nbtpj commited on
Commit
9db1f3e
1 Parent(s): 72688c9

End of training

Browse files
last-checkpoint/config.json DELETED
@@ -1,75 +0,0 @@
1
- {
2
- "_name_or_path": "facebook/bart-base",
3
- "activation_dropout": 0.1,
4
- "activation_function": "gelu",
5
- "add_bias_logits": false,
6
- "add_final_layer_norm": false,
7
- "architectures": [
8
- "EncoderDecoderForConditionalGeneration"
9
- ],
10
- "attention_dropout": 0.1,
11
- "bos_token_id": 0,
12
- "classif_dropout": 0.1,
13
- "classifier_dropout": 0.0,
14
- "d_model": 768,
15
- "decoder_attention_heads": 12,
16
- "decoder_ffn_dim": 3072,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 6,
19
- "decoder_start_token_id": 2,
20
- "dropout": 0.1,
21
- "early_stopping": true,
22
- "encoder_attention_heads": 12,
23
- "encoder_ffn_dim": 3072,
24
- "encoder_layerdrop": 0.0,
25
- "encoder_layers": 6,
26
- "eos_token_id": 2,
27
- "forced_bos_token_id": 0,
28
- "forced_eos_token_id": 2,
29
- "gradient_checkpointing": false,
30
- "id2label": {
31
- "0": "LABEL_0",
32
- "1": "LABEL_1",
33
- "2": "LABEL_2"
34
- },
35
- "init_std": 0.02,
36
- "is_encoder_decoder": true,
37
- "label2id": {
38
- "LABEL_0": 0,
39
- "LABEL_1": 1,
40
- "LABEL_2": 2
41
- },
42
- "max_position_embeddings": 1024,
43
- "model_type": "bart",
44
- "no_repeat_ngram_size": 3,
45
- "normalize_before": false,
46
- "normalize_embedding": true,
47
- "num_beams": 4,
48
- "num_hidden_layers": 6,
49
- "pad_token_id": 1,
50
- "scale_embedding": false,
51
- "task_specific_params": {
52
- "summarization": {
53
- "length_penalty": 1.0,
54
- "max_length": 128,
55
- "min_length": 12,
56
- "num_beams": 4
57
- },
58
- "summarization_cnn": {
59
- "length_penalty": 2.0,
60
- "max_length": 142,
61
- "min_length": 56,
62
- "num_beams": 4
63
- },
64
- "summarization_xsum": {
65
- "length_penalty": 1.0,
66
- "max_length": 62,
67
- "min_length": 11,
68
- "num_beams": 6
69
- }
70
- },
71
- "torch_dtype": "float32",
72
- "transformers_version": "4.20.1",
73
- "use_cache": true,
74
- "vocab_size": 50265
75
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
last-checkpoint/merges.txt DELETED
The diff for this file is too large to render. See raw diff
 
last-checkpoint/optimizer.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:2754a19deafed33f80b8bc9e1d496271fbdefd4dde6d02139929e3ddf01b008f
3
- size 1115513717
 
 
 
 
last-checkpoint/pytorch_model.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3916aa768b3547ed097d45447da1e5f44df02e98af7f4c9081b40d19b99444c0
3
- size 557969145
 
 
 
 
last-checkpoint/rng_state.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:0e6f2984904c3cd3157bcf9409a340770e020350dad6bed8f8d65bce2381442b
3
- size 15523
 
 
 
 
last-checkpoint/scheduler.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3d3cf89aef0bc5afe4fc7de6b797864a345f7acd73a830413d35a9d7579e5dbe
3
- size 623
 
 
 
 
last-checkpoint/special_tokens_map.json DELETED
@@ -1,15 +0,0 @@
1
- {
2
- "bos_token": "<s>",
3
- "cls_token": "<s>",
4
- "eos_token": "</s>",
5
- "mask_token": {
6
- "content": "<mask>",
7
- "lstrip": true,
8
- "normalized": false,
9
- "rstrip": false,
10
- "single_word": false
11
- },
12
- "pad_token": "<pad>",
13
- "sep_token": "</s>",
14
- "unk_token": "<unk>"
15
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
last-checkpoint/tokenizer.json DELETED
The diff for this file is too large to render. See raw diff
 
last-checkpoint/tokenizer_config.json DELETED
@@ -1,16 +0,0 @@
1
- {
2
- "add_prefix_space": false,
3
- "bos_token": "<s>",
4
- "cls_token": "<s>",
5
- "eos_token": "</s>",
6
- "errors": "replace",
7
- "mask_token": "<mask>",
8
- "model_max_length": 1024,
9
- "name_or_path": "facebook/bart-base",
10
- "pad_token": "<pad>",
11
- "sep_token": "</s>",
12
- "special_tokens_map_file": null,
13
- "tokenizer_class": "BartTokenizer",
14
- "trim_offsets": true,
15
- "unk_token": "<unk>"
16
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
last-checkpoint/trainer_state.json DELETED
@@ -1,1840 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 1.7695983011856309,
5
- "global_step": 30000,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.01,
12
- "learning_rate": 4.96526184702104e-05,
13
- "loss": 1.1303,
14
- "step": 100
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 4.932490004588058e-05,
19
- "loss": 0.8589,
20
- "step": 200
21
- },
22
- {
23
- "epoch": 0.02,
24
- "learning_rate": 4.8997181621550765e-05,
25
- "loss": 0.9136,
26
- "step": 300
27
- },
28
- {
29
- "epoch": 0.02,
30
- "learning_rate": 4.866946319722095e-05,
31
- "loss": 0.8209,
32
- "step": 400
33
- },
34
- {
35
- "epoch": 0.03,
36
- "learning_rate": 4.8341744772891136e-05,
37
- "loss": 0.8212,
38
- "step": 500
39
- },
40
- {
41
- "epoch": 0.04,
42
- "learning_rate": 4.801402634856132e-05,
43
- "loss": 0.7584,
44
- "step": 600
45
- },
46
- {
47
- "epoch": 0.04,
48
- "learning_rate": 4.768630792423151e-05,
49
- "loss": 0.8386,
50
- "step": 700
51
- },
52
- {
53
- "epoch": 0.05,
54
- "learning_rate": 4.735858949990168e-05,
55
- "loss": 0.7838,
56
- "step": 800
57
- },
58
- {
59
- "epoch": 0.05,
60
- "learning_rate": 4.703087107557187e-05,
61
- "loss": 0.7109,
62
- "step": 900
63
- },
64
- {
65
- "epoch": 0.06,
66
- "learning_rate": 4.670315265124205e-05,
67
- "loss": 0.8817,
68
- "step": 1000
69
- },
70
- {
71
- "epoch": 0.06,
72
- "learning_rate": 4.637543422691224e-05,
73
- "loss": 0.7,
74
- "step": 1100
75
- },
76
- {
77
- "epoch": 0.07,
78
- "learning_rate": 4.6047715802582424e-05,
79
- "loss": 0.7984,
80
- "step": 1200
81
- },
82
- {
83
- "epoch": 0.08,
84
- "learning_rate": 4.571999737825261e-05,
85
- "loss": 0.7983,
86
- "step": 1300
87
- },
88
- {
89
- "epoch": 0.08,
90
- "learning_rate": 4.539227895392279e-05,
91
- "loss": 0.8209,
92
- "step": 1400
93
- },
94
- {
95
- "epoch": 0.09,
96
- "learning_rate": 4.506456052959298e-05,
97
- "loss": 0.7698,
98
- "step": 1500
99
- },
100
- {
101
- "epoch": 0.09,
102
- "learning_rate": 4.473684210526316e-05,
103
- "loss": 0.6812,
104
- "step": 1600
105
- },
106
- {
107
- "epoch": 0.1,
108
- "learning_rate": 4.440912368093335e-05,
109
- "loss": 0.7507,
110
- "step": 1700
111
- },
112
- {
113
- "epoch": 0.11,
114
- "learning_rate": 4.408140525660353e-05,
115
- "loss": 0.7081,
116
- "step": 1800
117
- },
118
- {
119
- "epoch": 0.11,
120
- "learning_rate": 4.375368683227371e-05,
121
- "loss": 0.809,
122
- "step": 1900
123
- },
124
- {
125
- "epoch": 0.12,
126
- "learning_rate": 4.3425968407943895e-05,
127
- "loss": 0.7504,
128
- "step": 2000
129
- },
130
- {
131
- "epoch": 0.12,
132
- "learning_rate": 4.3098249983614084e-05,
133
- "loss": 0.764,
134
- "step": 2100
135
- },
136
- {
137
- "epoch": 0.13,
138
- "learning_rate": 4.2770531559284266e-05,
139
- "loss": 0.7374,
140
- "step": 2200
141
- },
142
- {
143
- "epoch": 0.14,
144
- "learning_rate": 4.244281313495445e-05,
145
- "loss": 0.786,
146
- "step": 2300
147
- },
148
- {
149
- "epoch": 0.14,
150
- "learning_rate": 4.211509471062463e-05,
151
- "loss": 0.8193,
152
- "step": 2400
153
- },
154
- {
155
- "epoch": 0.15,
156
- "learning_rate": 4.178737628629482e-05,
157
- "loss": 0.6467,
158
- "step": 2500
159
- },
160
- {
161
- "epoch": 0.15,
162
- "learning_rate": 4.1459657861965e-05,
163
- "loss": 0.7984,
164
- "step": 2600
165
- },
166
- {
167
- "epoch": 0.16,
168
- "learning_rate": 4.113193943763518e-05,
169
- "loss": 0.7313,
170
- "step": 2700
171
- },
172
- {
173
- "epoch": 0.17,
174
- "learning_rate": 4.080422101330537e-05,
175
- "loss": 0.6847,
176
- "step": 2800
177
- },
178
- {
179
- "epoch": 0.17,
180
- "learning_rate": 4.0476502588975554e-05,
181
- "loss": 0.7939,
182
- "step": 2900
183
- },
184
- {
185
- "epoch": 0.18,
186
- "learning_rate": 4.0148784164645736e-05,
187
- "loss": 0.7742,
188
- "step": 3000
189
- },
190
- {
191
- "epoch": 0.18,
192
- "learning_rate": 3.982106574031592e-05,
193
- "loss": 0.7562,
194
- "step": 3100
195
- },
196
- {
197
- "epoch": 0.19,
198
- "learning_rate": 3.949334731598611e-05,
199
- "loss": 0.6686,
200
- "step": 3200
201
- },
202
- {
203
- "epoch": 0.19,
204
- "learning_rate": 3.916562889165629e-05,
205
- "loss": 0.7455,
206
- "step": 3300
207
- },
208
- {
209
- "epoch": 0.2,
210
- "learning_rate": 3.883791046732648e-05,
211
- "loss": 0.7536,
212
- "step": 3400
213
- },
214
- {
215
- "epoch": 0.21,
216
- "learning_rate": 3.8510192042996654e-05,
217
- "loss": 0.8175,
218
- "step": 3500
219
- },
220
- {
221
- "epoch": 0.21,
222
- "learning_rate": 3.818247361866684e-05,
223
- "loss": 0.731,
224
- "step": 3600
225
- },
226
- {
227
- "epoch": 0.22,
228
- "learning_rate": 3.7854755194337025e-05,
229
- "loss": 0.7193,
230
- "step": 3700
231
- },
232
- {
233
- "epoch": 0.22,
234
- "learning_rate": 3.7527036770007214e-05,
235
- "loss": 0.8051,
236
- "step": 3800
237
- },
238
- {
239
- "epoch": 0.23,
240
- "learning_rate": 3.7199318345677396e-05,
241
- "loss": 0.7182,
242
- "step": 3900
243
- },
244
- {
245
- "epoch": 0.24,
246
- "learning_rate": 3.6871599921347585e-05,
247
- "loss": 0.7214,
248
- "step": 4000
249
- },
250
- {
251
- "epoch": 0.24,
252
- "learning_rate": 3.654388149701776e-05,
253
- "loss": 0.7336,
254
- "step": 4100
255
- },
256
- {
257
- "epoch": 0.25,
258
- "learning_rate": 3.621616307268795e-05,
259
- "loss": 0.8413,
260
- "step": 4200
261
- },
262
- {
263
- "epoch": 0.25,
264
- "learning_rate": 3.588844464835813e-05,
265
- "loss": 0.6897,
266
- "step": 4300
267
- },
268
- {
269
- "epoch": 0.26,
270
- "learning_rate": 3.556072622402832e-05,
271
- "loss": 0.6835,
272
- "step": 4400
273
- },
274
- {
275
- "epoch": 0.27,
276
- "learning_rate": 3.52330077996985e-05,
277
- "loss": 0.6818,
278
- "step": 4500
279
- },
280
- {
281
- "epoch": 0.27,
282
- "learning_rate": 3.4905289375368684e-05,
283
- "loss": 0.7128,
284
- "step": 4600
285
- },
286
- {
287
- "epoch": 0.28,
288
- "learning_rate": 3.4577570951038866e-05,
289
- "loss": 0.7883,
290
- "step": 4700
291
- },
292
- {
293
- "epoch": 0.28,
294
- "learning_rate": 3.4249852526709055e-05,
295
- "loss": 0.7018,
296
- "step": 4800
297
- },
298
- {
299
- "epoch": 0.29,
300
- "learning_rate": 3.392213410237924e-05,
301
- "loss": 0.6493,
302
- "step": 4900
303
- },
304
- {
305
- "epoch": 0.29,
306
- "learning_rate": 3.3594415678049426e-05,
307
- "loss": 0.7293,
308
- "step": 5000
309
- },
310
- {
311
- "epoch": 0.3,
312
- "learning_rate": 3.32666972537196e-05,
313
- "loss": 0.7594,
314
- "step": 5100
315
- },
316
- {
317
- "epoch": 0.31,
318
- "learning_rate": 3.293897882938979e-05,
319
- "loss": 0.6165,
320
- "step": 5200
321
- },
322
- {
323
- "epoch": 0.31,
324
- "learning_rate": 3.261126040505997e-05,
325
- "loss": 0.7486,
326
- "step": 5300
327
- },
328
- {
329
- "epoch": 0.32,
330
- "learning_rate": 3.228354198073016e-05,
331
- "loss": 0.7611,
332
- "step": 5400
333
- },
334
- {
335
- "epoch": 0.32,
336
- "learning_rate": 3.1955823556400344e-05,
337
- "loss": 0.7024,
338
- "step": 5500
339
- },
340
- {
341
- "epoch": 0.33,
342
- "learning_rate": 3.1628105132070526e-05,
343
- "loss": 0.6861,
344
- "step": 5600
345
- },
346
- {
347
- "epoch": 0.34,
348
- "learning_rate": 3.130038670774071e-05,
349
- "loss": 0.6787,
350
- "step": 5700
351
- },
352
- {
353
- "epoch": 0.34,
354
- "learning_rate": 3.09726682834109e-05,
355
- "loss": 0.7645,
356
- "step": 5800
357
- },
358
- {
359
- "epoch": 0.35,
360
- "learning_rate": 3.064494985908108e-05,
361
- "loss": 0.7743,
362
- "step": 5900
363
- },
364
- {
365
- "epoch": 0.35,
366
- "learning_rate": 3.0317231434751264e-05,
367
- "loss": 0.7202,
368
- "step": 6000
369
- },
370
- {
371
- "epoch": 0.36,
372
- "learning_rate": 2.998951301042145e-05,
373
- "loss": 0.6463,
374
- "step": 6100
375
- },
376
- {
377
- "epoch": 0.37,
378
- "learning_rate": 2.966179458609163e-05,
379
- "loss": 0.7259,
380
- "step": 6200
381
- },
382
- {
383
- "epoch": 0.37,
384
- "learning_rate": 2.9334076161761814e-05,
385
- "loss": 0.7776,
386
- "step": 6300
387
- },
388
- {
389
- "epoch": 0.38,
390
- "learning_rate": 2.9006357737432e-05,
391
- "loss": 0.7006,
392
- "step": 6400
393
- },
394
- {
395
- "epoch": 0.38,
396
- "learning_rate": 2.8678639313102185e-05,
397
- "loss": 0.7752,
398
- "step": 6500
399
- },
400
- {
401
- "epoch": 0.39,
402
- "learning_rate": 2.835092088877237e-05,
403
- "loss": 0.7215,
404
- "step": 6600
405
- },
406
- {
407
- "epoch": 0.4,
408
- "learning_rate": 2.8023202464442556e-05,
409
- "loss": 0.6722,
410
- "step": 6700
411
- },
412
- {
413
- "epoch": 0.4,
414
- "learning_rate": 2.7695484040112735e-05,
415
- "loss": 0.7795,
416
- "step": 6800
417
- },
418
- {
419
- "epoch": 0.41,
420
- "learning_rate": 2.736776561578292e-05,
421
- "loss": 0.6765,
422
- "step": 6900
423
- },
424
- {
425
- "epoch": 0.41,
426
- "learning_rate": 2.7040047191453106e-05,
427
- "loss": 0.6952,
428
- "step": 7000
429
- },
430
- {
431
- "epoch": 0.42,
432
- "learning_rate": 2.671232876712329e-05,
433
- "loss": 0.7187,
434
- "step": 7100
435
- },
436
- {
437
- "epoch": 0.42,
438
- "learning_rate": 2.6384610342793477e-05,
439
- "loss": 0.7934,
440
- "step": 7200
441
- },
442
- {
443
- "epoch": 0.43,
444
- "learning_rate": 2.6056891918463656e-05,
445
- "loss": 0.6715,
446
- "step": 7300
447
- },
448
- {
449
- "epoch": 0.44,
450
- "learning_rate": 2.572917349413384e-05,
451
- "loss": 0.6955,
452
- "step": 7400
453
- },
454
- {
455
- "epoch": 0.44,
456
- "learning_rate": 2.5401455069804027e-05,
457
- "loss": 0.7425,
458
- "step": 7500
459
- },
460
- {
461
- "epoch": 0.45,
462
- "learning_rate": 2.5073736645474212e-05,
463
- "loss": 0.7889,
464
- "step": 7600
465
- },
466
- {
467
- "epoch": 0.45,
468
- "learning_rate": 2.4746018221144394e-05,
469
- "loss": 0.6897,
470
- "step": 7700
471
- },
472
- {
473
- "epoch": 0.46,
474
- "learning_rate": 2.441829979681458e-05,
475
- "loss": 0.6955,
476
- "step": 7800
477
- },
478
- {
479
- "epoch": 0.47,
480
- "learning_rate": 2.4090581372484762e-05,
481
- "loss": 0.706,
482
- "step": 7900
483
- },
484
- {
485
- "epoch": 0.47,
486
- "learning_rate": 2.3762862948154948e-05,
487
- "loss": 0.7045,
488
- "step": 8000
489
- },
490
- {
491
- "epoch": 0.48,
492
- "learning_rate": 2.343514452382513e-05,
493
- "loss": 0.7654,
494
- "step": 8100
495
- },
496
- {
497
- "epoch": 0.48,
498
- "learning_rate": 2.3107426099495312e-05,
499
- "loss": 0.7681,
500
- "step": 8200
501
- },
502
- {
503
- "epoch": 0.49,
504
- "learning_rate": 2.2779707675165497e-05,
505
- "loss": 0.6995,
506
- "step": 8300
507
- },
508
- {
509
- "epoch": 0.5,
510
- "learning_rate": 2.2451989250835683e-05,
511
- "loss": 0.6944,
512
- "step": 8400
513
- },
514
- {
515
- "epoch": 0.5,
516
- "learning_rate": 2.2124270826505865e-05,
517
- "loss": 0.673,
518
- "step": 8500
519
- },
520
- {
521
- "epoch": 0.51,
522
- "learning_rate": 2.179655240217605e-05,
523
- "loss": 0.6161,
524
- "step": 8600
525
- },
526
- {
527
- "epoch": 0.51,
528
- "learning_rate": 2.1468833977846236e-05,
529
- "loss": 0.7853,
530
- "step": 8700
531
- },
532
- {
533
- "epoch": 0.52,
534
- "learning_rate": 2.1141115553516418e-05,
535
- "loss": 0.74,
536
- "step": 8800
537
- },
538
- {
539
- "epoch": 0.52,
540
- "learning_rate": 2.0813397129186604e-05,
541
- "loss": 0.6185,
542
- "step": 8900
543
- },
544
- {
545
- "epoch": 0.53,
546
- "learning_rate": 2.048567870485679e-05,
547
- "loss": 0.769,
548
- "step": 9000
549
- },
550
- {
551
- "epoch": 0.54,
552
- "learning_rate": 2.015796028052697e-05,
553
- "loss": 0.6912,
554
- "step": 9100
555
- },
556
- {
557
- "epoch": 0.54,
558
- "learning_rate": 1.9830241856197157e-05,
559
- "loss": 0.6794,
560
- "step": 9200
561
- },
562
- {
563
- "epoch": 0.55,
564
- "learning_rate": 1.950252343186734e-05,
565
- "loss": 0.6743,
566
- "step": 9300
567
- },
568
- {
569
- "epoch": 0.55,
570
- "learning_rate": 1.9174805007537524e-05,
571
- "loss": 0.762,
572
- "step": 9400
573
- },
574
- {
575
- "epoch": 0.56,
576
- "learning_rate": 1.884708658320771e-05,
577
- "loss": 0.6399,
578
- "step": 9500
579
- },
580
- {
581
- "epoch": 0.57,
582
- "learning_rate": 1.8519368158877892e-05,
583
- "loss": 0.638,
584
- "step": 9600
585
- },
586
- {
587
- "epoch": 0.57,
588
- "learning_rate": 1.8191649734548078e-05,
589
- "loss": 0.6879,
590
- "step": 9700
591
- },
592
- {
593
- "epoch": 0.58,
594
- "learning_rate": 1.7863931310218263e-05,
595
- "loss": 0.6728,
596
- "step": 9800
597
- },
598
- {
599
- "epoch": 0.58,
600
- "learning_rate": 1.7536212885888445e-05,
601
- "loss": 0.6807,
602
- "step": 9900
603
- },
604
- {
605
- "epoch": 0.59,
606
- "learning_rate": 1.720849446155863e-05,
607
- "loss": 0.6673,
608
- "step": 10000
609
- },
610
- {
611
- "epoch": 0.59,
612
- "eval_loss": 0.8870079517364502,
613
- "eval_runtime": 547.4308,
614
- "eval_samples_per_second": 29.116,
615
- "eval_steps_per_second": 1.821,
616
- "step": 10000
617
- },
618
- {
619
- "epoch": 0.6,
620
- "learning_rate": 1.6880776037228813e-05,
621
- "loss": 0.6613,
622
- "step": 10100
623
- },
624
- {
625
- "epoch": 0.6,
626
- "learning_rate": 1.6553057612899e-05,
627
- "loss": 0.6208,
628
- "step": 10200
629
- },
630
- {
631
- "epoch": 0.61,
632
- "learning_rate": 1.6225339188569184e-05,
633
- "loss": 0.6037,
634
- "step": 10300
635
- },
636
- {
637
- "epoch": 0.61,
638
- "learning_rate": 1.5897620764239366e-05,
639
- "loss": 0.7217,
640
- "step": 10400
641
- },
642
- {
643
- "epoch": 0.62,
644
- "learning_rate": 1.556990233990955e-05,
645
- "loss": 0.617,
646
- "step": 10500
647
- },
648
- {
649
- "epoch": 0.63,
650
- "learning_rate": 1.5242183915579735e-05,
651
- "loss": 0.6737,
652
- "step": 10600
653
- },
654
- {
655
- "epoch": 0.63,
656
- "learning_rate": 1.4914465491249917e-05,
657
- "loss": 0.6909,
658
- "step": 10700
659
- },
660
- {
661
- "epoch": 0.64,
662
- "learning_rate": 1.4586747066920103e-05,
663
- "loss": 0.6317,
664
- "step": 10800
665
- },
666
- {
667
- "epoch": 0.64,
668
- "learning_rate": 1.4259028642590285e-05,
669
- "loss": 0.7334,
670
- "step": 10900
671
- },
672
- {
673
- "epoch": 0.65,
674
- "learning_rate": 1.393131021826047e-05,
675
- "loss": 0.779,
676
- "step": 11000
677
- },
678
- {
679
- "epoch": 0.65,
680
- "learning_rate": 1.3603591793930656e-05,
681
- "loss": 0.5746,
682
- "step": 11100
683
- },
684
- {
685
- "epoch": 0.66,
686
- "learning_rate": 1.3275873369600838e-05,
687
- "loss": 0.6017,
688
- "step": 11200
689
- },
690
- {
691
- "epoch": 0.67,
692
- "learning_rate": 1.2948154945271024e-05,
693
- "loss": 0.714,
694
- "step": 11300
695
- },
696
- {
697
- "epoch": 0.67,
698
- "learning_rate": 1.262043652094121e-05,
699
- "loss": 0.6963,
700
- "step": 11400
701
- },
702
- {
703
- "epoch": 0.68,
704
- "learning_rate": 1.2292718096611391e-05,
705
- "loss": 0.6632,
706
- "step": 11500
707
- },
708
- {
709
- "epoch": 0.68,
710
- "learning_rate": 1.1964999672281577e-05,
711
- "loss": 0.6326,
712
- "step": 11600
713
- },
714
- {
715
- "epoch": 0.69,
716
- "learning_rate": 1.163728124795176e-05,
717
- "loss": 0.5966,
718
- "step": 11700
719
- },
720
- {
721
- "epoch": 0.7,
722
- "learning_rate": 1.1309562823621945e-05,
723
- "loss": 0.6477,
724
- "step": 11800
725
- },
726
- {
727
- "epoch": 0.7,
728
- "learning_rate": 1.098184439929213e-05,
729
- "loss": 0.7871,
730
- "step": 11900
731
- },
732
- {
733
- "epoch": 0.71,
734
- "learning_rate": 1.0654125974962314e-05,
735
- "loss": 0.5873,
736
- "step": 12000
737
- },
738
- {
739
- "epoch": 0.71,
740
- "learning_rate": 1.0326407550632498e-05,
741
- "loss": 0.6291,
742
- "step": 12100
743
- },
744
- {
745
- "epoch": 0.72,
746
- "learning_rate": 9.998689126302682e-06,
747
- "loss": 0.681,
748
- "step": 12200
749
- },
750
- {
751
- "epoch": 0.73,
752
- "learning_rate": 9.670970701972865e-06,
753
- "loss": 0.6848,
754
- "step": 12300
755
- },
756
- {
757
- "epoch": 0.73,
758
- "learning_rate": 9.34325227764305e-06,
759
- "loss": 0.6298,
760
- "step": 12400
761
- },
762
- {
763
- "epoch": 0.74,
764
- "learning_rate": 9.015533853313233e-06,
765
- "loss": 0.6394,
766
- "step": 12500
767
- },
768
- {
769
- "epoch": 0.74,
770
- "learning_rate": 8.687815428983417e-06,
771
- "loss": 0.6416,
772
- "step": 12600
773
- },
774
- {
775
- "epoch": 0.75,
776
- "learning_rate": 8.360097004653602e-06,
777
- "loss": 0.6916,
778
- "step": 12700
779
- },
780
- {
781
- "epoch": 0.76,
782
- "learning_rate": 8.032378580323786e-06,
783
- "loss": 0.7215,
784
- "step": 12800
785
- },
786
- {
787
- "epoch": 0.76,
788
- "learning_rate": 7.70466015599397e-06,
789
- "loss": 0.6666,
790
- "step": 12900
791
- },
792
- {
793
- "epoch": 0.77,
794
- "learning_rate": 7.376941731664154e-06,
795
- "loss": 0.715,
796
- "step": 13000
797
- },
798
- {
799
- "epoch": 0.77,
800
- "learning_rate": 7.049223307334339e-06,
801
- "loss": 0.6387,
802
- "step": 13100
803
- },
804
- {
805
- "epoch": 0.78,
806
- "learning_rate": 6.721504883004523e-06,
807
- "loss": 0.7107,
808
- "step": 13200
809
- },
810
- {
811
- "epoch": 0.78,
812
- "learning_rate": 6.393786458674707e-06,
813
- "loss": 0.64,
814
- "step": 13300
815
- },
816
- {
817
- "epoch": 0.79,
818
- "learning_rate": 6.066068034344892e-06,
819
- "loss": 0.6918,
820
- "step": 13400
821
- },
822
- {
823
- "epoch": 0.8,
824
- "learning_rate": 5.7383496100150754e-06,
825
- "loss": 0.769,
826
- "step": 13500
827
- },
828
- {
829
- "epoch": 0.8,
830
- "learning_rate": 5.410631185685259e-06,
831
- "loss": 0.5846,
832
- "step": 13600
833
- },
834
- {
835
- "epoch": 0.81,
836
- "learning_rate": 5.082912761355443e-06,
837
- "loss": 0.6344,
838
- "step": 13700
839
- },
840
- {
841
- "epoch": 0.81,
842
- "learning_rate": 4.755194337025628e-06,
843
- "loss": 0.6205,
844
- "step": 13800
845
- },
846
- {
847
- "epoch": 0.82,
848
- "learning_rate": 4.4274759126958116e-06,
849
- "loss": 0.6575,
850
- "step": 13900
851
- },
852
- {
853
- "epoch": 0.83,
854
- "learning_rate": 4.099757488365996e-06,
855
- "loss": 0.7167,
856
- "step": 14000
857
- },
858
- {
859
- "epoch": 0.83,
860
- "learning_rate": 3.77203906403618e-06,
861
- "loss": 0.6486,
862
- "step": 14100
863
- },
864
- {
865
- "epoch": 0.84,
866
- "learning_rate": 3.4443206397063643e-06,
867
- "loss": 0.6339,
868
- "step": 14200
869
- },
870
- {
871
- "epoch": 0.84,
872
- "learning_rate": 3.1166022153765485e-06,
873
- "loss": 0.6003,
874
- "step": 14300
875
- },
876
- {
877
- "epoch": 0.85,
878
- "learning_rate": 2.7888837910467328e-06,
879
- "loss": 0.6888,
880
- "step": 14400
881
- },
882
- {
883
- "epoch": 0.86,
884
- "learning_rate": 2.461165366716917e-06,
885
- "loss": 0.6196,
886
- "step": 14500
887
- },
888
- {
889
- "epoch": 0.86,
890
- "learning_rate": 2.133446942387101e-06,
891
- "loss": 0.7685,
892
- "step": 14600
893
- },
894
- {
895
- "epoch": 0.87,
896
- "learning_rate": 1.8057285180572853e-06,
897
- "loss": 0.6597,
898
- "step": 14700
899
- },
900
- {
901
- "epoch": 0.87,
902
- "learning_rate": 1.4780100937274693e-06,
903
- "loss": 0.6491,
904
- "step": 14800
905
- },
906
- {
907
- "epoch": 0.88,
908
- "learning_rate": 1.1502916693976536e-06,
909
- "loss": 0.5624,
910
- "step": 14900
911
- },
912
- {
913
- "epoch": 0.88,
914
- "learning_rate": 8.225732450678378e-07,
915
- "loss": 0.6941,
916
- "step": 15000
917
- },
918
- {
919
- "epoch": 0.89,
920
- "learning_rate": 4.948548207380218e-07,
921
- "loss": 0.7049,
922
- "step": 15100
923
- },
924
- {
925
- "epoch": 0.9,
926
- "learning_rate": 1.6713639640820606e-07,
927
- "loss": 0.6312,
928
- "step": 15200
929
- },
930
- {
931
- "epoch": 0.9,
932
- "learning_rate": 0.0,
933
- "loss": 0.6369,
934
- "step": 15300
935
- },
936
- {
937
- "epoch": 0.91,
938
- "learning_rate": 0.0,
939
- "loss": 0.6321,
940
- "step": 15400
941
- },
942
- {
943
- "epoch": 0.91,
944
- "learning_rate": 0.0,
945
- "loss": 0.7197,
946
- "step": 15500
947
- },
948
- {
949
- "epoch": 0.92,
950
- "learning_rate": 0.0,
951
- "loss": 0.6451,
952
- "step": 15600
953
- },
954
- {
955
- "epoch": 0.93,
956
- "learning_rate": 0.0,
957
- "loss": 0.6704,
958
- "step": 15700
959
- },
960
- {
961
- "epoch": 0.93,
962
- "learning_rate": 0.0,
963
- "loss": 0.6144,
964
- "step": 15800
965
- },
966
- {
967
- "epoch": 0.94,
968
- "learning_rate": 0.0,
969
- "loss": 0.6497,
970
- "step": 15900
971
- },
972
- {
973
- "epoch": 0.94,
974
- "learning_rate": 0.0,
975
- "loss": 0.7311,
976
- "step": 16000
977
- },
978
- {
979
- "epoch": 0.95,
980
- "learning_rate": 0.0,
981
- "loss": 0.6799,
982
- "step": 16100
983
- },
984
- {
985
- "epoch": 0.96,
986
- "learning_rate": 0.0,
987
- "loss": 0.6354,
988
- "step": 16200
989
- },
990
- {
991
- "epoch": 0.96,
992
- "learning_rate": 0.0,
993
- "loss": 0.6063,
994
- "step": 16300
995
- },
996
- {
997
- "epoch": 0.97,
998
- "learning_rate": 0.0,
999
- "loss": 0.6489,
1000
- "step": 16400
1001
- },
1002
- {
1003
- "epoch": 0.97,
1004
- "learning_rate": 0.0,
1005
- "loss": 0.6334,
1006
- "step": 16500
1007
- },
1008
- {
1009
- "epoch": 0.98,
1010
- "learning_rate": 0.0,
1011
- "loss": 0.689,
1012
- "step": 16600
1013
- },
1014
- {
1015
- "epoch": 0.99,
1016
- "learning_rate": 0.0,
1017
- "loss": 0.6852,
1018
- "step": 16700
1019
- },
1020
- {
1021
- "epoch": 0.99,
1022
- "learning_rate": 0.0,
1023
- "loss": 0.6481,
1024
- "step": 16800
1025
- },
1026
- {
1027
- "epoch": 1.0,
1028
- "learning_rate": 0.0,
1029
- "loss": 0.7362,
1030
- "step": 16900
1031
- },
1032
- {
1033
- "epoch": 1.0,
1034
- "learning_rate": 0.0,
1035
- "loss": 0.7102,
1036
- "step": 17000
1037
- },
1038
- {
1039
- "epoch": 1.01,
1040
- "learning_rate": 0.0,
1041
- "loss": 0.6313,
1042
- "step": 17100
1043
- },
1044
- {
1045
- "epoch": 1.01,
1046
- "learning_rate": 0.0,
1047
- "loss": 0.7327,
1048
- "step": 17200
1049
- },
1050
- {
1051
- "epoch": 1.02,
1052
- "learning_rate": 0.0,
1053
- "loss": 0.624,
1054
- "step": 17300
1055
- },
1056
- {
1057
- "epoch": 1.03,
1058
- "learning_rate": 0.0,
1059
- "loss": 0.6883,
1060
- "step": 17400
1061
- },
1062
- {
1063
- "epoch": 1.03,
1064
- "learning_rate": 0.0,
1065
- "loss": 0.5119,
1066
- "step": 17500
1067
- },
1068
- {
1069
- "epoch": 1.04,
1070
- "learning_rate": 0.0,
1071
- "loss": 0.6272,
1072
- "step": 17600
1073
- },
1074
- {
1075
- "epoch": 1.04,
1076
- "learning_rate": 0.0,
1077
- "loss": 0.6702,
1078
- "step": 17700
1079
- },
1080
- {
1081
- "epoch": 1.05,
1082
- "learning_rate": 0.0,
1083
- "loss": 0.5888,
1084
- "step": 17800
1085
- },
1086
- {
1087
- "epoch": 1.06,
1088
- "learning_rate": 0.0,
1089
- "loss": 0.6151,
1090
- "step": 17900
1091
- },
1092
- {
1093
- "epoch": 1.06,
1094
- "learning_rate": 0.0,
1095
- "loss": 0.6112,
1096
- "step": 18000
1097
- },
1098
- {
1099
- "epoch": 1.07,
1100
- "learning_rate": 0.0,
1101
- "loss": 0.6773,
1102
- "step": 18100
1103
- },
1104
- {
1105
- "epoch": 1.07,
1106
- "learning_rate": 0.0,
1107
- "loss": 0.6033,
1108
- "step": 18200
1109
- },
1110
- {
1111
- "epoch": 1.08,
1112
- "learning_rate": 0.0,
1113
- "loss": 0.6437,
1114
- "step": 18300
1115
- },
1116
- {
1117
- "epoch": 1.09,
1118
- "learning_rate": 0.0,
1119
- "loss": 0.63,
1120
- "step": 18400
1121
- },
1122
- {
1123
- "epoch": 1.09,
1124
- "learning_rate": 0.0,
1125
- "loss": 0.6846,
1126
- "step": 18500
1127
- },
1128
- {
1129
- "epoch": 1.1,
1130
- "learning_rate": 0.0,
1131
- "loss": 0.6727,
1132
- "step": 18600
1133
- },
1134
- {
1135
- "epoch": 1.1,
1136
- "learning_rate": 0.0,
1137
- "loss": 0.644,
1138
- "step": 18700
1139
- },
1140
- {
1141
- "epoch": 1.11,
1142
- "learning_rate": 0.0,
1143
- "loss": 0.5076,
1144
- "step": 18800
1145
- },
1146
- {
1147
- "epoch": 1.11,
1148
- "learning_rate": 0.0,
1149
- "loss": 0.6787,
1150
- "step": 18900
1151
- },
1152
- {
1153
- "epoch": 1.12,
1154
- "learning_rate": 0.0,
1155
- "loss": 0.6698,
1156
- "step": 19000
1157
- },
1158
- {
1159
- "epoch": 1.13,
1160
- "learning_rate": 0.0,
1161
- "loss": 0.6306,
1162
- "step": 19100
1163
- },
1164
- {
1165
- "epoch": 1.13,
1166
- "learning_rate": 0.0,
1167
- "loss": 0.618,
1168
- "step": 19200
1169
- },
1170
- {
1171
- "epoch": 1.14,
1172
- "learning_rate": 0.0,
1173
- "loss": 0.6077,
1174
- "step": 19300
1175
- },
1176
- {
1177
- "epoch": 1.14,
1178
- "learning_rate": 0.0,
1179
- "loss": 0.671,
1180
- "step": 19400
1181
- },
1182
- {
1183
- "epoch": 1.15,
1184
- "learning_rate": 0.0,
1185
- "loss": 0.564,
1186
- "step": 19500
1187
- },
1188
- {
1189
- "epoch": 1.16,
1190
- "learning_rate": 0.0,
1191
- "loss": 0.6521,
1192
- "step": 19600
1193
- },
1194
- {
1195
- "epoch": 1.16,
1196
- "learning_rate": 0.0,
1197
- "loss": 0.6126,
1198
- "step": 19700
1199
- },
1200
- {
1201
- "epoch": 1.17,
1202
- "learning_rate": 0.0,
1203
- "loss": 0.6015,
1204
- "step": 19800
1205
- },
1206
- {
1207
- "epoch": 1.17,
1208
- "learning_rate": 0.0,
1209
- "loss": 0.5994,
1210
- "step": 19900
1211
- },
1212
- {
1213
- "epoch": 1.18,
1214
- "learning_rate": 0.0,
1215
- "loss": 0.6969,
1216
- "step": 20000
1217
- },
1218
- {
1219
- "epoch": 1.18,
1220
- "eval_loss": 0.8651230335235596,
1221
- "eval_runtime": 547.573,
1222
- "eval_samples_per_second": 29.108,
1223
- "eval_steps_per_second": 1.821,
1224
- "step": 20000
1225
- },
1226
- {
1227
- "epoch": 1.19,
1228
- "learning_rate": 0.0,
1229
- "loss": 0.6145,
1230
- "step": 20100
1231
- },
1232
- {
1233
- "epoch": 1.19,
1234
- "learning_rate": 0.0,
1235
- "loss": 0.5923,
1236
- "step": 20200
1237
- },
1238
- {
1239
- "epoch": 1.2,
1240
- "learning_rate": 0.0,
1241
- "loss": 0.5869,
1242
- "step": 20300
1243
- },
1244
- {
1245
- "epoch": 1.2,
1246
- "learning_rate": 0.0,
1247
- "loss": 0.6092,
1248
- "step": 20400
1249
- },
1250
- {
1251
- "epoch": 1.21,
1252
- "learning_rate": 0.0,
1253
- "loss": 0.5804,
1254
- "step": 20500
1255
- },
1256
- {
1257
- "epoch": 1.22,
1258
- "learning_rate": 0.0,
1259
- "loss": 0.5638,
1260
- "step": 20600
1261
- },
1262
- {
1263
- "epoch": 1.22,
1264
- "learning_rate": 0.0,
1265
- "loss": 0.6938,
1266
- "step": 20700
1267
- },
1268
- {
1269
- "epoch": 1.23,
1270
- "learning_rate": 0.0,
1271
- "loss": 0.6669,
1272
- "step": 20800
1273
- },
1274
- {
1275
- "epoch": 1.23,
1276
- "learning_rate": 0.0,
1277
- "loss": 0.649,
1278
- "step": 20900
1279
- },
1280
- {
1281
- "epoch": 1.24,
1282
- "learning_rate": 0.0,
1283
- "loss": 0.6147,
1284
- "step": 21000
1285
- },
1286
- {
1287
- "epoch": 1.24,
1288
- "learning_rate": 0.0,
1289
- "loss": 0.63,
1290
- "step": 21100
1291
- },
1292
- {
1293
- "epoch": 1.25,
1294
- "learning_rate": 0.0,
1295
- "loss": 0.6312,
1296
- "step": 21200
1297
- },
1298
- {
1299
- "epoch": 1.26,
1300
- "learning_rate": 0.0,
1301
- "loss": 0.6575,
1302
- "step": 21300
1303
- },
1304
- {
1305
- "epoch": 1.26,
1306
- "learning_rate": 0.0,
1307
- "loss": 0.5937,
1308
- "step": 21400
1309
- },
1310
- {
1311
- "epoch": 1.27,
1312
- "learning_rate": 0.0,
1313
- "loss": 0.558,
1314
- "step": 21500
1315
- },
1316
- {
1317
- "epoch": 1.27,
1318
- "learning_rate": 0.0,
1319
- "loss": 0.6016,
1320
- "step": 21600
1321
- },
1322
- {
1323
- "epoch": 1.28,
1324
- "learning_rate": 0.0,
1325
- "loss": 0.6193,
1326
- "step": 21700
1327
- },
1328
- {
1329
- "epoch": 1.29,
1330
- "learning_rate": 0.0,
1331
- "loss": 0.6588,
1332
- "step": 21800
1333
- },
1334
- {
1335
- "epoch": 1.29,
1336
- "learning_rate": 0.0,
1337
- "loss": 0.614,
1338
- "step": 21900
1339
- },
1340
- {
1341
- "epoch": 1.3,
1342
- "learning_rate": 0.0,
1343
- "loss": 0.6077,
1344
- "step": 22000
1345
- },
1346
- {
1347
- "epoch": 1.3,
1348
- "learning_rate": 0.0,
1349
- "loss": 0.6539,
1350
- "step": 22100
1351
- },
1352
- {
1353
- "epoch": 1.31,
1354
- "learning_rate": 0.0,
1355
- "loss": 0.601,
1356
- "step": 22200
1357
- },
1358
- {
1359
- "epoch": 1.32,
1360
- "learning_rate": 0.0,
1361
- "loss": 0.6089,
1362
- "step": 22300
1363
- },
1364
- {
1365
- "epoch": 1.32,
1366
- "learning_rate": 0.0,
1367
- "loss": 0.5776,
1368
- "step": 22400
1369
- },
1370
- {
1371
- "epoch": 1.33,
1372
- "learning_rate": 0.0,
1373
- "loss": 0.5722,
1374
- "step": 22500
1375
- },
1376
- {
1377
- "epoch": 1.33,
1378
- "learning_rate": 0.0,
1379
- "loss": 0.5994,
1380
- "step": 22600
1381
- },
1382
- {
1383
- "epoch": 1.34,
1384
- "learning_rate": 0.0,
1385
- "loss": 0.6018,
1386
- "step": 22700
1387
- },
1388
- {
1389
- "epoch": 1.34,
1390
- "learning_rate": 0.0,
1391
- "loss": 0.6058,
1392
- "step": 22800
1393
- },
1394
- {
1395
- "epoch": 1.35,
1396
- "learning_rate": 0.0,
1397
- "loss": 0.5918,
1398
- "step": 22900
1399
- },
1400
- {
1401
- "epoch": 1.36,
1402
- "learning_rate": 0.0,
1403
- "loss": 0.5507,
1404
- "step": 23000
1405
- },
1406
- {
1407
- "epoch": 1.36,
1408
- "learning_rate": 0.0,
1409
- "loss": 0.5954,
1410
- "step": 23100
1411
- },
1412
- {
1413
- "epoch": 1.37,
1414
- "learning_rate": 0.0,
1415
- "loss": 0.6531,
1416
- "step": 23200
1417
- },
1418
- {
1419
- "epoch": 1.37,
1420
- "learning_rate": 0.0,
1421
- "loss": 0.6062,
1422
- "step": 23300
1423
- },
1424
- {
1425
- "epoch": 1.38,
1426
- "learning_rate": 0.0,
1427
- "loss": 0.6198,
1428
- "step": 23400
1429
- },
1430
- {
1431
- "epoch": 1.39,
1432
- "learning_rate": 0.0,
1433
- "loss": 0.5781,
1434
- "step": 23500
1435
- },
1436
- {
1437
- "epoch": 1.39,
1438
- "learning_rate": 0.0,
1439
- "loss": 0.6284,
1440
- "step": 23600
1441
- },
1442
- {
1443
- "epoch": 1.4,
1444
- "learning_rate": 0.0,
1445
- "loss": 0.658,
1446
- "step": 23700
1447
- },
1448
- {
1449
- "epoch": 1.4,
1450
- "learning_rate": 0.0,
1451
- "loss": 0.6053,
1452
- "step": 23800
1453
- },
1454
- {
1455
- "epoch": 1.41,
1456
- "learning_rate": 0.0,
1457
- "loss": 0.5701,
1458
- "step": 23900
1459
- },
1460
- {
1461
- "epoch": 1.42,
1462
- "learning_rate": 0.0,
1463
- "loss": 0.6044,
1464
- "step": 24000
1465
- },
1466
- {
1467
- "epoch": 1.42,
1468
- "learning_rate": 0.0,
1469
- "loss": 0.7357,
1470
- "step": 24100
1471
- },
1472
- {
1473
- "epoch": 1.43,
1474
- "learning_rate": 0.0,
1475
- "loss": 0.6144,
1476
- "step": 24200
1477
- },
1478
- {
1479
- "epoch": 1.43,
1480
- "learning_rate": 0.0,
1481
- "loss": 0.6445,
1482
- "step": 24300
1483
- },
1484
- {
1485
- "epoch": 1.44,
1486
- "learning_rate": 0.0,
1487
- "loss": 0.6284,
1488
- "step": 24400
1489
- },
1490
- {
1491
- "epoch": 1.45,
1492
- "learning_rate": 0.0,
1493
- "loss": 0.7069,
1494
- "step": 24500
1495
- },
1496
- {
1497
- "epoch": 1.45,
1498
- "learning_rate": 0.0,
1499
- "loss": 0.5919,
1500
- "step": 24600
1501
- },
1502
- {
1503
- "epoch": 1.46,
1504
- "learning_rate": 0.0,
1505
- "loss": 0.6415,
1506
- "step": 24700
1507
- },
1508
- {
1509
- "epoch": 1.46,
1510
- "learning_rate": 0.0,
1511
- "loss": 0.5846,
1512
- "step": 24800
1513
- },
1514
- {
1515
- "epoch": 1.47,
1516
- "learning_rate": 0.0,
1517
- "loss": 0.6158,
1518
- "step": 24900
1519
- },
1520
- {
1521
- "epoch": 1.47,
1522
- "learning_rate": 0.0,
1523
- "loss": 0.5761,
1524
- "step": 25000
1525
- },
1526
- {
1527
- "epoch": 1.48,
1528
- "learning_rate": 0.0,
1529
- "loss": 0.6457,
1530
- "step": 25100
1531
- },
1532
- {
1533
- "epoch": 1.49,
1534
- "learning_rate": 0.0,
1535
- "loss": 0.5787,
1536
- "step": 25200
1537
- },
1538
- {
1539
- "epoch": 1.49,
1540
- "learning_rate": 0.0,
1541
- "loss": 0.5936,
1542
- "step": 25300
1543
- },
1544
- {
1545
- "epoch": 1.5,
1546
- "learning_rate": 0.0,
1547
- "loss": 0.612,
1548
- "step": 25400
1549
- },
1550
- {
1551
- "epoch": 1.5,
1552
- "learning_rate": 0.0,
1553
- "loss": 0.5571,
1554
- "step": 25500
1555
- },
1556
- {
1557
- "epoch": 1.51,
1558
- "learning_rate": 0.0,
1559
- "loss": 0.5996,
1560
- "step": 25600
1561
- },
1562
- {
1563
- "epoch": 1.52,
1564
- "learning_rate": 0.0,
1565
- "loss": 0.6613,
1566
- "step": 25700
1567
- },
1568
- {
1569
- "epoch": 1.52,
1570
- "learning_rate": 0.0,
1571
- "loss": 0.5909,
1572
- "step": 25800
1573
- },
1574
- {
1575
- "epoch": 1.53,
1576
- "learning_rate": 0.0,
1577
- "loss": 0.6552,
1578
- "step": 25900
1579
- },
1580
- {
1581
- "epoch": 1.53,
1582
- "learning_rate": 0.0,
1583
- "loss": 0.6063,
1584
- "step": 26000
1585
- },
1586
- {
1587
- "epoch": 1.54,
1588
- "learning_rate": 0.0,
1589
- "loss": 0.7092,
1590
- "step": 26100
1591
- },
1592
- {
1593
- "epoch": 1.55,
1594
- "learning_rate": 0.0,
1595
- "loss": 0.6052,
1596
- "step": 26200
1597
- },
1598
- {
1599
- "epoch": 1.55,
1600
- "learning_rate": 0.0,
1601
- "loss": 0.6585,
1602
- "step": 26300
1603
- },
1604
- {
1605
- "epoch": 1.56,
1606
- "learning_rate": 0.0,
1607
- "loss": 0.578,
1608
- "step": 26400
1609
- },
1610
- {
1611
- "epoch": 1.56,
1612
- "learning_rate": 0.0,
1613
- "loss": 0.5763,
1614
- "step": 26500
1615
- },
1616
- {
1617
- "epoch": 1.57,
1618
- "learning_rate": 0.0,
1619
- "loss": 0.6718,
1620
- "step": 26600
1621
- },
1622
- {
1623
- "epoch": 1.57,
1624
- "learning_rate": 0.0,
1625
- "loss": 0.6693,
1626
- "step": 26700
1627
- },
1628
- {
1629
- "epoch": 1.58,
1630
- "learning_rate": 0.0,
1631
- "loss": 0.6546,
1632
- "step": 26800
1633
- },
1634
- {
1635
- "epoch": 1.59,
1636
- "learning_rate": 0.0,
1637
- "loss": 0.6595,
1638
- "step": 26900
1639
- },
1640
- {
1641
- "epoch": 1.59,
1642
- "learning_rate": 0.0,
1643
- "loss": 0.6535,
1644
- "step": 27000
1645
- },
1646
- {
1647
- "epoch": 1.6,
1648
- "learning_rate": 0.0,
1649
- "loss": 0.6744,
1650
- "step": 27100
1651
- },
1652
- {
1653
- "epoch": 1.6,
1654
- "learning_rate": 0.0,
1655
- "loss": 0.613,
1656
- "step": 27200
1657
- },
1658
- {
1659
- "epoch": 1.61,
1660
- "learning_rate": 0.0,
1661
- "loss": 0.5885,
1662
- "step": 27300
1663
- },
1664
- {
1665
- "epoch": 1.62,
1666
- "learning_rate": 0.0,
1667
- "loss": 0.6103,
1668
- "step": 27400
1669
- },
1670
- {
1671
- "epoch": 1.62,
1672
- "learning_rate": 0.0,
1673
- "loss": 0.6499,
1674
- "step": 27500
1675
- },
1676
- {
1677
- "epoch": 1.63,
1678
- "learning_rate": 0.0,
1679
- "loss": 0.6648,
1680
- "step": 27600
1681
- },
1682
- {
1683
- "epoch": 1.63,
1684
- "learning_rate": 0.0,
1685
- "loss": 0.6292,
1686
- "step": 27700
1687
- },
1688
- {
1689
- "epoch": 1.64,
1690
- "learning_rate": 0.0,
1691
- "loss": 0.5542,
1692
- "step": 27800
1693
- },
1694
- {
1695
- "epoch": 1.65,
1696
- "learning_rate": 0.0,
1697
- "loss": 0.5549,
1698
- "step": 27900
1699
- },
1700
- {
1701
- "epoch": 1.65,
1702
- "learning_rate": 0.0,
1703
- "loss": 0.6058,
1704
- "step": 28000
1705
- },
1706
- {
1707
- "epoch": 1.66,
1708
- "learning_rate": 0.0,
1709
- "loss": 0.565,
1710
- "step": 28100
1711
- },
1712
- {
1713
- "epoch": 1.66,
1714
- "learning_rate": 0.0,
1715
- "loss": 0.7065,
1716
- "step": 28200
1717
- },
1718
- {
1719
- "epoch": 1.67,
1720
- "learning_rate": 0.0,
1721
- "loss": 0.6621,
1722
- "step": 28300
1723
- },
1724
- {
1725
- "epoch": 1.68,
1726
- "learning_rate": 0.0,
1727
- "loss": 0.6052,
1728
- "step": 28400
1729
- },
1730
- {
1731
- "epoch": 1.68,
1732
- "learning_rate": 0.0,
1733
- "loss": 0.6506,
1734
- "step": 28500
1735
- },
1736
- {
1737
- "epoch": 1.69,
1738
- "learning_rate": 0.0,
1739
- "loss": 0.7008,
1740
- "step": 28600
1741
- },
1742
- {
1743
- "epoch": 1.69,
1744
- "learning_rate": 0.0,
1745
- "loss": 0.6847,
1746
- "step": 28700
1747
- },
1748
- {
1749
- "epoch": 1.7,
1750
- "learning_rate": 0.0,
1751
- "loss": 0.6591,
1752
- "step": 28800
1753
- },
1754
- {
1755
- "epoch": 1.7,
1756
- "learning_rate": 0.0,
1757
- "loss": 0.6289,
1758
- "step": 28900
1759
- },
1760
- {
1761
- "epoch": 1.71,
1762
- "learning_rate": 0.0,
1763
- "loss": 0.5949,
1764
- "step": 29000
1765
- },
1766
- {
1767
- "epoch": 1.72,
1768
- "learning_rate": 0.0,
1769
- "loss": 0.6497,
1770
- "step": 29100
1771
- },
1772
- {
1773
- "epoch": 1.72,
1774
- "learning_rate": 0.0,
1775
- "loss": 0.5947,
1776
- "step": 29200
1777
- },
1778
- {
1779
- "epoch": 1.73,
1780
- "learning_rate": 0.0,
1781
- "loss": 0.6483,
1782
- "step": 29300
1783
- },
1784
- {
1785
- "epoch": 1.73,
1786
- "learning_rate": 0.0,
1787
- "loss": 0.5743,
1788
- "step": 29400
1789
- },
1790
- {
1791
- "epoch": 1.74,
1792
- "learning_rate": 0.0,
1793
- "loss": 0.6876,
1794
- "step": 29500
1795
- },
1796
- {
1797
- "epoch": 1.75,
1798
- "learning_rate": 0.0,
1799
- "loss": 0.6503,
1800
- "step": 29600
1801
- },
1802
- {
1803
- "epoch": 1.75,
1804
- "learning_rate": 0.0,
1805
- "loss": 0.5963,
1806
- "step": 29700
1807
- },
1808
- {
1809
- "epoch": 1.76,
1810
- "learning_rate": 0.0,
1811
- "loss": 0.6361,
1812
- "step": 29800
1813
- },
1814
- {
1815
- "epoch": 1.76,
1816
- "learning_rate": 0.0,
1817
- "loss": 0.6624,
1818
- "step": 29900
1819
- },
1820
- {
1821
- "epoch": 1.77,
1822
- "learning_rate": 0.0,
1823
- "loss": 0.6298,
1824
- "step": 30000
1825
- },
1826
- {
1827
- "epoch": 1.77,
1828
- "eval_loss": 0.8651230335235596,
1829
- "eval_runtime": 546.7955,
1830
- "eval_samples_per_second": 29.15,
1831
- "eval_steps_per_second": 1.823,
1832
- "step": 30000
1833
- }
1834
- ],
1835
- "max_steps": 30516,
1836
- "num_train_epochs": 2,
1837
- "total_flos": 1.8284013786415104e+17,
1838
- "trial_name": null,
1839
- "trial_params": null
1840
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
last-checkpoint/training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:50d8747e43721b80b6e9bd368539717424fbb7049d6db13269fc09fffe501769
3
- size 3311
 
 
 
 
last-checkpoint/vocab.json DELETED
The diff for this file is too large to render. See raw diff
 
runs/Dec05_03-30-37_fbdce2302f52/events.out.tfevents.1670211053.fbdce2302f52.24.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:89b673d9ccde84e81880a851f0f5abf58604d05b17b5b045206afb9d74ba26fb
3
- size 57737
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9eb4b7c9e109f34ec0333f8271b93e1fa7dc0ab76432310a1664e4c84ebde19f
3
+ size 58897