning_rate": 9.063861958551987e-07,
206
+ "loss": 0.2379,
207
+ "step": 2112
208
+ },
209
+ {
210
+ "epoch": 0.1,
211
+ "learning_rate": 9.03539980432269e-07,
212
+ "loss": 0.3499,
213
+ "step": 2176
214
+ },
215
+ {
216
+ "epoch": 0.1,
217
+ "learning_rate": 9.006937650093391e-07,
218
+ "loss": 0.2133,
219
+ "step": 2240
220
+ },
221
+ {
222
+ "epoch": 0.1,
223
+ "learning_rate": 8.978475495864093e-07,
224
+ "loss": 0.283,
225
+ "step": 2304
226
+ },
227
+ {
228
+ "epoch": 0.11,
229
+ "learning_rate": 8.950013341634795e-07,
230
+ "loss": 0.2337,
231
+ "step": 2368
232
+ },
233
+ {
234
+ "epoch": 0.11,
235
+ "learning_rate": 8.921551187405496e-07,
236
+ "loss": 0.1282,
237
+ "step": 2432
238
+ },
239
+ {
240
+ "epoch": 0.11,
241
+ "learning_rate": 8.893089033176198e-07,
242
+ "loss": 0.1615,
243
+ "step": 2496
244
+ },
245
+ {
246
+ "epoch": 0.11,
247
+ "learning_rate": 8.864626878946901e-07,
248
+ "loss": 0.1537,
249
+ "step": 2560
250
+ },
251
+ {
252
+ "epoch": 0.12,
253
+ "learning_rate": 8.836164724717602e-07,
254
+ "loss": 0.2078,
255
+ "step": 2624
256
+ },
257
+ {
258
+ "epoch": 0.12,
259
+ "learning_rate": 8.807702570488304e-07,
260
+ "loss": 0.2886,
261
+ "step": 2688
262
+ },
263
+ {
264
+ "epoch": 0.12,
265
+ "learning_rate": 8.779240416259006e-07,
266
+ "loss": 0.2204,
267
+ "step": 2752
268
+ },
269
+ {
270
+ "epoch": 0.13,
271
+ "learning_rate": 8.750778262029706e-07,
272
+ "loss": 0.2415,
273
+ "step": 2816
274
+ },
275
+ {
276
+ "epoch": 0.13,
277
+ "learning_rate": 8.722316107800409e-07,
278
+ "loss": 0.1124,
279
+ "step": 2880
280
+ },
281
+ {
282
+ "epoch": 0.13,
283
+ "learning_rate": 8.69385395357111e-07,
284
+ "loss": 0.1638,
285
+ "step": 2944
286
+ },
287
+ {
288
+ "epoch": 0.13,
289
+ "learning_rate": 8.665391799341812e-07,
290
+ "loss": 0.2198,
291
+ "step": 3008
292
+ },
293
+ {
294
+ "epoch": 0.14,
295
+ "learning_rate": 8.636929645112514e-07,
296
+ "loss": 0.1793,
297
+ "step": 3072
298
+ },
299
+ {
300
+ "epoch": 0.14,
301
+ "learning_rate": 8.608467490883215e-07,
302
+ "loss": 0.1939,
303
+ "step": 3136
304
+ },
305
+ {
306
+ "epoch": 0.14,
307
+ "learning_rate": 8.580005336653917e-07,
308
+ "loss": 0.2771,
309
+ "step": 3200
310
+ },
311
+ {
312
+ "epoch": 0.15,
313
+ "learning_rate": 8.55154318242462e-07,
314
+ "loss": 0.1291,
315
+ "step": 3264
316
+ },
317
+ {
318
+ "epoch": 0.15,
319
+ "learning_rate": 8.523081028195322e-07,
320
+ "loss": 0.216,
321
+ "step": 3328
322
+ },
323
+ {
324
+ "epoch": 0.15,
325
+ "learning_rate": 8.494618873966023e-07,
326
+ "loss": 0.1985,
327
+ "step": 3392
328
+ },
329
+ {
330
+ "epoch": 0.15,
331
+ "learning_rate": 8.466156719736725e-07,
332
+ "loss": 0.2458,
333
+ "step": 3456
334
+ },
335
+ {
336
+ "epoch": 0.16,
337
+ "learning_rate": 8.437694565507427e-07,
338
+ "loss": 0.1253,
339
+ "step": 3520
340
+ },
341
+ {
342
+ "epoch": 0.16,
343
+ "learning_rate": 8.409232411278129e-07,
344
+ "loss": 0.2109,
345
+ "step": 3584
346
+ },
347
+ {
348
+ "epoch": 0.16,
349
+ "learning_rate": 8.38077025704883e-07,
350
+ "loss": 0.2105,
351
+ "step": 3648
352
+ },
353
+ {
354
+ "epoch": 0.17,
355
+ "learning_rate": 8.352308102819531e-07,
356
+ "loss": 0.2013,
357
+ "step": 3712
358
+ },
359
+ {
360
+ "epoch": 0.17,
361
+ "learning_rate": 8.323845948590233e-07,
362
+ "loss": 0.203,
363
+ "step": 3776
364
+ },
365
+ {
366
+ "epoch": 0.17,
367
+ "learning_rate": 8.295383794360935e-07,
368
+ "loss": 0.2366,
369
+ "step": 3840
370
+ },
371
+ {
372
+ "epoch": 0.17,
373
+ "learning_rate": 8.266921640131636e-07,
374
+ "loss": 0.2181,
375
+ "step": 3904
376
+ },
377
+ {
378
+ "epoch": 0.18,
379
+ "learning_rate": 8.238459485902339e-07,
380
+ "loss": 0.2611,
381
+ "step": 3968
382
+ },
383
+ {
384
+ "epoch": 0.18,
385
+ "learning_rate": 8.209997331673041e-07,
386
+ "loss": 0.1412,
387
+ "step": 4032
388
+ },
389
+ {
390
+ "epoch": 0.18,
391
+ "learning_rate": 8.181535177443743e-07,
392
+ "loss": 0.2678,
393
+ "step": 4096
394
+ },
395
+ {
396
+ "epoch": 0.19,
397
+ "learning_rate": 8.153073023214444e-07,
398
+ "loss": 0.1847,
399
+ "step": 4160
400
+ },
401
+ {
402
+ "epoch": 0.19,
403
+ "learning_rate": 8.124610868985146e-07,
404
+ "loss": 0.2395,
405
+ "step": 4224
406
+ },
407
+ {
408
+ "epoch": 0.19,
409
+ "learning_rate": 8.096148714755849e-07,
410
+ "loss": 0.1473,
411
+ "step": 4288
412
+ },
413
+ {
414
+ "epoch": 0.19,
415
+ "learning_rate": 8.06768656052655e-07,
416
+ "loss": 0.2675,
417
+ "step": 4352
418
+ },
419
+ {
420
+ "epoch": 0.2,
421
+ "learning_rate": 8.039224406297252e-07,
422
+ "loss": 0.2179,
423
+ "step": 4416
424
+ },
425
+ {
426
+ "epoch": 0.2,
427
+ "learning_rate": 8.010762252067952e-07,
428
+ "loss": 0.1667,
429
+ "step": 4480
430
+ },
431
+ {
432
+ "epoch": 0.2,
433
+ "learning_rate": 7.982300097838654e-07,
434
+ "loss": 0.2114,
435
+ "step": 4544
436
+ },
437
+ {
438
+ "epoch": 0.2,
439
+ "learning_rate": 7.953837943609356e-07,
440
+ "loss": 0.1929,
441
+ "step": 4608
442
+ },
443
+ {
444
+ "epoch": 0.21,
445
+ "learning_rate": 7.925375789380059e-07,
446
+ "loss": 0.2152,
447
+ "step": 4672
448
+ },
449
+ {
450
+ "epoch": 0.21,
451
+ "learning_rate": 7.89691363515076e-07,
452
+ "loss": 0.2065,
453
+ "step": 4736
454
+ },
455
+ {
456
+ "epoch": 0.21,
457
+ "learning_rate": 7.868451480921462e-07,
458
+ "loss": 0.1607,
459
+ "step": 4800
460
+ },
461
+ {
462
+ "epoch": 0.22,
463
+ "learning_rate": 7.839989326692163e-07,
464
+ "loss": 0.1592,
465
+ "step": 4864
466
+ },
467
+ {
468
+ "epoch": 0.22,
469
+ "learning_rate": 7.811527172462865e-07,
470
+ "loss": 0.2558,
471
+ "step": 4928
472
+ },
473
+ {
474
+ "epoch": 0.22,
475
+ "learning_rate": 7.783065018233567e-07,
476
+ "loss": 0.1564,
477
+ "step": 4992
478
+ },
479
+ {
480
+ "epoch": 0.22,
481
+ "learning_rate": 7.75460286400427e-07,
482
+ "loss": 0.2171,
483
+ "step": 5056
484
+ },
485
+ {
486
+ "epoch": 0.23,
487
+ "learning_rate": 7.726140709774971e-07,
488
+ "loss": 0.1973,
489
+ "step": 5120
490
+ },
491
+ {
492
+ "epoch": 0.23,
493
+ "learning_rate": 7.697678555545673e-07,
494
+ "loss": 0.1273,
495
+ "step": 5184
496
+ },
497
+ {
498
+ "epoch": 0.23,
499
+ "learning_rate": 7.669216401316375e-07,
500
+ "loss": 0.1158,
501
+ "step": 5248
502
+ },
503
+ {
504
+ "epoch": 0.24,
505
+ "learning_rate": 7.640754247087075e-07,
506
+ "loss": 0.151,
507
+ "step": 5312
508
+ },
509
+ {
510
+ "epoch": 0.24,
511
+ "learning_rate": 7.612292092857778e-07,
512
+ "loss": 0.2727,
513
+ "step": 5376
514
+ },
515
+ {
516
+ "epoch": 0.24,
517
+ "learning_rate": 7.58382993862848e-07,
518
+ "loss": 0.1271,
519
+ "step": 5440
520
+ },
521
+ {
522
+ "epoch": 0.24,
523
+ "learning_rate": 7.555367784399181e-07,
524
+ "loss": 0.2409,
525
+ "step": 5504
526
+ },
527
+ {
528
+ "epoch": 0.25,
529
+ "learning_rate": 7.526905630169883e-07,
530
+ "loss": 0.1456,
531
+ "step": 5568
532
+ },
533
+ {
534
+ "epoch": 0.25,
535
+ "learning_rate": 7.498443475940584e-07,
536
+ "loss": 0.249,
537
+ "step": 5632
538
+ },
539
+ {
540
+ "epoch": 0.25,
541
+ "learning_rate": 7.469981321711286e-07,
542
+ "loss": 0.2348,
543
+ "step": 5696
544
+ },
545
+ {
546
+ "epoch": 0.26,
547
+ "learning_rate": 7.441519167481989e-07,
548
+ "loss": 0.2074,
549
+ "step": 5760
550
+ },
551
+ {
552
+ "epoch": 0.26,
553
+ "learning_rate": 7.41305701325269e-07,
554
+ "loss": 0.1483,
555
+ "step": 5824
556
+ },
557
+ {
558
+ "epoch": 0.26,
559
+ "learning_rate": 7.384594859023392e-07,
560
+ "loss": 0.23,
561
+ "step": 5888
562
+ },
563
+ {
564
+ "epoch": 0.26,
565
+ "learning_rate": 7.356132704794094e-07,
566
+ "loss": 0.1294,
567
+ "step": 5952
568
+ },
569
+ {
570
+ "epoch": 0.27,
571
+ "learning_rate": 7.327670550564795e-07,
572
+ "loss": 0.2088,
573
+ "step": 6016
574
+ },
575
+ {
576
+ "epoch": 0.27,
577
+ "learning_rate": 7.299208396335498e-07,
578
+ "loss": 0.2233,
579
+ "step": 6080
580
+ },
581
+ {
582
+ "epoch": 0.27,
583
+ "learning_rate": 7.2707462421062e-07,
584
+ "loss": 0.1361,
585
+ "step": 6144
586
+ },
587
+ {
588
+ "epoch": 0.28,
589
+ "learning_rate": 7.2422840878769e-07,
590
+ "loss": 0.1664,
591
+ "step": 6208
592
+ },
593
+ {
594
+ "epoch": 0.28,
595
+ "learning_rate": 7.213821933647602e-07,
596
+ "loss": 0.1767,
597
+ "step": 6272
598
+ },
599
+ {
600
+ "epoch": 0.28,
601
+ "learning_rate": 7.185359779418304e-07,
602
+ "loss": 0.1613,
603
+ "step": 6336
604
+ },
605
+ {
606
+ "epoch": 0.28,
607
+ "learning_rate": 7.156897625189005e-07,
608
+ "loss": 0.1776,
609
+ "step": 6400
610
+ },
611
+ {
612
+ "epoch": 0.29,
613
+ "learning_rate": 7.128435470959708e-07,
614
+ "loss": 0.2314,
615
+ "step": 6464
616
+ },
617
+ {
618
+ "epoch": 0.29,
619
+ "learning_rate": 7.09997331673041e-07,
620
+ "loss": 0.1439,
621
+ "step": 6528
622
+ },
623
+ {
624
+ "epoch": 0.29,
625
+ "learning_rate": 7.071511162501111e-07,
626
+ "loss": 0.2195,
627
+ "step": 6592
628
+ },
629
+ {
630
+ "epoch": 0.3,
631
+ "learning_rate": 7.043049008271813e-07,
632
+ "loss": 0.248,
633
+ "step": 6656
634
+ },
635
+ {
636
+ "epoch": 0.3,
637
+ "learning_rate": 7.014586854042515e-07,
638
+ "loss": 0.2165,
639
+ "step": 6720
640
+ },
641
+ {
642
+ "epoch": 0.3,
643
+ "learning_rate": 6.986124699813218e-07,
644
+ "loss": 0.2642,
645
+ "step": 6784
646
+ },
647
+ {
648
+ "epoch": 0.3,
649
+ "learning_rate": 6.957662545583919e-07,
650
+ "loss": 0.236,
651
+ "step": 6848
652
+ },
653
+ {
654
+ "epoch": 0.31,
655
+ "learning_rate": 6.929200391354621e-07,
656
+ "loss": 0.2041,
657
+ "step": 6912
658
+ },
659
+ {
660
+ "epoch": 0.31,
661
+ "learning_rate": 6.900738237125323e-07,
662
+ "loss": 0.1925,
663
+ "step": 6976
664
+ },
665
+ {
666
+ "epoch": 0.31,
667
+ "learning_rate": 6.872276082896023e-07,
668
+ "loss": 0.2164,
669
+ "step": 7040
670
+ },
671
+ {
672
+ "epoch": 0.32,
673
+ "learning_rate": 6.843813928666725e-07,
674
+ "loss": 0.1399,
675
+ "step": 7104
676
+ },
677
+ {
678
+ "epoch": 0.32,
679
+ "learning_rate": 6.815351774437427e-07,
680
+ "loss": 0.1875,
681
+ "step": 7168
682
+ },
683
+ {
684
+ "epoch": 0.32,
685
+ "learning_rate": 6.786889620208129e-07,
686
+ "loss": 0.176,
687
+ "step": 7232
688
+ },
689
+ {
690
+ "epoch": 0.32,
691
+ "learning_rate": 6.758427465978831e-07,
692
+ "loss": 0.2326,
693
+ "step": 7296
694
+ },
695
+ {
696
+ "epoch": 0.33,
697
+ "learning_rate": 6.729965311749532e-07,
698
+ "loss": 0.1681,
699
+ "step": 7360
700
+ },
701
+ {
702
+ "epoch": 0.33,
703
+ "learning_rate": 6.701503157520234e-07,
704
+ "loss": 0.2422,
705
+ "step": 7424
706
+ },
707
+ {
708
+ "epoch": 0.33,
709
+ "learning_rate": 6.673041003290937e-07,
710
+ "loss": 0.1881,
711
+ "step": 7488
712
+ },
713
+ {
714
+ "epoch": 0.34,
715
+ "learning_rate": 6.644578849061639e-07,
716
+ "loss": 0.1716,
717
+ "step": 7552
718
+ },
719
+ {
720
+ "epoch": 0.34,
721
+ "learning_rate": 6.61611669483234e-07,
722
+ "loss": 0.1565,
723
+ "step": 7616
724
+ },
725
+ {
726
+ "epoch": 0.34,
727
+ "learning_rate": 6.587654540603042e-07,
728
+ "loss": 0.223,
729
+ "step": 7680
730
+ },
731
+ {
732
+ "epoch": 0.34,
733
+ "learning_rate": 6.559192386373743e-07,
734
+ "loss": 0.1473,
735
+ "step": 7744
736
+ },
737
+ {
738
+ "epoch": 0.35,
739
+ "learning_rate": 6.530730232144445e-07,
740
+ "loss": 0.1672,
741
+ "step": 7808
742
+ },
743
+ {
744
+ "epoch": 0.35,
745
+ "learning_rate": 6.502268077915147e-07,
746
+ "loss": 0.1719,
747
+ "step": 7872
748
+ },
749
+ {
750
+ "epoch": 0.35,
751
+ "learning_rate": 6.473805923685848e-07,
752
+ "loss": 0.1606,
753
+ "step": 7936
754
+ },
755
+ {
756
+ "epoch": 0.36,
757
+ "learning_rate": 6.44534376945655e-07,
758
+ "loss": 0.1673,
759
+ "step": 8000
760
+ },
761
+ {
762
+ "epoch": 0.36,
763
+ "learning_rate": 6.416881615227252e-07,
764
+ "loss": 0.1494,
765
+ "step": 8064
766
+ },
767
+ {
768
+ "epoch": 0.36,
769
+ "learning_rate": 6.388419460997953e-07,
770
+ "loss": 0.1167,
771
+ "step": 8128
772
+ },
773
+ {
774
+ "epoch": 0.36,
775
+ "learning_rate": 6.359957306768656e-07,
776
+ "loss": 0.1323,
777
+ "step": 8192
778
+ },
779
+ {
780
+ "epoch": 0.37,
781
+ "learning_rate": 6.331495152539358e-07,
782
+ "loss": 0.2502,
783
+ "step": 8256
784
+ },
785
+ {
786
+ "epoch": 0.37,
787
+ "learning_rate": 6.30303299831006e-07,
788
+ "loss": 0.1697,
789
+ "step": 8320
790
+ },
791
+ {
792
+ "epoch": 0.37,
793
+ "learning_rate": 6.274570844080761e-07,
794
+ "loss": 0.1476,
795
+ "step": 8384
796
+ },
797
+ {
798
+ "epoch": 0.38,
799
+ "learning_rate": 6.246108689851463e-07,
800
+ "loss": 0.1454,
801
+ "step": 8448
802
+ },
803
+ {
804
+ "epoch": 0.38,
805
+ "learning_rate": 6.217646535622164e-07,
806
+ "loss": 0.1052,
807
+ "step": 8512
808
+ },
809
+ {
810
+ "epoch": 0.38,
811
+ "learning_rate": 6.1896291025527e-07,
812
+ "loss": 0.1477,
813
+ "step": 8576
814
+ },
815
+ {
816
+ "epoch": 0.38,
817
+ "learning_rate": 6.161166948323401e-07,
818
+ "loss": 0.1992,
819
+ "step": 8640
820
+ },
821
+ {
822
+ "epoch": 0.39,
823
+ "learning_rate": 6.132704794094103e-07,
824
+ "loss": 0.1662,
825
+ "step": 8704
826
+ },
827
+ {
828
+ "epoch": 0.39,
829
+ "learning_rate": 6.104242639864806e-07,
830
+ "loss": 0.1351,
831
+ "step": 8768
832
+ },
833
+ {
834
+ "epoch": 0.39,
835
+ "learning_rate": 6.075780485635506e-07,
836
+ "loss": 0.1551,
837
+ "step": 8832
838
+ },
839
+ {
840
+ "epoch": 0.4,
841
+ "learning_rate": 6.047318331406208e-07,
842
+ "loss": 0.183,
843
+ "step": 8896
844
+ },
845
+ {
846
+ "epoch": 0.4,
847
+ "learning_rate": 6.018856177176909e-07,
848
+ "loss": 0.1767,
849
+ "step": 8960
850
+ },
851
+ {
852
+ "epoch": 0.4,
853
+ "learning_rate": 5.990394022947611e-07,
854
+ "loss": 0.1564,
855
+ "step": 9024
856
+ },
857
+ {
858
+ "epoch": 0.4,
859
+ "learning_rate": 5.961931868718313e-07,
860
+ "loss": 0.1972,
861
+ "step": 9088
862
+ },
863
+ {
864
+ "epoch": 0.41,
865
+ "learning_rate": 5.933469714489015e-07,
866
+ "loss": 0.1867,
867
+ "step": 9152
868
+ },
869
+ {
870
+ "epoch": 0.41,
871
+ "learning_rate": 5.905007560259717e-07,
872
+ "loss": 0.2061,
873
+ "step": 9216
874
+ },
875
+ {
876
+ "epoch": 0.41,
877
+ "learning_rate": 5.876545406030419e-07,
878
+ "loss": 0.1744,
879
+ "step": 9280
880
+ },
881
+ {
882
+ "epoch": 0.42,
883
+ "learning_rate": 5.84808325180112e-07,
884
+ "loss": 0.1119,
885
+ "step": 9344
886
+ },
887
+ {
888
+ "epoch": 0.42,
889
+ "learning_rate": 5.819621097571822e-07,
890
+ "loss": 0.1835,
891
+ "step": 9408
892
+ },
893
+ {
894
+ "epoch": 0.42,
895
+ "learning_rate": 5.791158943342525e-07,
896
+ "loss": 0.1975,
897
+ "step": 9472
898
+ },
899
+ {
900
+ "epoch": 0.42,
901
+ "learning_rate": 5.762696789113226e-07,
902
+ "loss": 0.1862,
903
+ "step": 9536
904
+ },
905
+ {
906
+ "epoch": 0.43,
907
+ "learning_rate": 5.734234634883928e-07,
908
+ "loss": 0.1508,
909
+ "step": 9600
910
+ },
911
+ {
912
+ "epoch": 0.43,
913
+ "learning_rate": 5.705772480654629e-07,
914
+ "loss": 0.144,
915
+ "step": 9664
916
+ },
917
+ {
918
+ "epoch": 0.43,
919
+ "learning_rate": 5.67731032642533e-07,
920
+ "loss": 0.1776,
921
+ "step": 9728
922
+ },
923
+ {
924
+ "epoch": 0.44,
925
+ "learning_rate": 5.648848172196032e-07,
926
+ "loss": 0.1483,
927
+ "step": 9792
928
+ },
929
+ {
930
+ "epoch": 0.44,
931
+ "learning_rate": 5.620386017966735e-07,
932
+ "loss": 0.1155,
933
+ "step": 9856
934
+ },
935
+ {
936
+ "epoch": 0.44,
937
+ "learning_rate": 5.591923863737436e-07,
938
+ "loss": 0.1515,
939
+ "step": 9920
940
+ },
941
+ {
942
+ "epoch": 0.44,
943
+ "learning_rate": 5.563461709508138e-07,
944
+ "loss": 0.2434,
945
+ "step": 9984
946
+ },
947
+ {
948
+ "epoch": 0.45,
949
+ "learning_rate": 5.53499955527884e-07,
950
+ "loss": 0.1918,
951
+ "step": 10048
952
+ },
953
+ {
954
+ "epoch": 0.45,
955
+ "learning_rate": 5.506537401049541e-07,
956
+ "loss": 0.1944,
957
+ "step": 10112
958
+ },
959
+ {
960
+ "epoch": 0.45,
961
+ "learning_rate": 5.478075246820243e-07,
962
+ "loss": 0.1781,
963
+ "step": 10176
964
+ },
965
+ {
966
+ "epoch": 0.46,
967
+ "learning_rate": 5.449613092590946e-07,
968
+ "loss": 0.2186,
969
+ "step": 10240
970
+ },
971
+ {
972
+ "epoch": 0.46,
973
+ "learning_rate": 5.421150938361647e-07,
974
+ "loss": 0.1591,
975
+ "step": 10304
976
+ },
977
+ {
978
+ "epoch": 0.46,
979
+ "learning_rate": 5.392688784132349e-07,
980
+ "loss": 0.1282,
981
+ "step": 10368
982
+ },
983
+ {
984
+ "epoch": 0.46,
985
+ "learning_rate": 5.364226629903051e-07,
986
+ "loss": 0.23,
987
+ "step": 10432
988
+ },
989
+ {
990
+ "epoch": 0.47,
991
+ "learning_rate": 5.335764475673751e-07,
992
+ "loss": 0.1707,
993
+ "step": 10496
994
+ },
995
+ {
996
+ "epoch": 0.47,
997
+ "learning_rate": 5.307302321444454e-07,
998
+ "loss": 0.179,
999
+ "step": 10560
1000
+ },
1001
+ {
1002
+ "epoch": 0.47,
1003
+ "learning_rate": 5.278840167215156e-07,
1004
+ "loss": 0.1568,
1005
+ "step": 10624
1006
+ },
1007
+ {
1008
+ "epoch": 0.48,
1009
+ "learning_rate": 5.250378012985857e-07,
1010
+ "loss": 0.0832,
1011
+ "step": 10688
1012
+ },
1013
+ {
1014
+ "epoch": 0.48,
1015
+ "learning_rate": 5.221915858756559e-07,
1016
+ "loss": 0.1286,
1017
+ "step": 10752
1018
+ },
1019
+ {
1020
+ "epoch": 0.48,
1021
+ "learning_rate": 5.193898425687094e-07,
1022
+ "loss": 0.1397,
1023
+ "step": 10816
1024
+ },
1025
+ {
1026
+ "epoch": 0.48,
1027
+ "learning_rate": 5.165436271457796e-07,
1028
+ "loss": 0.1383,
1029
+ "step": 10880
1030
+ },
1031
+ {
1032
+ "epoch": 0.49,
1033
+ "learning_rate": 5.136974117228497e-07,
1034
+ "loss": 0.2332,
1035
+ "step": 10944
1036
+ },
1037
+ {
1038
+ "epoch": 0.49,
1039
+ "learning_rate": 5.108511962999199e-07,
1040
+ "loss": 0.1151,
1041
+ "step": 11008
1042
+ },
1043
+ {
1044
+ "epoch": 0.49,
1045
+ "learning_rate": 5.080049808769901e-07,
1046
+ "loss": 0.1769,
1047
+ "step": 11072
1048
+ },
1049
+ {
1050
+ "epoch": 0.5,
1051
+ "learning_rate": 5.051587654540603e-07,
1052
+ "loss": 0.1346,
1053
+ "step": 11136
1054
+ },
1055
+ {
1056
+ "epoch": 0.5,
1057
+ "learning_rate": 5.023125500311305e-07,
1058
+ "loss": 0.2568,
1059
+ "step": 11200
1060
+ },
1061
+ {
1062
+ "epoch": 0.5,
1063
+ "learning_rate": 4.994663346082007e-07,
1064
+ "loss": 0.1216,
1065
+ "step": 11264
1066
+ },
1067
+ {
1068
+ "epoch": 0.5,
1069
+ "learning_rate": 4.966201191852708e-07,
1070
+ "loss": 0.1167,
1071
+ "step": 11328
1072
+ },
1073
+ {
1074
+ "epoch": 0.51,
1075
+ "learning_rate": 4.93773903762341e-07,
1076
+ "loss": 0.1752,
1077
+ "step": 11392
1078
+ },
1079
+ {
1080
+ "epoch": 0.51,
1081
+ "learning_rate": 4.909276883394112e-07,
1082
+ "loss": 0.3198,
1083
+ "step": 11456
1084
+ },
1085
+ {
1086
+ "epoch": 0.51,
1087
+ "learning_rate": 4.880814729164813e-07,
1088
+ "loss": 0.2023,
1089
+ "step": 11520
1090
+ },
1091
+ {
1092
+ "epoch": 0.52,
1093
+ "learning_rate": 4.852352574935515e-07,
1094
+ "loss": 0.1261,
1095
+ "step": 11584
1096
+ },
1097
+ {
1098
+ "epoch": 0.52,
1099
+ "learning_rate": 4.823890420706217e-07,
1100
+ "loss": 0.0972,
1101
+ "step": 11648
1102
+ },
1103
+ {
1104
+ "epoch": 0.52,
1105
+ "learning_rate": 4.795428266476918e-07,
1106
+ "loss": 0.1692,
1107
+ "step": 11712
1108
+ },
1109
+ {
1110
+ "epoch": 0.52,
1111
+ "learning_rate": 4.7669661122476206e-07,
1112
+ "loss": 0.2509,
1113
+ "step": 11776
1114
+ },
1115
+ {
1116
+ "epoch": 0.53,
1117
+ "learning_rate": 4.7385039580183223e-07,
1118
+ "loss": 0.165,
1119
+ "step": 11840
1120
+ },
1121
+ {
1122
+ "epoch": 0.53,
1123
+ "learning_rate": 4.710041803789024e-07,
1124
+ "loss": 0.102,
1125
+ "step": 11904
1126
+ },
1127
+ {
1128
+ "epoch": 0.53,
1129
+ "learning_rate": 4.681579649559726e-07,
1130
+ "loss": 0.1197,
1131
+ "step": 11968
1132
+ },
1133
+ {
1134
+ "epoch": 0.54,
1135
+ "learning_rate": 4.653117495330428e-07,
1136
+ "loss": 0.2289,
1137
+ "step": 12032
1138
+ },
1139
+ {
1140
+ "epoch": 0.54,
1141
+ "learning_rate": 4.6246553411011294e-07,
1142
+ "loss": 0.1477,
1143
+ "step": 12096
1144
+ },
1145
+ {
1146
+ "epoch": 0.54,
1147
+ "learning_rate": 4.596193186871831e-07,
1148
+ "loss": 0.1813,
1149
+ "step": 12160
1150
+ },
1151
+ {
1152
+ "epoch": 0.54,
1153
+ "learning_rate": 4.567731032642533e-07,
1154
+ "loss": 0.2105,
1155
+ "step": 12224
1156
+ },
1157
+ {
1158
+ "epoch": 0.55,
1159
+ "learning_rate": 4.539268878413235e-07,
1160
+ "loss": 0.2905,
1161
+ "step": 12288
1162
+ },
1163
+ {
1164
+ "epoch": 0.55,
1165
+ "learning_rate": 4.5112514453437694e-07,
1166
+ "loss": 0.1313,
1167
+ "step": 12352
1168
+ },
1169
+ {
1170
+ "epoch": 0.55,
1171
+ "learning_rate": 4.482789291114471e-07,
1172
+ "loss": 0.188,
1173
+ "step": 12416
1174
+ },
1175
+ {
1176
+ "epoch": 0.56,
1177
+ "learning_rate": 4.454327136885173e-07,
1178
+ "loss": 0.1242,
1179
+ "step": 12480
1180
+ },
1181
+ {
1182
+ "epoch": 0.56,
1183
+ "learning_rate": 4.425864982655875e-07,
1184
+ "loss": 0.1764,
1185
+ "step": 12544
1186
+ },
1187
+ {
1188
+ "epoch": 0.56,
1189
+ "learning_rate": 4.397402828426576e-07,
1190
+ "loss": 0.2115,
1191
+ "step": 12608
1192
+ },
1193
+ {
1194
+ "epoch": 0.56,
1195
+ "learning_rate": 4.368940674197278e-07,
1196
+ "loss": 0.1478,
1197
+ "step": 12672
1198
+ },
1199
+ {
1200
+ "epoch": 0.57,
1201
+ "learning_rate": 4.34047851996798e-07,
1202
+ "loss": 0.2097,
1203
+ "step": 12736
1204
+ },
1205
+ {
1206
+ "epoch": 0.57,
1207
+ "learning_rate": 4.3120163657386816e-07,
1208
+ "loss": 0.2395,
1209
+ "step": 12800
1210
+ },
1211
+ {
1212
+ "epoch": 0.57,
1213
+ "learning_rate": 4.283554211509384e-07,
1214
+ "loss": 0.1679,
1215
+ "step": 12864
1216
+ },
1217
+ {
1218
+ "epoch": 0.57,
1219
+ "learning_rate": 4.2550920572800854e-07,
1220
+ "loss": 0.1273,
1221
+ "step": 12928
1222
+ },
1223
+ {
1224
+ "epoch": 0.58,
1225
+ "learning_rate": 4.2266299030507866e-07,
1226
+ "loss": 0.1552,
1227
+ "step": 12992
1228
+ },
1229
+ {
1230
+ "epoch": 0.58,
1231
+ "learning_rate": 4.198167748821489e-07,
1232
+ "loss": 0.1337,
1233
+ "step": 13056
1234
+ },
1235
+ {
1236
+ "epoch": 0.58,
1237
+ "learning_rate": 4.1697055945921904e-07,
1238
+ "loss": 0.1604,
1239
+ "step": 13120
1240
+ },
1241
+ {
1242
+ "epoch": 0.59,
1243
+ "learning_rate": 4.141243440362892e-07,
1244
+ "loss": 0.2796,
1245
+ "step": 13184
1246
+ },
1247
+ {
1248
+ "epoch": 0.59,
1249
+ "learning_rate": 4.1127812861335943e-07,
1250
+ "loss": 0.1497,
1251
+ "step": 13248
1252
+ },
1253
+ {
1254
+ "epoch": 0.59,
1255
+ "learning_rate": 4.084319131904296e-07,
1256
+ "loss": 0.1414,
1257
+ "step": 13312
1258
+ },
1259
+ {
1260
+ "epoch": 0.59,
1261
+ "learning_rate": 4.0558569776749976e-07,
1262
+ "loss": 0.1435,
1263
+ "step": 13376
1264
+ },
1265
+ {
1266
+ "epoch": 0.6,
1267
+ "learning_rate": 4.027394823445699e-07,
1268
+ "loss": 0.1435,
1269
+ "step": 13440
1270
+ },
1271
+ {
1272
+ "epoch": 0.6,
1273
+ "learning_rate": 3.998932669216401e-07,
1274
+ "loss": 0.222,
1275
+ "step": 13504
1276
+ },
1277
+ {
1278
+ "epoch": 0.6,
1279
+ "learning_rate": 3.970470514987103e-07,
1280
+ "loss": 0.1387,
1281
+ "step": 13568
1282
+ },
1283
+ {
1284
+ "epoch": 0.61,
1285
+ "learning_rate": 3.942008360757805e-07,
1286
+ "loss": 0.144,
1287
+ "step": 13632
1288
+ },
1289
+ {
1290
+ "epoch": 0.61,
1291
+ "learning_rate": 3.9135462065285064e-07,
1292
+ "loss": 0.1839,
1293
+ "step": 13696
1294
+ },
1295
+ {
1296
+ "epoch": 0.61,
1297
+ "learning_rate": 3.8850840522992086e-07,
1298
+ "loss": 0.3017,
1299
+ "step": 13760
1300
+ },
1301
+ {
1302
+ "epoch": 0.61,
1303
+ "learning_rate": 3.85662189806991e-07,
1304
+ "loss": 0.1267,
1305
+ "step": 13824
1306
+ },
1307
+ {
1308
+ "epoch": 0.62,
1309
+ "learning_rate": 3.8281597438406114e-07,
1310
+ "loss": 0.1733,
1311
+ "step": 13888
1312
+ },
1313
+ {
1314
+ "epoch": 0.62,
1315
+ "learning_rate": 3.7996975896113136e-07,
1316
+ "loss": 0.2575,
1317
+ "step": 13952
1318
+ },
1319
+ {
1320
+ "epoch": 0.62,
1321
+ "learning_rate": 3.771235435382015e-07,
1322
+ "loss": 0.1702,
1323
+ "step": 14016
1324
+ },
1325
+ {
1326
+ "epoch": 0.63,
1327
+ "learning_rate": 3.742773281152717e-07,
1328
+ "loss": 0.1609,
1329
+ "step": 14080
1330
+ },
1331
+ {
1332
+ "epoch": 0.63,
1333
+ "learning_rate": 3.714311126923419e-07,
1334
+ "loss": 0.1917,
1335
+ "step": 14144
1336
+ },
1337
+ {
1338
+ "epoch": 0.63,
1339
+ "learning_rate": 3.685848972694121e-07,
1340
+ "loss": 0.1087,
1341
+ "step": 14208
1342
+ },
1343
+ {
1344
+ "epoch": 0.63,
1345
+ "learning_rate": 3.6573868184648224e-07,
1346
+ "loss": 0.1452,
1347
+ "step": 14272
1348
+ },
1349
+ {
1350
+ "epoch": 0.64,
1351
+ "learning_rate": 3.628924664235524e-07,
1352
+ "loss": 0.1491,
1353
+ "step": 14336
1354
+ },
1355
+ {
1356
+ "epoch": 0.64,
1357
+ "learning_rate": 3.6004625100062257e-07,
1358
+ "loss": 0.1735,
1359
+ "step": 14400
1360
+ },
1361
+ {
1362
+ "epoch": 0.64,
1363
+ "learning_rate": 3.572000355776928e-07,
1364
+ "loss": 0.2468,
1365
+ "step": 14464
1366
+ },
1367
+ {
1368
+ "epoch": 0.65,
1369
+ "learning_rate": 3.5435382015476296e-07,
1370
+ "loss": 0.1958,
1371
+ "step": 14528
1372
+ },
1373
+ {
1374
+ "epoch": 0.65,
1375
+ "learning_rate": 3.515076047318331e-07,
1376
+ "loss": 0.1487,
1377
+ "step": 14592
1378
+ },
1379
+ {
1380
+ "epoch": 0.65,
1381
+ "learning_rate": 3.4866138930890334e-07,
1382
+ "loss": 0.1503,
1383
+ "step": 14656
1384
+ },
1385
+ {
1386
+ "epoch": 0.65,
1387
+ "learning_rate": 3.4581517388597346e-07,
1388
+ "loss": 0.184,
1389
+ "step": 14720
1390
+ },
1391
+ {
1392
+ "epoch": 0.66,
1393
+ "learning_rate": 3.429689584630436e-07,
1394
+ "loss": 0.2188,
1395
+ "step": 14784
1396
+ },
1397
+ {
1398
+ "epoch": 0.66,
1399
+ "learning_rate": 3.4012274304011384e-07,
1400
+ "loss": 0.1971,
1401
+ "step": 14848
1402
+ },
1403
+ {
1404
+ "epoch": 0.66,
1405
+ "learning_rate": 3.37276527617184e-07,
1406
+ "loss": 0.1236,
1407
+ "step": 14912
1408
+ },
1409
+ {
1410
+ "epoch": 0.67,
1411
+ "learning_rate": 3.3443031219425417e-07,
1412
+ "loss": 0.2584,
1413
+ "step": 14976
1414
+ }
1415
+ ],
1416
+ "logging_steps": 64,
1417
+ "max_steps": 22486,
1418
+ "num_input_tokens_seen": 0,
1419
+ "num_train_epochs": 1,
1420
+ "save_steps": 7495,
1421
+ "total_flos": 3.1826333601792e+19,
1422
+ "train_batch_size": 1,
1423
+ "trial_name": null,
1424
+ "trial_params": null
1425
+ }
checkpoint-14990/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5dd66348aef7664171d12847300a1c626bfa2b1ede8e5141c5a2fc70bac4b83
3
+ size 5048
checkpoint-22485/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-22485/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-22485/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a3ea9b7dc3ada6e3f7ff8b69da24968fc459e62e8357a7ec7b21a299a0a1e16
3
+ size 4992706480
checkpoint-22485/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbab5dbea0448455ca6638599f8024a646277efe8c2cf7039b3ecbbb0f31e691
3
+ size 1180663192
checkpoint-22485/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-22485/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa9821acf2b783e9cf8e44422c5f79c3e8a537c62b88ec752e27c26c835fd002
3
+ size 3095074288
checkpoint-22485/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-22485/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eba910cb5dfcf9065fef36e165af0c0afe2ea49edd17d1923e981ff6fa250eb2
3
+ size 14244
checkpoint-22485/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a579e384f224a8b5f499118fbc635fa6555a3f4e189b3b7d08d3351420e2f0a4
3
+ size 1064
checkpoint-22485/trainer_state.json ADDED
@@ -0,0 +1,2127 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9999555278840168,
5
+ "eval_steps": 500,
6
+ "global_step": 22485,
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": 9.973761451569865e-07,
14
+ "loss": 3.933,
15
+ "step": 64
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 9.945299297340568e-07,
20
+ "loss": 1.6429,
21
+ "step": 128
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 9.917726585430935e-07,
26
+ "loss": 1.1012,
27
+ "step": 192
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 9.889264431201636e-07,
32
+ "loss": 0.3376,
33
+ "step": 256
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 9.860802276972338e-07,
38
+ "loss": 0.136,
39
+ "step": 320
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 9.832340122743039e-07,
44
+ "loss": 0.2753,
45
+ "step": 384
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 9.803877968513742e-07,
50
+ "loss": 0.25,
51
+ "step": 448
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 9.775415814284444e-07,
56
+ "loss": 0.3117,
57
+ "step": 512
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 9.746953660055145e-07,
62
+ "loss": 0.2853,
63
+ "step": 576
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 9.718491505825848e-07,
68
+ "loss": 0.2489,
69
+ "step": 640
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 9.690029351596548e-07,
74
+ "loss": 0.217,
75
+ "step": 704
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 9.66156719736725e-07,
80
+ "loss": 0.1804,
81
+ "step": 768
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 9.633105043137952e-07,
86
+ "loss": 0.2842,
87
+ "step": 832
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 9.604642888908654e-07,
92
+ "loss": 0.2621,
93
+ "step": 896
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 9.576180734679355e-07,
98
+ "loss": 0.3316,
99
+ "step": 960
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 9.547718580450058e-07,
104
+ "loss": 0.1856,
105
+ "step": 1024
106
+ },
107
+ {
108
+ "epoch": 0.05,
109
+ "learning_rate": 9.519256426220758e-07,
110
+ "loss": 0.2112,
111
+ "step": 1088
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 9.490794271991461e-07,
116
+ "loss": 0.1778,
117
+ "step": 1152
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 9.462332117762163e-07,
122
+ "loss": 0.319,
123
+ "step": 1216
124
+ },
125
+ {
126
+ "epoch": 0.06,
127
+ "learning_rate": 9.433869963532864e-07,
128
+ "loss": 0.1678,
129
+ "step": 1280
130
+ },
131
+ {
132
+ "epoch": 0.06,
133
+ "learning_rate": 9.405407809303566e-07,
134
+ "loss": 0.2549,
135
+ "step": 1344
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.376945655074268e-07,
140
+ "loss": 0.1723,
141
+ "step": 1408
142
+ },
143
+ {
144
+ "epoch": 0.07,
145
+ "learning_rate": 9.34848350084497e-07,
146
+ "loss": 0.149,
147
+ "step": 1472
148
+ },
149
+ {
150
+ "epoch": 0.07,
151
+ "learning_rate": 9.320021346615672e-07,
152
+ "loss": 0.2097,
153
+ "step": 1536
154
+ },
155
+ {
156
+ "epoch": 0.07,
157
+ "learning_rate": 9.291559192386374e-07,
158
+ "loss": 0.3716,
159
+ "step": 1600
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.263097038157075e-07,
164
+ "loss": 0.1303,
165
+ "step": 1664
166
+ },
167
+ {
168
+ "epoch": 0.08,
169
+ "learning_rate": 9.234634883927777e-07,
170
+ "loss": 0.2041,
171
+ "step": 1728
172
+ },
173
+ {
174
+ "epoch": 0.08,
175
+ "learning_rate": 9.206172729698479e-07,
176
+ "loss": 0.2529,
177
+ "step": 1792
178
+ },
179
+ {
180
+ "epoch": 0.08,
181
+ "learning_rate": 9.177710575469181e-07,
182
+ "loss": 0.2241,
183
+ "step": 1856
184
+ },
185
+ {
186
+ "epoch": 0.09,
187
+ "learning_rate": 9.149248421239883e-07,
188
+ "loss": 0.2056,
189
+ "step": 1920
190
+ },
191
+ {
192
+ "epoch": 0.09,
193
+ "learning_rate": 9.120786267010584e-07,
194
+ "loss": 0.2956,
195
+ "step": 1984
196
+ },
197
+ {
198
+ "epoch": 0.09,
199
+ "learning_rate": 9.092324112781285e-07,
200
+ "loss": 0.2177,
201
+ "step": 2048
202
+ },
203
+ {
204
+ "epoch": 0.09,
205
+ "learning_rate": 9.063861958551987e-07,
206
+ "loss": 0.2379,
207
+ "step": 2112
208
+ },
209
+ {
210
+ "epoch": 0.1,
211
+ "learning_rate": 9.03539980432269e-07,
212
+ "loss": 0.3499,
213
+ "step": 2176
214
+ },
215
+ {
216
+ "epoch": 0.1,
217
+ "learning_rate": 9.006937650093391e-07,
218
+ "loss": 0.2133,
219
+ "step": 2240
220
+ },
221
+ {
222
+ "epoch": 0.1,
223
+ "learning_rate": 8.978475495864093e-07,
224
+ "loss": 0.283,
225
+ "step": 2304
226
+ },
227
+ {
228
+ "epoch": 0.11,
229
+ "learning_rate": 8.950013341634795e-07,
230
+ "loss": 0.2337,
231
+ "step": 2368
232
+ },
233
+ {
234
+ "epoch": 0.11,
235
+ "learning_rate": 8.921551187405496e-07,
236
+ "loss": 0.1282,
237
+ "step": 2432
238
+ },
239
+ {
240
+ "epoch": 0.11,
241
+ "learning_rate": 8.893089033176198e-07,
242
+ "loss": 0.1615,
243
+ "step": 2496
244
+ },
245
+ {
246
+ "epoch": 0.11,
247
+ "learning_rate": 8.864626878946901e-07,
248
+ "loss": 0.1537,
249
+ "step": 2560
250
+ },
251
+ {
252
+ "epoch": 0.12,
253
+ "learning_rate": 8.836164724717602e-07,
254
+ "loss": 0.2078,
255
+ "step": 2624
256
+ },
257
+ {
258
+ "epoch": 0.12,
259
+ "learning_rate": 8.807702570488304e-07,
260
+ "loss": 0.2886,
261
+ "step": 2688
262
+ },
263
+ {
264
+ "epoch": 0.12,
265
+ "learning_rate": 8.779240416259006e-07,
266
+ "loss": 0.2204,
267
+ "step": 2752
268
+ },
269
+ {
270
+ "epoch": 0.13,
271
+ "learning_rate": 8.750778262029706e-07,
272
+ "loss": 0.2415,
273
+ "step": 2816
274
+ },
275
+ {
276
+ "epoch": 0.13,
277
+ "learning_rate": 8.722316107800409e-07,
278
+ "loss": 0.1124,
279
+ "step": 2880
280
+ },
281
+ {
282
+ "epoch": 0.13,
283
+ "learning_rate": 8.69385395357111e-07,
284
+ "loss": 0.1638,
285
+ "step": 2944
286
+ },
287
+ {
288
+ "epoch": 0.13,
289
+ "learning_rate": 8.665391799341812e-07,
290
+ "loss": 0.2198,
291
+ "step": 3008
292
+ },
293
+ {
294
+ "epoch": 0.14,
295
+ "learning_rate": 8.636929645112514e-07,
296
+ "loss": 0.1793,
297
+ "step": 3072
298
+ },
299
+ {
300
+ "epoch": 0.14,
301
+ "learning_rate": 8.608467490883215e-07,
302
+ "loss": 0.1939,
303
+ "step": 3136
304
+ },
305
+ {
306
+ "epoch": 0.14,
307
+ "learning_rate": 8.580005336653917e-07,
308
+ "loss": 0.2771,
309
+ "step": 3200
310
+ },
311
+ {
312
+ "epoch": 0.15,
313
+ "learning_rate": 8.55154318242462e-07,
314
+ "loss": 0.1291,
315
+ "step": 3264
316
+ },
317
+ {
318
+ "epoch": 0.15,
319
+ "learning_rate": 8.523081028195322e-07,
320
+ "loss": 0.216,
321
+ "step": 3328
322
+ },
323
+ {
324
+ "epoch": 0.15,
325
+ "learning_rate": 8.494618873966023e-07,
326
+ "loss": 0.1985,
327
+ "step": 3392
328
+ },
329
+ {
330
+ "epoch": 0.15,
331
+ "learning_rate": 8.466156719736725e-07,
332
+ "loss": 0.2458,
333
+ "step": 3456
334
+ },
335
+ {
336
+ "epoch": 0.16,
337
+ "learning_rate": 8.437694565507427e-07,
338
+ "loss": 0.1253,
339
+ "step": 3520
340
+ },
341
+ {
342
+ "epoch": 0.16,
343
+ "learning_rate": 8.409232411278129e-07,
344
+ "loss": 0.2109,
345
+ "step": 3584
346
+ },
347
+ {
348
+ "epoch": 0.16,
349
+ "learning_rate": 8.38077025704883e-07,
350
+ "loss": 0.2105,
351
+ "step": 3648
352
+ },
353
+ {
354
+ "epoch": 0.17,
355
+ "learning_rate": 8.352308102819531e-07,
356
+ "loss": 0.2013,
357
+ "step": 3712
358
+ },
359
+ {
360
+ "epoch": 0.17,
361
+ "learning_rate": 8.323845948590233e-07,
362
+ "loss": 0.203,
363
+ "step": 3776
364
+ },
365
+ {
366
+ "epoch": 0.17,
367
+ "learning_rate": 8.295383794360935e-07,
368
+ "loss": 0.2366,
369
+ "step": 3840
370
+ },
371
+ {
372
+ "epoch": 0.17,
373
+ "learning_rate": 8.266921640131636e-07,
374
+ "loss": 0.2181,
375
+ "step": 3904
376
+ },
377
+ {
378
+ "epoch": 0.18,
379
+ "learning_rate": 8.238459485902339e-07,
380
+ "loss": 0.2611,
381
+ "step": 3968
382
+ },
383
+ {
384
+ "epoch": 0.18,
385
+ "learning_rate": 8.209997331673041e-07,
386
+ "loss": 0.1412,
387
+ "step": 4032
388
+ },
389
+ {
390
+ "epoch": 0.18,
391
+ "learning_rate": 8.181535177443743e-07,
392
+ "loss": 0.2678,
393
+ "step": 4096
394
+ },
395
+ {
396
+ "epoch": 0.19,
397
+ "learning_rate": 8.153073023214444e-07,
398
+ "loss": 0.1847,
399
+ "step": 4160
400
+ },
401
+ {
402
+ "epoch": 0.19,
403
+ "learning_rate": 8.124610868985146e-07,
404
+ "loss": 0.2395,
405
+ "step": 4224
406
+ },
407
+ {
408
+ "epoch": 0.19,
409
+ "learning_rate": 8.096148714755849e-07,
410
+ "loss": 0.1473,
411
+ "step": 4288
412
+ },
413
+ {
414
+ "epoch": 0.19,
415
+ "learning_rate": 8.06768656052655e-07,
416
+ "loss": 0.2675,
417
+ "step": 4352
418
+ },
419
+ {
420
+ "epoch": 0.2,
421
+ "learning_rate": 8.039224406297252e-07,
422
+ "loss": 0.2179,
423
+ "step": 4416
424
+ },
425
+ {
426
+ "epoch": 0.2,
427
+ "learning_rate": 8.010762252067952e-07,
428
+ "loss": 0.1667,
429
+ "step": 4480
430
+ },
431
+ {
432
+ "epoch": 0.2,
433
+ "learning_rate": 7.982300097838654e-07,
434
+ "loss": 0.2114,
435
+ "step": 4544
436
+ },
437
+ {
438
+ "epoch": 0.2,
439
+ "learning_rate": 7.953837943609356e-07,
440
+ "loss": 0.1929,
441
+ "step": 4608
442
+ },
443
+ {
444
+ "epoch": 0.21,
445
+ "learning_rate": 7.925375789380059e-07,
446
+ "loss": 0.2152,
447
+ "step": 4672
448
+ },
449
+ {
450
+ "epoch": 0.21,
451
+ "learning_rate": 7.89691363515076e-07,
452
+ "loss": 0.2065,
453
+ "step": 4736
454
+ },
455
+ {
456
+ "epoch": 0.21,
457
+ "learning_rate": 7.868451480921462e-07,
458
+ "loss": 0.1607,
459
+ "step": 4800
460
+ },
461
+ {
462
+ "epoch": 0.22,
463
+ "learning_rate": 7.839989326692163e-07,
464
+ "loss": 0.1592,
465
+ "step": 4864
466
+ },
467
+ {
468
+ "epoch": 0.22,
469
+ "learning_rate": 7.811527172462865e-07,
470
+ "loss": 0.2558,
471
+ "step": 4928
472
+ },
473
+ {
474
+ "epoch": 0.22,
475
+ "learning_rate": 7.783065018233567e-07,
476
+ "loss": 0.1564,
477
+ "step": 4992
478
+ },
479
+ {
480
+ "epoch": 0.22,
481
+ "learning_rate": 7.75460286400427e-07,
482
+ "loss": 0.2171,
483
+ "step": 5056
484
+ },
485
+ {
486
+ "epoch": 0.23,
487
+ "learning_rate": 7.726140709774971e-07,
488
+ "loss": 0.1973,
489
+ "step": 5120
490
+ },
491
+ {
492
+ "epoch": 0.23,
493
+ "learning_rate": 7.697678555545673e-07,
494
+ "loss": 0.1273,
495
+ "step": 5184
496
+ },
497
+ {
498
+ "epoch": 0.23,
499
+ "learning_rate": 7.669216401316375e-07,
500
+ "loss": 0.1158,
501
+ "step": 5248
502
+ },
503
+ {
504
+ "epoch": 0.24,
505
+ "learning_rate": 7.640754247087075e-07,
506
+ "loss": 0.151,
507
+ "step": 5312
508
+ },
509
+ {
510
+ "epoch": 0.24,
511
+ "learning_rate": 7.612292092857778e-07,
512
+ "loss": 0.2727,
513
+ "step": 5376
514
+ },
515
+ {
516
+ "epoch": 0.24,
517
+ "learning_rate": 7.58382993862848e-07,
518
+ "loss": 0.1271,
519
+ "step": 5440
520
+ },
521
+ {
522
+ "epoch": 0.24,
523
+ "learning_rate": 7.555367784399181e-07,
524
+ "loss": 0.2409,
525
+ "step": 5504
526
+ },
527
+ {
528
+ "epoch": 0.25,
529
+ "learning_rate": 7.526905630169883e-07,
530
+ "loss": 0.1456,
531
+ "step": 5568
532
+ },
533
+ {
534
+ "epoch": 0.25,
535
+ "learning_rate": 7.498443475940584e-07,
536
+ "loss": 0.249,
537
+ "step": 5632
538
+ },
539
+ {
540
+ "epoch": 0.25,
541
+ "learning_rate": 7.469981321711286e-07,
542
+ "loss": 0.2348,
543
+ "step": 5696
544
+ },
545
+ {
546
+ "epoch": 0.26,
547
+ "learning_rate": 7.441519167481989e-07,
548
+ "loss": 0.2074,
549
+ "step": 5760
550
+ },
551
+ {
552
+ "epoch": 0.26,
553
+ "learning_rate": 7.41305701325269e-07,
554
+ "loss": 0.1483,
555
+ "step": 5824
556
+ },
557
+ {
558
+ "epoch": 0.26,
559
+ "learning_rate": 7.384594859023392e-07,
560
+ "loss": 0.23,
561
+ "step": 5888
562
+ },
563
+ {
564
+ "epoch": 0.26,
565
+ "learning_rate": 7.356132704794094e-07,
566
+ "loss": 0.1294,
567
+ "step": 5952
568
+ },
569
+ {
570
+ "epoch": 0.27,
571
+ "learning_rate": 7.327670550564795e-07,
572
+ "loss": 0.2088,
573
+ "step": 6016
574
+ },
575
+ {
576
+ "epoch": 0.27,
577
+ "learning_rate": 7.299208396335498e-07,
578
+ "loss": 0.2233,
579
+ "step": 6080
580
+ },
581
+ {
582
+ "epoch": 0.27,
583
+ "learning_rate": 7.2707462421062e-07,
584
+ "loss": 0.1361,
585
+ "step": 6144
586
+ },
587
+ {
588
+ "epoch": 0.28,
589
+ "learning_rate": 7.2422840878769e-07,
590
+ "loss": 0.1664,
591
+ "step": 6208
592
+ },
593
+ {
594
+ "epoch": 0.28,
595
+ "learning_rate": 7.213821933647602e-07,
596
+ "loss": 0.1767,
597
+ "step": 6272
598
+ },
599
+ {
600
+ "epoch": 0.28,
601
+ "learning_rate": 7.185359779418304e-07,
602
+ "loss": 0.1613,
603
+ "step": 6336
604
+ },
605
+ {
606
+ "epoch": 0.28,
607
+ "learning_rate": 7.156897625189005e-07,
608
+ "loss": 0.1776,
609
+ "step": 6400
610
+ },
611
+ {
612
+ "epoch": 0.29,
613
+ "learning_rate": 7.128435470959708e-07,
614
+ "loss": 0.2314,
615
+ "step": 6464
616
+ },
617
+ {
618
+ "epoch": 0.29,
619
+ "learning_rate": 7.09997331673041e-07,
620
+ "loss": 0.1439,
621
+ "step": 6528
622
+ },
623
+ {
624
+ "epoch": 0.29,
625
+ "learning_rate": 7.071511162501111e-07,
626
+ "loss": 0.2195,
627
+ "step": 6592
628
+ },
629
+ {
630
+ "epoch": 0.3,
631
+ "learning_rate": 7.043049008271813e-07,
632
+ "loss": 0.248,
633
+ "step": 6656
634
+ },
635
+ {
636
+ "epoch": 0.3,
637
+ "learning_rate": 7.014586854042515e-07,
638
+ "loss": 0.2165,
639
+ "step": 6720
640
+ },
641
+ {
642
+ "epoch": 0.3,
643
+ "learning_rate": 6.986124699813218e-07,
644
+ "loss": 0.2642,
645
+ "step": 6784
646
+ },
647
+ {
648
+ "epoch": 0.3,
649
+ "learning_rate": 6.957662545583919e-07,
650
+ "loss": 0.236,
651
+ "step": 6848
652
+ },
653
+ {
654
+ "epoch": 0.31,
655
+ "learning_rate": 6.929200391354621e-07,
656
+ "loss": 0.2041,
657
+ "step": 6912
658
+ },
659
+ {
660
+ "epoch": 0.31,
661
+ "learning_rate": 6.900738237125323e-07,
662
+ "loss": 0.1925,
663
+ "step": 6976
664
+ },
665
+ {
666
+ "epoch": 0.31,
667
+ "learning_rate": 6.872276082896023e-07,
668
+ "loss": 0.2164,
669
+ "step": 7040
670
+ },
671
+ {
672
+ "epoch": 0.32,
673
+ "learning_rate": 6.843813928666725e-07,
674
+ "loss": 0.1399,
675
+ "step": 7104
676
+ },
677
+ {
678
+ "epoch": 0.32,
679
+ "learning_rate": 6.815351774437427e-07,
680
+ "loss": 0.1875,
681
+ "step": 7168
682
+ },
683
+ {
684
+ "epoch": 0.32,
685
+ "learning_rate": 6.786889620208129e-07,
686
+ "loss": 0.176,
687
+ "step": 7232
688
+ },
689
+ {
690
+ "epoch": 0.32,
691
+ "learning_rate": 6.758427465978831e-07,
692
+ "loss": 0.2326,
693
+ "step": 7296
694
+ },
695
+ {
696
+ "epoch": 0.33,
697
+ "learning_rate": 6.729965311749532e-07,
698
+ "loss": 0.1681,
699
+ "step": 7360
700
+ },
701
+ {
702
+ "epoch": 0.33,
703
+ "learning_rate": 6.701503157520234e-07,
704
+ "loss": 0.2422,
705
+ "step": 7424
706
+ },
707
+ {
708
+ "epoch": 0.33,
709
+ "learning_rate": 6.673041003290937e-07,
710
+ "loss": 0.1881,
711
+ "step": 7488
712
+ },
713
+ {
714
+ "epoch": 0.34,
715
+ "learning_rate": 6.644578849061639e-07,
716
+ "loss": 0.1716,
717
+ "step": 7552
718
+ },
719
+ {
720
+ "epoch": 0.34,
721
+ "learning_rate": 6.61611669483234e-07,
722
+ "loss": 0.1565,
723
+ "step": 7616
724
+ },
725
+ {
726
+ "epoch": 0.34,
727
+ "learning_rate": 6.587654540603042e-07,
728
+ "loss": 0.223,
729
+ "step": 7680
730
+ },
731
+ {
732
+ "epoch": 0.34,
733
+ "learning_rate": 6.559192386373743e-07,
734
+ "loss": 0.1473,
735
+ "step": 7744
736
+ },
737
+ {
738
+ "epoch": 0.35,
739
+ "learning_rate": 6.530730232144445e-07,
740
+ "loss": 0.1672,
741
+ "step": 7808
742
+ },
743
+ {
744
+ "epoch": 0.35,
745
+ "learning_rate": 6.502268077915147e-07,
746
+ "loss": 0.1719,
747
+ "step": 7872
748
+ },
749
+ {
750
+ "epoch": 0.35,
751
+ "learning_rate": 6.473805923685848e-07,
752
+ "loss": 0.1606,
753
+ "step": 7936
754
+ },
755
+ {
756
+ "epoch": 0.36,
757
+ "learning_rate": 6.44534376945655e-07,
758
+ "loss": 0.1673,
759
+ "step": 8000
760
+ },
761
+ {
762
+ "epoch": 0.36,
763
+ "learning_rate": 6.416881615227252e-07,
764
+ "loss": 0.1494,
765
+ "step": 8064
766
+ },
767
+ {
768
+ "epoch": 0.36,
769
+ "learning_rate": 6.388419460997953e-07,
770
+ "loss": 0.1167,
771
+ "step": 8128
772
+ },
773
+ {
774
+ "epoch": 0.36,
775
+ "learning_rate": 6.359957306768656e-07,
776
+ "loss": 0.1323,
777
+ "step": 8192
778
+ },
779
+ {
780
+ "epoch": 0.37,
781
+ "learning_rate": 6.331495152539358e-07,
782
+ "loss": 0.2502,
783
+ "step": 8256
784
+ },
785
+ {
786
+ "epoch": 0.37,
787
+ "learning_rate": 6.30303299831006e-07,
788
+ "loss": 0.1697,
789
+ "step": 8320
790
+ },
791
+ {
792
+ "epoch": 0.37,
793
+ "learning_rate": 6.274570844080761e-07,
794
+ "loss": 0.1476,
795
+ "step": 8384
796
+ },
797
+ {
798
+ "epoch": 0.38,
799
+ "learning_rate": 6.246108689851463e-07,
800
+ "loss": 0.1454,
801
+ "step": 8448
802
+ },
803
+ {
804
+ "epoch": 0.38,
805
+ "learning_rate": 6.217646535622164e-07,
806
+ "loss": 0.1052,
807
+ "step": 8512
808
+ },
809
+ {
810
+ "epoch": 0.38,
811
+ "learning_rate": 6.1896291025527e-07,
812
+ "loss": 0.1477,
813
+ "step": 8576
814
+ },
815
+ {
816
+ "epoch": 0.38,
817
+ "learning_rate": 6.161166948323401e-07,
818
+ "loss": 0.1992,
819
+ "step": 8640
820
+ },
821
+ {
822
+ "epoch": 0.39,
823
+ "learning_rate": 6.132704794094103e-07,
824
+ "loss": 0.1662,
825
+ "step": 8704
826
+ },
827
+ {
828
+ "epoch": 0.39,
829
+ "learning_rate": 6.104242639864806e-07,
830
+ "loss": 0.1351,
831
+ "step": 8768
832
+ },
833
+ {
834
+ "epoch": 0.39,
835
+ "learning_rate": 6.075780485635506e-07,
836
+ "loss": 0.1551,
837
+ "step": 8832
838
+ },
839
+ {
840
+ "epoch": 0.4,
841
+ "learning_rate": 6.047318331406208e-07,
842
+ "loss": 0.183,
843
+ "step": 8896
844
+ },
845
+ {
846
+ "epoch": 0.4,
847
+ "learning_rate": 6.018856177176909e-07,
848
+ "loss": 0.1767,
849
+ "step": 8960
850
+ },
851
+ {
852
+ "epoch": 0.4,
853
+ "learning_rate": 5.990394022947611e-07,
854
+ "loss": 0.1564,
855
+ "step": 9024
856
+ },
857
+ {
858
+ "epoch": 0.4,
859
+ "learning_rate": 5.961931868718313e-07,
860
+ "loss": 0.1972,
861
+ "step": 9088
862
+ },
863
+ {
864
+ "epoch": 0.41,
865
+ "learning_rate": 5.933469714489015e-07,
866
+ "loss": 0.1867,
867
+ "step": 9152
868
+ },
869
+ {
870
+ "epoch": 0.41,
871
+ "learning_rate": 5.905007560259717e-07,
872
+ "loss": 0.2061,
873
+ "step": 9216
874
+ },
875
+ {
876
+ "epoch": 0.41,
877
+ "learning_rate": 5.876545406030419e-07,
878
+ "loss": 0.1744,
879
+ "step": 9280
880
+ },
881
+ {
882
+ "epoch": 0.42,
883
+ "learning_rate": 5.84808325180112e-07,
884
+ "loss": 0.1119,
885
+ "step": 9344
886
+ },
887
+ {
888
+ "epoch": 0.42,
889
+ "learning_rate": 5.819621097571822e-07,
890
+ "loss": 0.1835,
891
+ "step": 9408
892
+ },
893
+ {
894
+ "epoch": 0.42,
895
+ "learning_rate": 5.791158943342525e-07,
896
+ "loss": 0.1975,
897
+ "step": 9472
898
+ },
899
+ {
900
+ "epoch": 0.42,
901
+ "learning_rate": 5.762696789113226e-07,
902
+ "loss": 0.1862,
903
+ "step": 9536
904
+ },
905
+ {
906
+ "epoch": 0.43,
907
+ "learning_rate": 5.734234634883928e-07,
908
+ "loss": 0.1508,
909
+ "step": 9600
910
+ },
911
+ {
912
+ "epoch": 0.43,
913
+ "learning_rate": 5.705772480654629e-07,
914
+ "loss": 0.144,
915
+ "step": 9664
916
+ },
917
+ {
918
+ "epoch": 0.43,
919
+ "learning_rate": 5.67731032642533e-07,
920
+ "loss": 0.1776,
921
+ "step": 9728
922
+ },
923
+ {
924
+ "epoch": 0.44,
925
+ "learning_rate": 5.648848172196032e-07,
926
+ "loss": 0.1483,
927
+ "step": 9792
928
+ },
929
+ {
930
+ "epoch": 0.44,
931
+ "learning_rate": 5.620386017966735e-07,
932
+ "loss": 0.1155,
933
+ "step": 9856
934
+ },
935
+ {
936
+ "epoch": 0.44,
937
+ "learning_rate": 5.591923863737436e-07,
938
+ "loss": 0.1515,
939
+ "step": 9920
940
+ },
941
+ {
942
+ "epoch": 0.44,
943
+ "learning_rate": 5.563461709508138e-07,
944
+ "loss": 0.2434,
945
+ "step": 9984
946
+ },
947
+ {
948
+ "epoch": 0.45,
949
+ "learning_rate": 5.53499955527884e-07,
950
+ "loss": 0.1918,
951
+ "step": 10048
952
+ },
953
+ {
954
+ "epoch": 0.45,
955
+ "learning_rate": 5.506537401049541e-07,
956
+ "loss": 0.1944,
957
+ "step": 10112
958
+ },
959
+ {
960
+ "epoch": 0.45,
961
+ "learning_rate": 5.478075246820243e-07,
962
+ "loss": 0.1781,
963
+ "step": 10176
964
+ },
965
+ {
966
+ "epoch": 0.46,
967
+ "learning_rate": 5.449613092590946e-07,
968
+ "loss": 0.2186,
969
+ "step": 10240
970
+ },
971
+ {
972
+ "epoch": 0.46,
973
+ "learning_rate": 5.421150938361647e-07,
974
+ "loss": 0.1591,
975
+ "step": 10304
976
+ },
977
+ {
978
+ "epoch": 0.46,
979
+ "learning_rate": 5.392688784132349e-07,
980
+ "loss": 0.1282,
981
+ "step": 10368
982
+ },
983
+ {
984
+ "epoch": 0.46,
985
+ "learning_rate": 5.364226629903051e-07,
986
+ "loss": 0.23,
987
+ "step": 10432
988
+ },
989
+ {
990
+ "epoch": 0.47,
991
+ "learning_rate": 5.335764475673751e-07,
992
+ "loss": 0.1707,
993
+ "step": 10496
994
+ },
995
+ {
996
+ "epoch": 0.47,
997
+ "learning_rate": 5.307302321444454e-07,
998
+ "loss": 0.179,
999
+ "step": 10560
1000
+ },
1001
+ {
1002
+ "epoch": 0.47,
1003
+ "learning_rate": 5.278840167215156e-07,
1004
+ "loss": 0.1568,
1005
+ "step": 10624
1006
+ },
1007
+ {
1008
+ "epoch": 0.48,
1009
+ "learning_rate": 5.250378012985857e-07,
1010
+ "loss": 0.0832,
1011
+ "step": 10688
1012
+ },
1013
+ {
1014
+ "epoch": 0.48,
1015
+ "learning_rate": 5.221915858756559e-07,
1016
+ "loss": 0.1286,
1017
+ "step": 10752
1018
+ },
1019
+ {
1020
+ "epoch": 0.48,
1021
+ "learning_rate": 5.193898425687094e-07,
1022
+ "loss": 0.1397,
1023
+ "step": 10816
1024
+ },
1025
+ {
1026
+ "epoch": 0.48,
1027
+ "learning_rate": 5.165436271457796e-07,
1028
+ "loss": 0.1383,
1029
+ "step": 10880
1030
+ },
1031
+ {
1032
+ "epoch": 0.49,
1033
+ "learning_rate": 5.136974117228497e-07,
1034
+ "loss": 0.2332,
1035
+ "step": 10944
1036
+ },
1037
+ {
1038
+ "epoch": 0.49,
1039
+ "learning_rate": 5.108511962999199e-07,
1040
+ "loss": 0.1151,
1041
+ "step": 11008
1042
+ },
1043
+ {
1044
+ "epoch": 0.49,
1045
+ "learning_rate": 5.080049808769901e-07,
1046
+ "loss": 0.1769,
1047
+ "step": 11072
1048
+ },
1049
+ {
1050
+ "epoch": 0.5,
1051
+ "learning_rate": 5.051587654540603e-07,
1052
+ "loss": 0.1346,
1053
+ "step": 11136
1054
+ },
1055
+ {
1056
+ "epoch": 0.5,
1057
+ "learning_rate": 5.023125500311305e-07,
1058
+ "loss": 0.2568,
1059
+ "step": 11200
1060
+ },
1061
+ {
1062
+ "epoch": 0.5,
1063
+ "learning_rate": 4.994663346082007e-07,
1064
+ "loss": 0.1216,
1065
+ "step": 11264
1066
+ },
1067
+ {
1068
+ "epoch": 0.5,
1069
+ "learning_rate": 4.966201191852708e-07,
1070
+ "loss": 0.1167,
1071
+ "step": 11328
1072
+ },
1073
+ {
1074
+ "epoch": 0.51,
1075
+ "learning_rate": 4.93773903762341e-07,
1076
+ "loss": 0.1752,
1077
+ "step": 11392
1078
+ },
1079
+ {
1080
+ "epoch": 0.51,
1081
+ "learning_rate": 4.909276883394112e-07,
1082
+ "loss": 0.3198,
1083
+ "step": 11456
1084
+ },
1085
+ {
1086
+ "epoch": 0.51,
1087
+ "learning_rate": 4.880814729164813e-07,
1088
+ "loss": 0.2023,
1089
+ "step": 11520
1090
+ },
1091
+ {
1092
+ "epoch": 0.52,
1093
+ "learning_rate": 4.852352574935515e-07,
1094
+ "loss": 0.1261,
1095
+ "step": 11584
1096
+ },
1097
+ {
1098
+ "epoch": 0.52,
1099
+ "learning_rate": 4.823890420706217e-07,
1100
+ "loss": 0.0972,
1101
+ "step": 11648
1102
+ },
1103
+ {
1104
+ "epoch": 0.52,
1105
+ "learning_rate": 4.795428266476918e-07,
1106
+ "loss": 0.1692,
1107
+ "step": 11712
1108
+ },
1109
+ {
1110
+ "epoch": 0.52,
1111
+ "learning_rate": 4.7669661122476206e-07,
1112
+ "loss": 0.2509,
1113
+ "step": 11776
1114
+ },
1115
+ {
1116
+ "epoch": 0.53,
1117
+ "learning_rate": 4.7385039580183223e-07,
1118
+ "loss": 0.165,
1119
+ "step": 11840
1120
+ },
1121
+ {
1122
+ "epoch": 0.53,
1123
+ "learning_rate": 4.710041803789024e-07,
1124
+ "loss": 0.102,
1125
+ "step": 11904
1126
+ },
1127
+ {
1128
+ "epoch": 0.53,
1129
+ "learning_rate": 4.681579649559726e-07,
1130
+ "loss": 0.1197,
1131
+ "step": 11968
1132
+ },
1133
+ {
1134
+ "epoch": 0.54,
1135
+ "learning_rate": 4.653117495330428e-07,
1136
+ "loss": 0.2289,
1137
+ "step": 12032
1138
+ },
1139
+ {
1140
+ "epoch": 0.54,
1141
+ "learning_rate": 4.6246553411011294e-07,
1142
+ "loss": 0.1477,
1143
+ "step": 12096
1144
+ },
1145
+ {
1146
+ "epoch": 0.54,
1147
+ "learning_rate": 4.596193186871831e-07,
1148
+ "loss": 0.1813,
1149
+ "step": 12160
1150
+ },
1151
+ {
1152
+ "epoch": 0.54,
1153
+ "learning_rate": 4.567731032642533e-07,
1154
+ "loss": 0.2105,
1155
+ "step": 12224
1156
+ },
1157
+ {
1158
+ "epoch": 0.55,
1159
+ "learning_rate": 4.539268878413235e-07,
1160
+ "loss": 0.2905,
1161
+ "step": 12288
1162
+ },
1163
+ {
1164
+ "epoch": 0.55,
1165
+ "learning_rate": 4.5112514453437694e-07,
1166
+ "loss": 0.1313,
1167
+ "step": 12352
1168
+ },
1169
+ {
1170
+ "epoch": 0.55,
1171
+ "learning_rate": 4.482789291114471e-07,
1172
+ "loss": 0.188,
1173
+ "step": 12416
1174
+ },
1175
+ {
1176
+ "epoch": 0.56,
1177
+ "learning_rate": 4.454327136885173e-07,
1178
+ "loss": 0.1242,
1179
+ "step": 12480
1180
+ },
1181
+ {
1182
+ "epoch": 0.56,
1183
+ "learning_rate": 4.425864982655875e-07,
1184
+ "loss": 0.1764,
1185
+ "step": 12544
1186
+ },
1187
+ {
1188
+ "epoch": 0.56,
1189
+ "learning_rate": 4.397402828426576e-07,
1190
+ "loss": 0.2115,
1191
+ "step": 12608
1192
+ },
1193
+ {
1194
+ "epoch": 0.56,
1195
+ "learning_rate": 4.368940674197278e-07,
1196
+ "loss": 0.1478,
1197
+ "step": 12672
1198
+ },
1199
+ {
1200
+ "epoch": 0.57,
1201
+ "learning_rate": 4.34047851996798e-07,
1202
+ "loss": 0.2097,
1203
+ "step": 12736
1204
+ },
1205
+ {
1206
+ "epoch": 0.57,
1207
+ "learning_rate": 4.3120163657386816e-07,
1208
+ "loss": 0.2395,
1209
+ "step": 12800
1210
+ },
1211
+ {
1212
+ "epoch": 0.57,
1213
+ "learning_rate": 4.283554211509384e-07,
1214
+ "loss": 0.1679,
1215
+ "step": 12864
1216
+ },
1217
+ {
1218
+ "epoch": 0.57,
1219
+ "learning_rate": 4.2550920572800854e-07,
1220
+ "loss": 0.1273,
1221
+ "step": 12928
1222
+ },
1223
+ {
1224
+ "epoch": 0.58,
1225
+ "learning_rate": 4.2266299030507866e-07,
1226
+ "loss": 0.1552,
1227
+ "step": 12992
1228
+ },
1229
+ {
1230
+ "epoch": 0.58,
1231
+ "learning_rate": 4.198167748821489e-07,
1232
+ "loss": 0.1337,
1233
+ "step": 13056
1234
+ },
1235
+ {
1236
+ "epoch": 0.58,
1237
+ "learning_rate": 4.1697055945921904e-07,
1238
+ "loss": 0.1604,
1239
+ "step": 13120
1240
+ },
1241
+ {
1242
+ "epoch": 0.59,
1243
+ "learning_rate": 4.141243440362892e-07,
1244
+ "loss": 0.2796,
1245
+ "step": 13184
1246
+ },
1247
+ {
1248
+ "epoch": 0.59,
1249
+ "learning_rate": 4.1127812861335943e-07,
1250
+ "loss": 0.1497,
1251
+ "step": 13248
1252
+ },
1253
+ {
1254
+ "epoch": 0.59,
1255
+ "learning_rate": 4.084319131904296e-07,
1256
+ "loss": 0.1414,
1257
+ "step": 13312
1258
+ },
1259
+ {
1260
+ "epoch": 0.59,
1261
+ "learning_rate": 4.0558569776749976e-07,
1262
+ "loss": 0.1435,
1263
+ "step": 13376
1264
+ },
1265
+ {
1266
+ "epoch": 0.6,
1267
+ "learning_rate": 4.027394823445699e-07,
1268
+ "loss": 0.1435,
1269
+ "step": 13440
1270
+ },
1271
+ {
1272
+ "epoch": 0.6,
1273
+ "learning_rate": 3.998932669216401e-07,
1274
+ "loss": 0.222,
1275
+ "step": 13504
1276
+ },
1277
+ {
1278
+ "epoch": 0.6,
1279
+ "learning_rate": 3.970470514987103e-07,
1280
+ "loss": 0.1387,
1281
+ "step": 13568
1282
+ },
1283
+ {
1284
+ "epoch": 0.61,
1285
+ "learning_rate": 3.942008360757805e-07,
1286
+ "loss": 0.144,
1287
+ "step": 13632
1288
+ },
1289
+ {
1290
+ "epoch": 0.61,
1291
+ "learning_rate": 3.9135462065285064e-07,
1292
+ "loss": 0.1839,
1293
+ "step": 13696
1294
+ },
1295
+ {
1296
+ "epoch": 0.61,
1297
+ "learning_rate": 3.8850840522992086e-07,
1298
+ "loss": 0.3017,
1299
+ "step": 13760
1300
+ },
1301
+ {
1302
+ "epoch": 0.61,
1303
+ "learning_rate": 3.85662189806991e-07,
1304
+ "loss": 0.1267,
1305
+ "step": 13824
1306
+ },
1307
+ {
1308
+ "epoch": 0.62,
1309
+ "learning_rate": 3.8281597438406114e-07,
1310
+ "loss": 0.1733,
1311
+ "step": 13888
1312
+ },
1313
+ {
1314
+ "epoch": 0.62,
1315
+ "learning_rate": 3.7996975896113136e-07,
1316
+ "loss": 0.2575,
1317
+ "step": 13952
1318
+ },
1319
+ {
1320
+ "epoch": 0.62,
1321
+ "learning_rate": 3.771235435382015e-07,
1322
+ "loss": 0.1702,
1323
+ "step": 14016
1324
+ },
1325
+ {
1326
+ "epoch": 0.63,
1327
+ "learning_rate": 3.742773281152717e-07,
1328
+ "loss": 0.1609,
1329
+ "step": 14080
1330
+ },
1331
+ {
1332
+ "epoch": 0.63,
1333
+ "learning_rate": 3.714311126923419e-07,
1334
+ "loss": 0.1917,
1335
+ "step": 14144
1336
+ },
1337
+ {
1338
+ "epoch": 0.63,
1339
+ "learning_rate": 3.685848972694121e-07,
1340
+ "loss": 0.1087,
1341
+ "step": 14208
1342
+ },
1343
+ {
1344
+ "epoch": 0.63,
1345
+ "learning_rate": 3.6573868184648224e-07,
1346
+ "loss": 0.1452,
1347
+ "step": 14272
1348
+ },
1349
+ {
1350
+ "epoch": 0.64,
1351
+ "learning_rate": 3.628924664235524e-07,
1352
+ "loss": 0.1491,
1353
+ "step": 14336
1354
+ },
1355
+ {
1356
+ "epoch": 0.64,
1357
+ "learning_rate": 3.6004625100062257e-07,
1358
+ "loss": 0.1735,
1359
+ "step": 14400
1360
+ },
1361
+ {
1362
+ "epoch": 0.64,
1363
+ "learning_rate": 3.572000355776928e-07,
1364
+ "loss": 0.2468,
1365
+ "step": 14464
1366
+ },
1367
+ {
1368
+ "epoch": 0.65,
1369
+ "learning_rate": 3.5435382015476296e-07,
1370
+ "loss": 0.1958,
1371
+ "step": 14528
1372
+ },
1373
+ {
1374
+ "epoch": 0.65,
1375
+ "learning_rate": 3.515076047318331e-07,
1376
+ "loss": 0.1487,
1377
+ "step": 14592
1378
+ },
1379
+ {
1380
+ "epoch": 0.65,
1381
+ "learning_rate": 3.4866138930890334e-07,
1382
+ "loss": 0.1503,
1383
+ "step": 14656
1384
+ },
1385
+ {
1386
+ "epoch": 0.65,
1387
+ "learning_rate": 3.4581517388597346e-07,
1388
+ "loss": 0.184,
1389
+ "step": 14720
1390
+ },
1391
+ {
1392
+ "epoch": 0.66,
1393
+ "learning_rate": 3.429689584630436e-07,
1394
+ "loss": 0.2188,
1395
+ "step": 14784
1396
+ },
1397
+ {
1398
+ "epoch": 0.66,
1399
+ "learning_rate": 3.4012274304011384e-07,
1400
+ "loss": 0.1971,
1401
+ "step": 14848
1402
+ },
1403
+ {
1404
+ "epoch": 0.66,
1405
+ "learning_rate": 3.37276527617184e-07,
1406
+ "loss": 0.1236,
1407
+ "step": 14912
1408
+ },
1409
+ {
1410
+ "epoch": 0.67,
1411
+ "learning_rate": 3.3443031219425417e-07,
1412
+ "loss": 0.2584,
1413
+ "step": 14976
1414
+ },
1415
+ {
1416
+ "epoch": 0.67,
1417
+ "learning_rate": 3.315840967713244e-07,
1418
+ "loss": 0.174,
1419
+ "step": 15040
1420
+ },
1421
+ {
1422
+ "epoch": 0.67,
1423
+ "learning_rate": 3.287378813483945e-07,
1424
+ "loss": 0.1952,
1425
+ "step": 15104
1426
+ },
1427
+ {
1428
+ "epoch": 0.67,
1429
+ "learning_rate": 3.258916659254647e-07,
1430
+ "loss": 0.1841,
1431
+ "step": 15168
1432
+ },
1433
+ {
1434
+ "epoch": 0.68,
1435
+ "learning_rate": 3.230454505025349e-07,
1436
+ "loss": 0.2457,
1437
+ "step": 15232
1438
+ },
1439
+ {
1440
+ "epoch": 0.68,
1441
+ "learning_rate": 3.2019923507960506e-07,
1442
+ "loss": 0.1458,
1443
+ "step": 15296
1444
+ },
1445
+ {
1446
+ "epoch": 0.68,
1447
+ "learning_rate": 3.173530196566753e-07,
1448
+ "loss": 0.1952,
1449
+ "step": 15360
1450
+ },
1451
+ {
1452
+ "epoch": 0.69,
1453
+ "learning_rate": 3.1450680423374544e-07,
1454
+ "loss": 0.1276,
1455
+ "step": 15424
1456
+ },
1457
+ {
1458
+ "epoch": 0.69,
1459
+ "learning_rate": 3.116605888108156e-07,
1460
+ "loss": 0.1784,
1461
+ "step": 15488
1462
+ },
1463
+ {
1464
+ "epoch": 0.69,
1465
+ "learning_rate": 3.0881437338788577e-07,
1466
+ "loss": 0.2179,
1467
+ "step": 15552
1468
+ },
1469
+ {
1470
+ "epoch": 0.69,
1471
+ "learning_rate": 3.0596815796495594e-07,
1472
+ "loss": 0.1603,
1473
+ "step": 15616
1474
+ },
1475
+ {
1476
+ "epoch": 0.7,
1477
+ "learning_rate": 3.031219425420261e-07,
1478
+ "loss": 0.1517,
1479
+ "step": 15680
1480
+ },
1481
+ {
1482
+ "epoch": 0.7,
1483
+ "learning_rate": 3.003201992350796e-07,
1484
+ "loss": 0.1842,
1485
+ "step": 15744
1486
+ },
1487
+ {
1488
+ "epoch": 0.7,
1489
+ "learning_rate": 2.9747398381214977e-07,
1490
+ "loss": 0.2172,
1491
+ "step": 15808
1492
+ },
1493
+ {
1494
+ "epoch": 0.71,
1495
+ "learning_rate": 2.9462776838921994e-07,
1496
+ "loss": 0.1086,
1497
+ "step": 15872
1498
+ },
1499
+ {
1500
+ "epoch": 0.71,
1501
+ "learning_rate": 2.9178155296629016e-07,
1502
+ "loss": 0.1628,
1503
+ "step": 15936
1504
+ },
1505
+ {
1506
+ "epoch": 0.71,
1507
+ "learning_rate": 2.889353375433603e-07,
1508
+ "loss": 0.1529,
1509
+ "step": 16000
1510
+ },
1511
+ {
1512
+ "epoch": 0.71,
1513
+ "learning_rate": 2.8608912212043044e-07,
1514
+ "loss": 0.2353,
1515
+ "step": 16064
1516
+ },
1517
+ {
1518
+ "epoch": 0.72,
1519
+ "learning_rate": 2.8324290669750066e-07,
1520
+ "loss": 0.2485,
1521
+ "step": 16128
1522
+ },
1523
+ {
1524
+ "epoch": 0.72,
1525
+ "learning_rate": 2.803966912745708e-07,
1526
+ "loss": 0.118,
1527
+ "step": 16192
1528
+ },
1529
+ {
1530
+ "epoch": 0.72,
1531
+ "learning_rate": 2.77550475851641e-07,
1532
+ "loss": 0.1164,
1533
+ "step": 16256
1534
+ },
1535
+ {
1536
+ "epoch": 0.73,
1537
+ "learning_rate": 2.747042604287112e-07,
1538
+ "loss": 0.1365,
1539
+ "step": 16320
1540
+ },
1541
+ {
1542
+ "epoch": 0.73,
1543
+ "learning_rate": 2.7185804500578137e-07,
1544
+ "loss": 0.1181,
1545
+ "step": 16384
1546
+ },
1547
+ {
1548
+ "epoch": 0.73,
1549
+ "learning_rate": 2.6901182958285154e-07,
1550
+ "loss": 0.1266,
1551
+ "step": 16448
1552
+ },
1553
+ {
1554
+ "epoch": 0.73,
1555
+ "learning_rate": 2.661656141599217e-07,
1556
+ "loss": 0.1228,
1557
+ "step": 16512
1558
+ },
1559
+ {
1560
+ "epoch": 0.74,
1561
+ "learning_rate": 2.6331939873699187e-07,
1562
+ "loss": 0.2486,
1563
+ "step": 16576
1564
+ },
1565
+ {
1566
+ "epoch": 0.74,
1567
+ "learning_rate": 2.604731833140621e-07,
1568
+ "loss": 0.1087,
1569
+ "step": 16640
1570
+ },
1571
+ {
1572
+ "epoch": 0.74,
1573
+ "learning_rate": 2.5762696789113226e-07,
1574
+ "loss": 0.2053,
1575
+ "step": 16704
1576
+ },
1577
+ {
1578
+ "epoch": 0.75,
1579
+ "learning_rate": 2.547807524682024e-07,
1580
+ "loss": 0.1459,
1581
+ "step": 16768
1582
+ },
1583
+ {
1584
+ "epoch": 0.75,
1585
+ "learning_rate": 2.5193453704527264e-07,
1586
+ "loss": 0.1386,
1587
+ "step": 16832
1588
+ },
1589
+ {
1590
+ "epoch": 0.75,
1591
+ "learning_rate": 2.4908832162234275e-07,
1592
+ "loss": 0.1425,
1593
+ "step": 16896
1594
+ },
1595
+ {
1596
+ "epoch": 0.75,
1597
+ "learning_rate": 2.4624210619941297e-07,
1598
+ "loss": 0.2195,
1599
+ "step": 16960
1600
+ },
1601
+ {
1602
+ "epoch": 0.76,
1603
+ "learning_rate": 2.4339589077648314e-07,
1604
+ "loss": 0.199,
1605
+ "step": 17024
1606
+ },
1607
+ {
1608
+ "epoch": 0.76,
1609
+ "learning_rate": 2.405496753535533e-07,
1610
+ "loss": 0.1927,
1611
+ "step": 17088
1612
+ },
1613
+ {
1614
+ "epoch": 0.76,
1615
+ "learning_rate": 2.377034599306235e-07,
1616
+ "loss": 0.1302,
1617
+ "step": 17152
1618
+ },
1619
+ {
1620
+ "epoch": 0.77,
1621
+ "learning_rate": 2.3485724450769366e-07,
1622
+ "loss": 0.1006,
1623
+ "step": 17216
1624
+ },
1625
+ {
1626
+ "epoch": 0.77,
1627
+ "learning_rate": 2.3201102908476383e-07,
1628
+ "loss": 0.1549,
1629
+ "step": 17280
1630
+ },
1631
+ {
1632
+ "epoch": 0.77,
1633
+ "learning_rate": 2.2916481366183402e-07,
1634
+ "loss": 0.1823,
1635
+ "step": 17344
1636
+ },
1637
+ {
1638
+ "epoch": 0.77,
1639
+ "learning_rate": 2.2631859823890421e-07,
1640
+ "loss": 0.1862,
1641
+ "step": 17408
1642
+ },
1643
+ {
1644
+ "epoch": 0.78,
1645
+ "learning_rate": 2.2347238281597438e-07,
1646
+ "loss": 0.1273,
1647
+ "step": 17472
1648
+ },
1649
+ {
1650
+ "epoch": 0.78,
1651
+ "learning_rate": 2.2062616739304455e-07,
1652
+ "loss": 0.1278,
1653
+ "step": 17536
1654
+ },
1655
+ {
1656
+ "epoch": 0.78,
1657
+ "learning_rate": 2.1777995197011474e-07,
1658
+ "loss": 0.1448,
1659
+ "step": 17600
1660
+ },
1661
+ {
1662
+ "epoch": 0.79,
1663
+ "learning_rate": 2.149337365471849e-07,
1664
+ "loss": 0.1611,
1665
+ "step": 17664
1666
+ },
1667
+ {
1668
+ "epoch": 0.79,
1669
+ "learning_rate": 2.1208752112425507e-07,
1670
+ "loss": 0.1766,
1671
+ "step": 17728
1672
+ },
1673
+ {
1674
+ "epoch": 0.79,
1675
+ "learning_rate": 2.0924130570132526e-07,
1676
+ "loss": 0.1391,
1677
+ "step": 17792
1678
+ },
1679
+ {
1680
+ "epoch": 0.79,
1681
+ "learning_rate": 2.0639509027839543e-07,
1682
+ "loss": 0.1674,
1683
+ "step": 17856
1684
+ },
1685
+ {
1686
+ "epoch": 0.8,
1687
+ "learning_rate": 2.0354887485546562e-07,
1688
+ "loss": 0.1679,
1689
+ "step": 17920
1690
+ },
1691
+ {
1692
+ "epoch": 0.8,
1693
+ "learning_rate": 2.007026594325358e-07,
1694
+ "loss": 0.1686,
1695
+ "step": 17984
1696
+ },
1697
+ {
1698
+ "epoch": 0.8,
1699
+ "learning_rate": 1.9785644400960595e-07,
1700
+ "loss": 0.1433,
1701
+ "step": 18048
1702
+ },
1703
+ {
1704
+ "epoch": 0.81,
1705
+ "learning_rate": 1.9501022858667615e-07,
1706
+ "loss": 0.1909,
1707
+ "step": 18112
1708
+ },
1709
+ {
1710
+ "epoch": 0.81,
1711
+ "learning_rate": 1.9216401316374634e-07,
1712
+ "loss": 0.1912,
1713
+ "step": 18176
1714
+ },
1715
+ {
1716
+ "epoch": 0.81,
1717
+ "learning_rate": 1.893177977408165e-07,
1718
+ "loss": 0.1547,
1719
+ "step": 18240
1720
+ },
1721
+ {
1722
+ "epoch": 0.81,
1723
+ "learning_rate": 1.8647158231788667e-07,
1724
+ "loss": 0.1751,
1725
+ "step": 18304
1726
+ },
1727
+ {
1728
+ "epoch": 0.82,
1729
+ "learning_rate": 1.8362536689495686e-07,
1730
+ "loss": 0.199,
1731
+ "step": 18368
1732
+ },
1733
+ {
1734
+ "epoch": 0.82,
1735
+ "learning_rate": 1.8077915147202703e-07,
1736
+ "loss": 0.0896,
1737
+ "step": 18432
1738
+ },
1739
+ {
1740
+ "epoch": 0.82,
1741
+ "learning_rate": 1.779329360490972e-07,
1742
+ "loss": 0.1783,
1743
+ "step": 18496
1744
+ },
1745
+ {
1746
+ "epoch": 0.83,
1747
+ "learning_rate": 1.750867206261674e-07,
1748
+ "loss": 0.1862,
1749
+ "step": 18560
1750
+ },
1751
+ {
1752
+ "epoch": 0.83,
1753
+ "learning_rate": 1.7224050520323758e-07,
1754
+ "loss": 0.1291,
1755
+ "step": 18624
1756
+ },
1757
+ {
1758
+ "epoch": 0.83,
1759
+ "learning_rate": 1.6939428978030772e-07,
1760
+ "loss": 0.1866,
1761
+ "step": 18688
1762
+ },
1763
+ {
1764
+ "epoch": 0.83,
1765
+ "learning_rate": 1.665480743573779e-07,
1766
+ "loss": 0.1292,
1767
+ "step": 18752
1768
+ },
1769
+ {
1770
+ "epoch": 0.84,
1771
+ "learning_rate": 1.637018589344481e-07,
1772
+ "loss": 0.2217,
1773
+ "step": 18816
1774
+ },
1775
+ {
1776
+ "epoch": 0.84,
1777
+ "learning_rate": 1.6085564351151827e-07,
1778
+ "loss": 0.188,
1779
+ "step": 18880
1780
+ },
1781
+ {
1782
+ "epoch": 0.84,
1783
+ "learning_rate": 1.5805390020457175e-07,
1784
+ "loss": 0.1794,
1785
+ "step": 18944
1786
+ },
1787
+ {
1788
+ "epoch": 0.85,
1789
+ "learning_rate": 1.5520768478164189e-07,
1790
+ "loss": 0.1387,
1791
+ "step": 19008
1792
+ },
1793
+ {
1794
+ "epoch": 0.85,
1795
+ "learning_rate": 1.5236146935871208e-07,
1796
+ "loss": 0.1688,
1797
+ "step": 19072
1798
+ },
1799
+ {
1800
+ "epoch": 0.85,
1801
+ "learning_rate": 1.4951525393578227e-07,
1802
+ "loss": 0.1593,
1803
+ "step": 19136
1804
+ },
1805
+ {
1806
+ "epoch": 0.85,
1807
+ "learning_rate": 1.466690385128524e-07,
1808
+ "loss": 0.1521,
1809
+ "step": 19200
1810
+ },
1811
+ {
1812
+ "epoch": 0.86,
1813
+ "learning_rate": 1.438228230899226e-07,
1814
+ "loss": 0.279,
1815
+ "step": 19264
1816
+ },
1817
+ {
1818
+ "epoch": 0.86,
1819
+ "learning_rate": 1.409766076669928e-07,
1820
+ "loss": 0.1985,
1821
+ "step": 19328
1822
+ },
1823
+ {
1824
+ "epoch": 0.86,
1825
+ "learning_rate": 1.38130392244063e-07,
1826
+ "loss": 0.1708,
1827
+ "step": 19392
1828
+ },
1829
+ {
1830
+ "epoch": 0.87,
1831
+ "learning_rate": 1.3528417682113313e-07,
1832
+ "loss": 0.126,
1833
+ "step": 19456
1834
+ },
1835
+ {
1836
+ "epoch": 0.87,
1837
+ "learning_rate": 1.3243796139820332e-07,
1838
+ "loss": 0.0776,
1839
+ "step": 19520
1840
+ },
1841
+ {
1842
+ "epoch": 0.87,
1843
+ "learning_rate": 1.295917459752735e-07,
1844
+ "loss": 0.1608,
1845
+ "step": 19584
1846
+ },
1847
+ {
1848
+ "epoch": 0.87,
1849
+ "learning_rate": 1.2674553055234368e-07,
1850
+ "loss": 0.2249,
1851
+ "step": 19648
1852
+ },
1853
+ {
1854
+ "epoch": 0.88,
1855
+ "learning_rate": 1.2389931512941384e-07,
1856
+ "loss": 0.1545,
1857
+ "step": 19712
1858
+ },
1859
+ {
1860
+ "epoch": 0.88,
1861
+ "learning_rate": 1.2105309970648404e-07,
1862
+ "loss": 0.1853,
1863
+ "step": 19776
1864
+ },
1865
+ {
1866
+ "epoch": 0.88,
1867
+ "learning_rate": 1.182068842835542e-07,
1868
+ "loss": 0.1838,
1869
+ "step": 19840
1870
+ },
1871
+ {
1872
+ "epoch": 0.89,
1873
+ "learning_rate": 1.1536066886062438e-07,
1874
+ "loss": 0.1543,
1875
+ "step": 19904
1876
+ },
1877
+ {
1878
+ "epoch": 0.89,
1879
+ "learning_rate": 1.1251445343769456e-07,
1880
+ "loss": 0.1488,
1881
+ "step": 19968
1882
+ },
1883
+ {
1884
+ "epoch": 0.89,
1885
+ "learning_rate": 1.0966823801476474e-07,
1886
+ "loss": 0.1933,
1887
+ "step": 20032
1888
+ },
1889
+ {
1890
+ "epoch": 0.89,
1891
+ "learning_rate": 1.068220225918349e-07,
1892
+ "loss": 0.2114,
1893
+ "step": 20096
1894
+ },
1895
+ {
1896
+ "epoch": 0.9,
1897
+ "learning_rate": 1.039758071689051e-07,
1898
+ "loss": 0.2539,
1899
+ "step": 20160
1900
+ },
1901
+ {
1902
+ "epoch": 0.9,
1903
+ "learning_rate": 1.0112959174597526e-07,
1904
+ "loss": 0.2068,
1905
+ "step": 20224
1906
+ },
1907
+ {
1908
+ "epoch": 0.9,
1909
+ "learning_rate": 9.828337632304546e-08,
1910
+ "loss": 0.1288,
1911
+ "step": 20288
1912
+ },
1913
+ {
1914
+ "epoch": 0.91,
1915
+ "learning_rate": 9.543716090011562e-08,
1916
+ "loss": 0.1821,
1917
+ "step": 20352
1918
+ },
1919
+ {
1920
+ "epoch": 0.91,
1921
+ "learning_rate": 9.259094547718579e-08,
1922
+ "loss": 0.2054,
1923
+ "step": 20416
1924
+ },
1925
+ {
1926
+ "epoch": 0.91,
1927
+ "learning_rate": 8.974473005425598e-08,
1928
+ "loss": 0.1729,
1929
+ "step": 20480
1930
+ },
1931
+ {
1932
+ "epoch": 0.91,
1933
+ "learning_rate": 8.689851463132615e-08,
1934
+ "loss": 0.1045,
1935
+ "step": 20544
1936
+ },
1937
+ {
1938
+ "epoch": 0.92,
1939
+ "learning_rate": 8.405229920839634e-08,
1940
+ "loss": 0.2292,
1941
+ "step": 20608
1942
+ },
1943
+ {
1944
+ "epoch": 0.92,
1945
+ "learning_rate": 8.12060837854665e-08,
1946
+ "loss": 0.1159,
1947
+ "step": 20672
1948
+ },
1949
+ {
1950
+ "epoch": 0.92,
1951
+ "learning_rate": 7.835986836253668e-08,
1952
+ "loss": 0.1653,
1953
+ "step": 20736
1954
+ },
1955
+ {
1956
+ "epoch": 0.93,
1957
+ "learning_rate": 7.551365293960686e-08,
1958
+ "loss": 0.1501,
1959
+ "step": 20800
1960
+ },
1961
+ {
1962
+ "epoch": 0.93,
1963
+ "learning_rate": 7.266743751667704e-08,
1964
+ "loss": 0.1996,
1965
+ "step": 20864
1966
+ },
1967
+ {
1968
+ "epoch": 0.93,
1969
+ "learning_rate": 6.982122209374722e-08,
1970
+ "loss": 0.1906,
1971
+ "step": 20928
1972
+ },
1973
+ {
1974
+ "epoch": 0.93,
1975
+ "learning_rate": 6.697500667081739e-08,
1976
+ "loss": 0.1375,
1977
+ "step": 20992
1978
+ },
1979
+ {
1980
+ "epoch": 0.94,
1981
+ "learning_rate": 6.412879124788757e-08,
1982
+ "loss": 0.1735,
1983
+ "step": 21056
1984
+ },
1985
+ {
1986
+ "epoch": 0.94,
1987
+ "learning_rate": 6.128257582495775e-08,
1988
+ "loss": 0.1695,
1989
+ "step": 21120
1990
+ },
1991
+ {
1992
+ "epoch": 0.94,
1993
+ "learning_rate": 5.8436360402027926e-08,
1994
+ "loss": 0.1745,
1995
+ "step": 21184
1996
+ },
1997
+ {
1998
+ "epoch": 0.94,
1999
+ "learning_rate": 5.55901449790981e-08,
2000
+ "loss": 0.1534,
2001
+ "step": 21248
2002
+ },
2003
+ {
2004
+ "epoch": 0.95,
2005
+ "learning_rate": 5.274392955616828e-08,
2006
+ "loss": 0.1317,
2007
+ "step": 21312
2008
+ },
2009
+ {
2010
+ "epoch": 0.95,
2011
+ "learning_rate": 4.989771413323846e-08,
2012
+ "loss": 0.2353,
2013
+ "step": 21376
2014
+ },
2015
+ {
2016
+ "epoch": 0.95,
2017
+ "learning_rate": 4.7051498710308637e-08,
2018
+ "loss": 0.1399,
2019
+ "step": 21440
2020
+ },
2021
+ {
2022
+ "epoch": 0.96,
2023
+ "learning_rate": 4.420528328737881e-08,
2024
+ "loss": 0.1618,
2025
+ "step": 21504
2026
+ },
2027
+ {
2028
+ "epoch": 0.96,
2029
+ "learning_rate": 4.135906786444899e-08,
2030
+ "loss": 0.0871,
2031
+ "step": 21568
2032
+ },
2033
+ {
2034
+ "epoch": 0.96,
2035
+ "learning_rate": 3.851285244151917e-08,
2036
+ "loss": 0.1977,
2037
+ "step": 21632
2038
+ },
2039
+ {
2040
+ "epoch": 0.96,
2041
+ "learning_rate": 3.566663701858935e-08,
2042
+ "loss": 0.2695,
2043
+ "step": 21696
2044
+ },
2045
+ {
2046
+ "epoch": 0.97,
2047
+ "learning_rate": 3.282042159565952e-08,
2048
+ "loss": 0.1834,
2049
+ "step": 21760
2050
+ },
2051
+ {
2052
+ "epoch": 0.97,
2053
+ "learning_rate": 2.99742061727297e-08,
2054
+ "loss": 0.1823,
2055
+ "step": 21824
2056
+ },
2057
+ {
2058
+ "epoch": 0.97,
2059
+ "learning_rate": 2.7127990749799875e-08,
2060
+ "loss": 0.1985,
2061
+ "step": 21888
2062
+ },
2063
+ {
2064
+ "epoch": 0.98,
2065
+ "learning_rate": 2.4281775326870054e-08,
2066
+ "loss": 0.1322,
2067
+ "step": 21952
2068
+ },
2069
+ {
2070
+ "epoch": 0.98,
2071
+ "learning_rate": 2.1435559903940226e-08,
2072
+ "loss": 0.1736,
2073
+ "step": 22016
2074
+ },
2075
+ {
2076
+ "epoch": 0.98,
2077
+ "learning_rate": 1.8589344481010406e-08,
2078
+ "loss": 0.2192,
2079
+ "step": 22080
2080
+ },
2081
+ {
2082
+ "epoch": 0.98,
2083
+ "learning_rate": 1.574312905808058e-08,
2084
+ "loss": 0.2138,
2085
+ "step": 22144
2086
+ },
2087
+ {
2088
+ "epoch": 0.99,
2089
+ "learning_rate": 1.289691363515076e-08,
2090
+ "loss": 0.117,
2091
+ "step": 22208
2092
+ },
2093
+ {
2094
+ "epoch": 0.99,
2095
+ "learning_rate": 1.0050698212220937e-08,
2096
+ "loss": 0.1349,
2097
+ "step": 22272
2098
+ },
2099
+ {
2100
+ "epoch": 0.99,
2101
+ "learning_rate": 7.204482789291114e-09,
2102
+ "loss": 0.24,
2103
+ "step": 22336
2104
+ },
2105
+ {
2106
+ "epoch": 1.0,
2107
+ "learning_rate": 4.358267366361292e-09,
2108
+ "loss": 0.1303,
2109
+ "step": 22400
2110
+ },
2111
+ {
2112
+ "epoch": 1.0,
2113
+ "learning_rate": 1.5120519434314685e-09,
2114
+ "loss": 0.1459,
2115
+ "step": 22464
2116
+ }
2117
+ ],
2118
+ "logging_steps": 64,
2119
+ "max_steps": 22486,
2120
+ "num_input_tokens_seen": 0,
2121
+ "num_train_epochs": 1,
2122
+ "save_steps": 7495,
2123
+ "total_flos": 4.7739500402688e+19,
2124
+ "train_batch_size": 1,
2125
+ "trial_name": null,
2126
+ "trial_params": null
2127
+ }
checkpoint-22485/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5dd66348aef7664171d12847300a1c626bfa2b1ede8e5141c5a2fc70bac4b83
3
+ size 5048
checkpoint-7495/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-7495/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-7495/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6fe3a60d4bb74b9285e1e9b91b78aa44ea486900d7cf668f2c194ccbef2470b
3
+ size 4992706480
checkpoint-7495/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5db06f6498b03583951f9c1f5ecaf74edcf74018880905a6c4bb30dee37ce6eb
3
+ size 1180663192
checkpoint-7495/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7495/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:efcc7935bed85f721eaee8db9dc3f016d9193d2643cf065855cee3d02bdbaafb
3
+ size 3095074288
checkpoint-7495/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-7495/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e52f6a4a50e09e4a6a5814e2e491a8928aa1fd4170083050a50d3ec7c4bd039
3
+ size 14244
checkpoint-7495/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd3b0b7b331acf67dc2b463ab51d489e25ca5c6e23b8d0af52cd0b414b82b514
3
+ size 1064
checkpoint-7495/trainer_state.json ADDED
@@ -0,0 +1,723 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.3333185092946722,
5
+ "eval_steps": 500,
6
+ "global_step": 7495,
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": 9.973761451569865e-07,
14
+ "loss": 3.933,
15
+ "step": 64
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 9.945299297340568e-07,
20
+ "loss": 1.6429,
21
+ "step": 128
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 9.917726585430935e-07,
26
+ "loss": 1.1012,
27
+ "step": 192
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 9.889264431201636e-07,
32
+ "loss": 0.3376,
33
+ "step": 256
34
+ },
35
+ {
36
+ "epoch": 0.01,
37
+ "learning_rate": 9.860802276972338e-07,
38
+ "loss": 0.136,
39
+ "step": 320
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 9.832340122743039e-07,
44
+ "loss": 0.2753,
45
+ "step": 384
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 9.803877968513742e-07,
50
+ "loss": 0.25,
51
+ "step": 448
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 9.775415814284444e-07,
56
+ "loss": 0.3117,
57
+ "step": 512
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 9.746953660055145e-07,
62
+ "loss": 0.2853,
63
+ "step": 576
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 9.718491505825848e-07,
68
+ "loss": 0.2489,
69
+ "step": 640
70
+ },
71
+ {
72
+ "epoch": 0.03,
73
+ "learning_rate": 9.690029351596548e-07,
74
+ "loss": 0.217,
75
+ "step": 704
76
+ },
77
+ {
78
+ "epoch": 0.03,
79
+ "learning_rate": 9.66156719736725e-07,
80
+ "loss": 0.1804,
81
+ "step": 768
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 9.633105043137952e-07,
86
+ "loss": 0.2842,
87
+ "step": 832
88
+ },
89
+ {
90
+ "epoch": 0.04,
91
+ "learning_rate": 9.604642888908654e-07,
92
+ "loss": 0.2621,
93
+ "step": 896
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "learning_rate": 9.576180734679355e-07,
98
+ "loss": 0.3316,
99
+ "step": 960
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 9.547718580450058e-07,
104
+ "loss": 0.1856,
105
+ "step": 1024
106
+ },
107
+ {
108
+ "epoch": 0.05,
109
+ "learning_rate": 9.519256426220758e-07,
110
+ "loss": 0.2112,
111
+ "step": 1088
112
+ },
113
+ {
114
+ "epoch": 0.05,
115
+ "learning_rate": 9.490794271991461e-07,
116
+ "loss": 0.1778,
117
+ "step": 1152
118
+ },
119
+ {
120
+ "epoch": 0.05,
121
+ "learning_rate": 9.462332117762163e-07,
122
+ "loss": 0.319,
123
+ "step": 1216
124
+ },
125
+ {
126
+ "epoch": 0.06,
127
+ "learning_rate": 9.433869963532864e-07,
128
+ "loss": 0.1678,
129
+ "step": 1280
130
+ },
131
+ {
132
+ "epoch": 0.06,
133
+ "learning_rate": 9.405407809303566e-07,
134
+ "loss": 0.2549,
135
+ "step": 1344
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "learning_rate": 9.376945655074268e-07,
140
+ "loss": 0.1723,
141
+ "step": 1408
142
+ },
143
+ {
144
+ "epoch": 0.07,
145
+ "learning_rate": 9.34848350084497e-07,
146
+ "loss": 0.149,
147
+ "step": 1472
148
+ },
149
+ {
150
+ "epoch": 0.07,
151
+ "learning_rate": 9.320021346615672e-07,
152
+ "loss": 0.2097,
153
+ "step": 1536
154
+ },
155
+ {
156
+ "epoch": 0.07,
157
+ "learning_rate": 9.291559192386374e-07,
158
+ "loss": 0.3716,
159
+ "step": 1600
160
+ },
161
+ {
162
+ "epoch": 0.07,
163
+ "learning_rate": 9.263097038157075e-07,
164
+ "loss": 0.1303,
165
+ "step": 1664
166
+ },
167
+ {
168
+ "epoch": 0.08,
169
+ "learning_rate": 9.234634883927777e-07,
170
+ "loss": 0.2041,
171
+ "step": 1728
172
+ },
173
+ {
174
+ "epoch": 0.08,
175
+ "learning_rate": 9.206172729698479e-07,
176
+ "loss": 0.2529,
177
+ "step": 1792
178
+ },
179
+ {
180
+ "epoch": 0.08,
181
+ "learning_rate": 9.177710575469181e-07,
182
+ "loss": 0.2241,
183
+ "step": 1856
184
+ },
185
+ {
186
+ "epoch": 0.09,
187
+ "learning_rate": 9.149248421239883e-07,
188
+ "loss": 0.2056,
189
+ "step": 1920
190
+ },
191
+ {
192
+ "epoch": 0.09,
193
+ "learning_rate": 9.120786267010584e-07,
194
+ "loss": 0.2956,
195
+ "step": 1984
196
+ },
197
+ {
198
+ "epoch": 0.09,
199
+ "learning_rate": 9.092324112781285e-07,
200
+ "loss": 0.2177,
201
+ "step": 2048
202
+ },
203
+ {
204
+ "epoch": 0.09,
205
+ "learning_rate": 9.063861958551987e-07,
206
+ "loss": 0.2379,
207
+ "step": 2112
208
+ },
209
+ {
210
+ "epoch": 0.1,
211
+ "learning_rate": 9.03539980432269e-07,
212
+ "loss": 0.3499,
213
+ "step": 2176
214
+ },
215
+ {
216
+ "epoch": 0.1,
217
+ "learning_rate": 9.006937650093391e-07,
218
+ "loss": 0.2133,
219
+ "step": 2240
220
+ },
221
+ {
222
+ "epoch": 0.1,
223
+ "learning_rate": 8.978475495864093e-07,
224
+ "loss": 0.283,
225
+ "step": 2304
226
+ },
227
+ {
228
+ "epoch": 0.11,
229
+ "learning_rate": 8.950013341634795e-07,
230
+ "loss": 0.2337,
231
+ "step": 2368
232
+ },
233
+ {
234
+ "epoch": 0.11,
235
+ "learning_rate": 8.921551187405496e-07,
236
+ "loss": 0.1282,
237
+ "step": 2432
238
+ },
239
+ {
240
+ "epoch": 0.11,
241
+ "learning_rate": 8.893089033176198e-07,
242
+ "loss": 0.1615,
243
+ "step": 2496
244
+ },
245
+ {
246
+ "epoch": 0.11,
247
+ "learning_rate": 8.864626878946901e-07,
248
+ "loss": 0.1537,
249
+ "step": 2560
250
+ },
251
+ {
252
+ "epoch": 0.12,
253
+ "learning_rate": 8.836164724717602e-07,
254
+ "loss": 0.2078,
255
+ "step": 2624
256
+ },
257
+ {
258
+ "epoch": 0.12,
259
+ "learning_rate": 8.807702570488304e-07,
260
+ "loss": 0.2886,
261
+ "step": 2688
262
+ },
263
+ {
264
+ "epoch": 0.12,
265
+ "learning_rate": 8.779240416259006e-07,
266
+ "loss": 0.2204,
267
+ "step": 2752
268
+ },
269
+ {
270
+ "epoch": 0.13,
271
+ "learning_rate": 8.750778262029706e-07,
272
+ "loss": 0.2415,
273
+ "step": 2816
274
+ },
275
+ {
276
+ "epoch": 0.13,
277
+ "learning_rate": 8.722316107800409e-07,
278
+ "loss": 0.1124,
279
+ "step": 2880
280
+ },
281
+ {
282
+ "epoch": 0.13,
283
+ "learning_rate": 8.69385395357111e-07,
284
+ "loss": 0.1638,
285
+ "step": 2944
286
+ },
287
+ {
288
+ "epoch": 0.13,
289
+ "learning_rate": 8.665391799341812e-07,
290
+ "loss": 0.2198,
291
+ "step": 3008
292
+ },
293
+ {
294
+ "epoch": 0.14,
295
+ "learning_rate": 8.636929645112514e-07,
296
+ "loss": 0.1793,
297
+ "step": 3072
298
+ },
299
+ {
300
+ "epoch": 0.14,
301
+ "learning_rate": 8.608467490883215e-07,
302
+ "loss": 0.1939,
303
+ "step": 3136
304
+ },
305
+ {
306
+ "epoch": 0.14,
307
+ "learning_rate": 8.580005336653917e-07,
308
+ "loss": 0.2771,
309
+ "step": 3200
310
+ },
311
+ {
312
+ "epoch": 0.15,
313
+ "learning_rate": 8.55154318242462e-07,
314
+ "loss": 0.1291,
315
+ "step": 3264
316
+ },
317
+ {
318
+ "epoch": 0.15,
319
+ "learning_rate": 8.523081028195322e-07,
320
+ "loss": 0.216,
321
+ "step": 3328
322
+ },
323
+ {
324
+ "epoch": 0.15,
325
+ "learning_rate": 8.494618873966023e-07,
326
+ "loss": 0.1985,
327
+ "step": 3392
328
+ },
329
+ {
330
+ "epoch": 0.15,
331
+ "learning_rate": 8.466156719736725e-07,
332
+ "loss": 0.2458,
333
+ "step": 3456
334
+ },
335
+ {
336
+ "epoch": 0.16,
337
+ "learning_rate": 8.437694565507427e-07,
338
+ "loss": 0.1253,
339
+ "step": 3520
340
+ },
341
+ {
342
+ "epoch": 0.16,
343
+ "learning_rate": 8.409232411278129e-07,
344
+ "loss": 0.2109,
345
+ "step": 3584
346
+ },
347
+ {
348
+ "epoch": 0.16,
349
+ "learning_rate": 8.38077025704883e-07,
350
+ "loss": 0.2105,
351
+ "step": 3648
352
+ },
353
+ {
354
+ "epoch": 0.17,
355
+ "learning_rate": 8.352308102819531e-07,
356
+ "loss": 0.2013,
357
+ "step": 3712
358
+ },
359
+ {
360
+ "epoch": 0.17,
361
+ "learning_rate": 8.323845948590233e-07,
362
+ "loss": 0.203,
363
+ "step": 3776
364
+ },
365
+ {
366
+ "epoch": 0.17,
367
+ "learning_rate": 8.295383794360935e-07,
368
+ "loss": 0.2366,
369
+ "step": 3840
370
+ },
371
+ {
372
+ "epoch": 0.17,
373
+ "learning_rate": 8.266921640131636e-07,
374
+ "loss": 0.2181,
375
+ "step": 3904
376
+ },
377
+ {
378
+ "epoch": 0.18,
379
+ "learning_rate": 8.238459485902339e-07,
380
+ "loss": 0.2611,
381
+ "step": 3968
382
+ },
383
+ {
384
+ "epoch": 0.18,
385
+ "learning_rate": 8.209997331673041e-07,
386
+ "loss": 0.1412,
387
+ "step": 4032
388
+ },
389
+ {
390
+ "epoch": 0.18,
391
+ "learning_rate": 8.181535177443743e-07,
392
+ "loss": 0.2678,
393
+ "step": 4096
394
+ },
395
+ {
396
+ "epoch": 0.19,
397
+ "learning_rate": 8.153073023214444e-07,
398
+ "loss": 0.1847,
399
+ "step": 4160
400
+ },
401
+ {
402
+ "epoch": 0.19,
403
+ "learning_rate": 8.124610868985146e-07,
404
+ "loss": 0.2395,
405
+ "step": 4224
406
+ },
407
+ {
408
+ "epoch": 0.19,
409
+ "learning_rate": 8.096148714755849e-07,
410
+ "loss": 0.1473,
411
+ "step": 4288
412
+ },
413
+ {
414
+ "epoch": 0.19,
415
+ "learning_rate": 8.06768656052655e-07,
416
+ "loss": 0.2675,
417
+ "step": 4352
418
+ },
419
+ {
420
+ "epoch": 0.2,
421
+ "learning_rate": 8.039224406297252e-07,
422
+ "loss": 0.2179,
423
+ "step": 4416
424
+ },
425
+ {
426
+ "epoch": 0.2,
427
+ "learning_rate": 8.010762252067952e-07,
428
+ "loss": 0.1667,
429
+ "step": 4480
430
+ },
431
+ {
432
+ "epoch": 0.2,
433
+ "learning_rate": 7.982300097838654e-07,
434
+ "loss": 0.2114,
435
+ "step": 4544
436
+ },
437
+ {
438
+ "epoch": 0.2,
439
+ "learning_rate": 7.953837943609356e-07,
440
+ "loss": 0.1929,
441
+ "step": 4608
442
+ },
443
+ {
444
+ "epoch": 0.21,
445
+ "learning_rate": 7.925375789380059e-07,
446
+ "loss": 0.2152,
447
+ "step": 4672
448
+ },
449
+ {
450
+ "epoch": 0.21,
451
+ "learning_rate": 7.89691363515076e-07,
452
+ "loss": 0.2065,
453
+ "step": 4736
454
+ },
455
+ {
456
+ "epoch": 0.21,
457
+ "learning_rate": 7.868451480921462e-07,
458
+ "loss": 0.1607,
459
+ "step": 4800
460
+ },
461
+ {
462
+ "epoch": 0.22,
463
+ "learning_rate": 7.839989326692163e-07,
464
+ "loss": 0.1592,
465
+ "step": 4864
466
+ },
467
+ {
468
+ "epoch": 0.22,
469
+ "learning_rate": 7.811527172462865e-07,
470
+ "loss": 0.2558,
471
+ "step": 4928
472
+ },
473
+ {
474
+ "epoch": 0.22,
475
+ "learning_rate": 7.783065018233567e-07,
476
+ "loss": 0.1564,
477
+ "step": 4992
478
+ },
479
+ {
480
+ "epoch": 0.22,
481
+ "learning_rate": 7.75460286400427e-07,
482
+ "loss": 0.2171,
483
+ "step": 5056
484
+ },
485
+ {
486
+ "epoch": 0.23,
487
+ "learning_rate": 7.726140709774971e-07,
488
+ "loss": 0.1973,
489
+ "step": 5120
490
+ },
491
+ {
492
+ "epoch": 0.23,
493
+ "learning_rate": 7.697678555545673e-07,
494
+ "loss": 0.1273,
495
+ "step": 5184
496
+ },
497
+ {
498
+ "epoch": 0.23,
499
+ "learning_rate": 7.669216401316375e-07,
500
+ "loss": 0.1158,
501
+ "step": 5248
502
+ },
503
+ {
504
+ "epoch": 0.24,
505
+ "learning_rate": 7.640754247087075e-07,
506
+ "loss": 0.151,
507
+ "step": 5312
508
+ },
509
+ {
510
+ "epoch": 0.24,
511
+ "learning_rate": 7.612292092857778e-07,
512
+ "loss": 0.2727,
513
+ "step": 5376
514
+ },
515
+ {
516
+ "epoch": 0.24,
517
+ "learning_rate": 7.58382993862848e-07,
518
+ "loss": 0.1271,
519
+ "step": 5440
520
+ },
521
+ {
522
+ "epoch": 0.24,
523
+ "learning_rate": 7.555367784399181e-07,
524
+ "loss": 0.2409,
525
+ "step": 5504
526
+ },
527
+ {
528
+ "epoch": 0.25,
529
+ "learning_rate": 7.526905630169883e-07,
530
+ "loss": 0.1456,
531
+ "step": 5568
532
+ },
533
+ {
534
+ "epoch": 0.25,
535
+ "learning_rate": 7.498443475940584e-07,
536
+ "loss": 0.249,
537
+ "step": 5632
538
+ },
539
+ {
540
+ "epoch": 0.25,
541
+ "learning_rate": 7.469981321711286e-07,
542
+ "loss": 0.2348,
543
+ "step": 5696
544
+ },
545
+ {
546
+ "epoch": 0.26,
547
+ "learning_rate": 7.441519167481989e-07,
548
+ "loss": 0.2074,
549
+ "step": 5760
550
+ },
551
+ {
552
+ "epoch": 0.26,
553
+ "learning_rate": 7.41305701325269e-07,
554
+ "loss": 0.1483,
555
+ "step": 5824
556
+ },
557
+ {
558
+ "epoch": 0.26,
559
+ "learning_rate": 7.384594859023392e-07,
560
+ "loss": 0.23,
561
+ "step": 5888
562
+ },
563
+ {
564
+ "epoch": 0.26,
565
+ "learning_rate": 7.356132704794094e-07,
566
+ "loss": 0.1294,
567
+ "step": 5952
568
+ },
569
+ {
570
+ "epoch": 0.27,
571
+ "learning_rate": 7.327670550564795e-07,
572
+ "loss": 0.2088,
573
+ "step": 6016
574
+ },
575
+ {
576
+ "epoch": 0.27,
577
+ "learning_rate": 7.299208396335498e-07,
578
+ "loss": 0.2233,
579
+ "step": 6080
580
+ },
581
+ {
582
+ "epoch": 0.27,
583
+ "learning_rate": 7.2707462421062e-07,
584
+ "loss": 0.1361,
585
+ "step": 6144
586
+ },
587
+ {
588
+ "epoch": 0.28,
589
+ "learning_rate": 7.2422840878769e-07,
590
+ "loss": 0.1664,
591
+ "step": 6208
592
+ },
593
+ {
594
+ "epoch": 0.28,
595
+ "learning_rate": 7.213821933647602e-07,
596
+ "loss": 0.1767,
597
+ "step": 6272
598
+ },
599
+ {
600
+ "epoch": 0.28,
601
+ "learning_rate": 7.185359779418304e-07,
602
+ "loss": 0.1613,
603
+ "step": 6336
604
+ },
605
+ {
606
+ "epoch": 0.28,
607
+ "learning_rate": 7.156897625189005e-07,
608
+ "loss": 0.1776,
609
+ "step": 6400
610
+ },
611
+ {
612
+ "epoch": 0.29,
613
+ "learning_rate": 7.128435470959708e-07,
614
+ "loss": 0.2314,
615
+ "step": 6464
616
+ },
617
+ {
618
+ "epoch": 0.29,
619
+ "learning_rate": 7.09997331673041e-07,
620
+ "loss": 0.1439,
621
+ "step": 6528
622
+ },
623
+ {
624
+ "epoch": 0.29,
625
+ "learning_rate": 7.071511162501111e-07,
626
+ "loss": 0.2195,
627
+ "step": 6592
628
+ },
629
+ {
630
+ "epoch": 0.3,
631
+ "learning_rate": 7.043049008271813e-07,
632
+ "loss": 0.248,
633
+ "step": 6656
634
+ },
635
+ {
636
+ "epoch": 0.3,
637
+ "learning_rate": 7.014586854042515e-07,
638
+ "loss": 0.2165,
639
+ "step": 6720
640
+ },
641
+ {
642
+ "epoch": 0.3,
643
+ "learning_rate": 6.986124699813218e-07,
644
+ "loss": 0.2642,
645
+ "step": 6784
646
+ },
647
+ {
648
+ "epoch": 0.3,
649
+ "learning_rate": 6.957662545583919e-07,
650
+ "loss": 0.236,
651
+ "step": 6848
652
+ },
653
+ {
654
+ "epoch": 0.31,
655
+ "learning_rate": 6.929200391354621e-07,
656
+ "loss": 0.2041,
657
+ "step": 6912
658
+ },
659
+ {
660
+ "epoch": 0.31,
661
+ "learning_rate": 6.900738237125323e-07,
662
+ "loss": 0.1925,
663
+ "step": 6976
664
+ },
665
+ {
666
+ "epoch": 0.31,
667
+ "learning_rate": 6.872276082896023e-07,
668
+ "loss": 0.2164,
669
+ "step": 7040
670
+ },
671
+ {
672
+ "epoch": 0.32,
673
+ "learning_rate": 6.843813928666725e-07,
674
+ "loss": 0.1399,
675
+ "step": 7104
676
+ },
677
+ {
678
+ "epoch": 0.32,
679
+ "learning_rate": 6.815351774437427e-07,
680
+ "loss": 0.1875,
681
+ "step": 7168
682
+ },
683
+ {
684
+ "epoch": 0.32,
685
+ "learning_rate": 6.786889620208129e-07,
686
+ "loss": 0.176,
687
+ "step": 7232
688
+ },
689
+ {
690
+ "epoch": 0.32,
691
+ "learning_rate": 6.758427465978831e-07,
692
+ "loss": 0.2326,
693
+ "step": 7296
694
+ },
695
+ {
696
+ "epoch": 0.33,
697
+ "learning_rate": 6.729965311749532e-07,
698
+ "loss": 0.1681,
699
+ "step": 7360
700
+ },
701
+ {
702
+ "epoch": 0.33,
703
+ "learning_rate": 6.701503157520234e-07,
704
+ "loss": 0.2422,
705
+ "step": 7424
706
+ },
707
+ {
708
+ "epoch": 0.33,
709
+ "learning_rate": 6.673041003290937e-07,
710
+ "loss": 0.1881,
711
+ "step": 7488
712
+ }
713
+ ],
714
+ "logging_steps": 64,
715
+ "max_steps": 22486,
716
+ "num_input_tokens_seen": 0,
717
+ "num_train_epochs": 1,
718
+ "save_steps": 7495,
719
+ "total_flos": 1.5913166800896e+19,
720
+ "train_batch_size": 1,
721
+ "trial_name": null,
722
+ "trial_params": null
723
+ }
checkpoint-7495/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5dd66348aef7664171d12847300a1c626bfa2b1ede8e5141c5a2fc70bac4b83
3
+ size 5048
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
+ }
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
+ }
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f39b5ef561305e7a99fcdb96673580927609f662f2ba94c4c5c29493f86d7c2e
3
+ size 4992706480
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f69ac3bbcf11b45c5d70d2298ab7f9355f25da0584b8fd1f1ed3bf04d63a5d35
3
+ size 1180663192
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
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
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5dd66348aef7664171d12847300a1c626bfa2b1ede8e5141c5a2fc70bac4b83
3
+ size 5048