lIlBrother commited on
Commit
d526a62
1 Parent(s): acb6b59

Update: bleu Score 로 선택된 모델로 변경

Browse files
Files changed (4) hide show
  1. all_results.json +15 -15
  2. pytorch_model.bin +2 -2
  3. trainer_state.json +0 -1576
  4. training_args.bin +2 -2
all_results.json CHANGED
@@ -1,20 +1,20 @@
1
  {
2
  "epoch": 5.13,
3
- "eval_bleu": 0.9529006548919251,
4
  "eval_brevity_penalty": 1.0,
5
- "eval_length_ratio": 1.0129792088807863,
6
- "eval_loss": 0.040760207921266556,
7
  "eval_reference_length": 688948,
8
- "eval_rouge1": 0.9693520563208838,
9
- "eval_rouge2": 0.9444220599246154,
10
- "eval_rougeL": 0.9692485601662657,
11
- "eval_rougeLsum": 0.9692422603343052,
12
- "eval_runtime": 348.5598,
13
- "eval_samples_per_second": 122.995,
14
- "eval_steps_per_second": 10.251,
15
- "eval_translation_length": 697890,
16
- "train_loss": 0.08529333166642622,
17
- "train_runtime": 78966.8059,
18
- "train_samples_per_second": 25.074,
19
- "train_steps_per_second": 4.179
20
  }
 
1
  {
2
  "epoch": 5.13,
3
+ "eval_bleu": 0.958234790096092,
4
  "eval_brevity_penalty": 1.0,
5
+ "eval_length_ratio": 1.0078162647979239,
6
+ "eval_loss": 0.04153120145201683,
7
  "eval_reference_length": 688948,
8
+ "eval_rouge1": 0.9735361877162854,
9
+ "eval_rouge2": 0.9493975212378124,
10
+ "eval_rougeL": 0.9734558938864928,
11
+ "eval_rougeLsum": 0.9734350757552404,
12
+ "eval_runtime": 375.3753,
13
+ "eval_samples_per_second": 114.208,
14
+ "eval_steps_per_second": 9.518,
15
+ "eval_translation_length": 694333,
16
+ "train_loss": 0.08534714673360189,
17
+ "train_runtime": 94029.0071,
18
+ "train_samples_per_second": 21.057,
19
+ "train_steps_per_second": 3.51
20
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:002675439d3f8ffec4c9410dcb0992e50f1db850b5e49a2e207ec30b08c86997
3
- size 495646265
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7609dd43a5af3140dae1306e904c7a022c492bbbac9478a91e69604054d18e31
3
+ size 495648413
trainer_state.json DELETED
@@ -1,1576 +0,0 @@
1
- {
2
- "best_metric": 0.040760207921266556,
3
- "best_model_checkpoint": "/data2/bart/temp_workspace/nlp/output_dir/checkpoint-250000",
4
- "epoch": 5.131554394476582,
5
- "global_step": 330000,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.03,
12
- "learning_rate": 2.844285714285714e-07,
13
- "loss": 2.9266,
14
- "step": 2000
15
- },
16
- {
17
- "epoch": 0.06,
18
- "learning_rate": 5.701428571428572e-07,
19
- "loss": 0.9726,
20
- "step": 4000
21
- },
22
- {
23
- "epoch": 0.09,
24
- "learning_rate": 8.558571428571428e-07,
25
- "loss": 0.5831,
26
- "step": 6000
27
- },
28
- {
29
- "epoch": 0.12,
30
- "learning_rate": 1.1415714285714287e-06,
31
- "loss": 0.4407,
32
- "step": 8000
33
- },
34
- {
35
- "epoch": 0.16,
36
- "learning_rate": 1.4272857142857143e-06,
37
- "loss": 0.361,
38
- "step": 10000
39
- },
40
- {
41
- "epoch": 0.16,
42
- "eval_bleu": 0.7861044602086867,
43
- "eval_brevity_penalty": 1.0,
44
- "eval_length_ratio": 1.0976053345100065,
45
- "eval_loss": 0.24034307897090912,
46
- "eval_reference_length": 688948,
47
- "eval_rouge1": 0.8478815702507803,
48
- "eval_rouge2": 0.770393954032274,
49
- "eval_rougeL": 0.8471496607249986,
50
- "eval_rougeLsum": 0.8471595113950809,
51
- "eval_runtime": 357.1063,
52
- "eval_samples_per_second": 120.051,
53
- "eval_steps_per_second": 10.005,
54
- "eval_translation_length": 756193,
55
- "step": 10000
56
- },
57
- {
58
- "epoch": 0.19,
59
- "learning_rate": 1.712857142857143e-06,
60
- "loss": 0.2936,
61
- "step": 12000
62
- },
63
- {
64
- "epoch": 0.22,
65
- "learning_rate": 1.9985714285714287e-06,
66
- "loss": 0.255,
67
- "step": 14000
68
- },
69
- {
70
- "epoch": 0.25,
71
- "learning_rate": 2.2841428571428574e-06,
72
- "loss": 0.2187,
73
- "step": 16000
74
- },
75
- {
76
- "epoch": 0.28,
77
- "learning_rate": 2.569714285714286e-06,
78
- "loss": 0.1976,
79
- "step": 18000
80
- },
81
- {
82
- "epoch": 0.31,
83
- "learning_rate": 2.8552857142857144e-06,
84
- "loss": 0.1795,
85
- "step": 20000
86
- },
87
- {
88
- "epoch": 0.31,
89
- "eval_bleu": 0.8725717687820272,
90
- "eval_brevity_penalty": 1.0,
91
- "eval_length_ratio": 1.0512491508793116,
92
- "eval_loss": 0.12307040393352509,
93
- "eval_reference_length": 688948,
94
- "eval_rouge1": 0.9065960080443258,
95
- "eval_rouge2": 0.8543292052866391,
96
- "eval_rougeL": 0.9063849761037827,
97
- "eval_rougeLsum": 0.9063953832846721,
98
- "eval_runtime": 357.2954,
99
- "eval_samples_per_second": 119.988,
100
- "eval_steps_per_second": 10.0,
101
- "eval_translation_length": 724256,
102
- "step": 20000
103
- },
104
- {
105
- "epoch": 0.34,
106
- "learning_rate": 3.140857142857143e-06,
107
- "loss": 0.1669,
108
- "step": 22000
109
- },
110
- {
111
- "epoch": 0.37,
112
- "learning_rate": 3.4264285714285715e-06,
113
- "loss": 0.151,
114
- "step": 24000
115
- },
116
- {
117
- "epoch": 0.4,
118
- "learning_rate": 3.7121428571428575e-06,
119
- "loss": 0.1387,
120
- "step": 26000
121
- },
122
- {
123
- "epoch": 0.44,
124
- "learning_rate": 3.997714285714286e-06,
125
- "loss": 0.1354,
126
- "step": 28000
127
- },
128
- {
129
- "epoch": 0.47,
130
- "learning_rate": 4.2834285714285715e-06,
131
- "loss": 0.1244,
132
- "step": 30000
133
- },
134
- {
135
- "epoch": 0.47,
136
- "eval_bleu": 0.8899834257776226,
137
- "eval_brevity_penalty": 1.0,
138
- "eval_length_ratio": 1.050031352148493,
139
- "eval_loss": 0.09066224843263626,
140
- "eval_reference_length": 688948,
141
- "eval_rouge1": 0.9246723733764061,
142
- "eval_rouge2": 0.8828517220657879,
143
- "eval_rougeL": 0.9244225429282074,
144
- "eval_rougeLsum": 0.9244091547844906,
145
- "eval_runtime": 347.7759,
146
- "eval_samples_per_second": 123.272,
147
- "eval_steps_per_second": 10.274,
148
- "eval_translation_length": 723417,
149
- "step": 30000
150
- },
151
- {
152
- "epoch": 0.5,
153
- "learning_rate": 4.569e-06,
154
- "loss": 0.1206,
155
- "step": 32000
156
- },
157
- {
158
- "epoch": 0.53,
159
- "learning_rate": 4.854714285714286e-06,
160
- "loss": 0.1136,
161
- "step": 34000
162
- },
163
- {
164
- "epoch": 0.56,
165
- "learning_rate": 5.140285714285715e-06,
166
- "loss": 0.1128,
167
- "step": 36000
168
- },
169
- {
170
- "epoch": 0.59,
171
- "learning_rate": 5.426e-06,
172
- "loss": 0.1048,
173
- "step": 38000
174
- },
175
- {
176
- "epoch": 0.62,
177
- "learning_rate": 5.7115714285714285e-06,
178
- "loss": 0.1012,
179
- "step": 40000
180
- },
181
- {
182
- "epoch": 0.62,
183
- "eval_bleu": 0.9208253246491475,
184
- "eval_brevity_penalty": 1.0,
185
- "eval_length_ratio": 1.0275782787670478,
186
- "eval_loss": 0.07363951206207275,
187
- "eval_reference_length": 688948,
188
- "eval_rouge1": 0.9455229652203183,
189
- "eval_rouge2": 0.9101496822244501,
190
- "eval_rougeL": 0.9454059770773995,
191
- "eval_rougeLsum": 0.9454056293148791,
192
- "eval_runtime": 359.949,
193
- "eval_samples_per_second": 119.103,
194
- "eval_steps_per_second": 9.926,
195
- "eval_translation_length": 707948,
196
- "step": 40000
197
- },
198
- {
199
- "epoch": 0.65,
200
- "learning_rate": 5.997285714285714e-06,
201
- "loss": 0.0962,
202
- "step": 42000
203
- },
204
- {
205
- "epoch": 0.68,
206
- "learning_rate": 6.282571428571429e-06,
207
- "loss": 0.0964,
208
- "step": 44000
209
- },
210
- {
211
- "epoch": 0.72,
212
- "learning_rate": 6.568285714285715e-06,
213
- "loss": 0.095,
214
- "step": 46000
215
- },
216
- {
217
- "epoch": 0.75,
218
- "learning_rate": 6.8538571428571434e-06,
219
- "loss": 0.0913,
220
- "step": 48000
221
- },
222
- {
223
- "epoch": 0.78,
224
- "learning_rate": 7.139428571428572e-06,
225
- "loss": 0.0906,
226
- "step": 50000
227
- },
228
- {
229
- "epoch": 0.78,
230
- "eval_bleu": 0.9072012805362628,
231
- "eval_brevity_penalty": 1.0,
232
- "eval_length_ratio": 1.0453851379204235,
233
- "eval_loss": 0.06683389842510223,
234
- "eval_reference_length": 688948,
235
- "eval_rouge1": 0.936913196858357,
236
- "eval_rouge2": 0.9021025087676598,
237
- "eval_rougeL": 0.9368038216303922,
238
- "eval_rougeLsum": 0.936786757645766,
239
- "eval_runtime": 350.5992,
240
- "eval_samples_per_second": 122.279,
241
- "eval_steps_per_second": 10.191,
242
- "eval_translation_length": 720216,
243
- "step": 50000
244
- },
245
- {
246
- "epoch": 0.81,
247
- "learning_rate": 7.425142857142857e-06,
248
- "loss": 0.0881,
249
- "step": 52000
250
- },
251
- {
252
- "epoch": 0.84,
253
- "learning_rate": 7.710714285714287e-06,
254
- "loss": 0.088,
255
- "step": 54000
256
- },
257
- {
258
- "epoch": 0.87,
259
- "learning_rate": 7.996428571428572e-06,
260
- "loss": 0.0852,
261
- "step": 56000
262
- },
263
- {
264
- "epoch": 0.9,
265
- "learning_rate": 8.281857142857143e-06,
266
- "loss": 0.0836,
267
- "step": 58000
268
- },
269
- {
270
- "epoch": 0.93,
271
- "learning_rate": 8.567428571428572e-06,
272
- "loss": 0.0837,
273
- "step": 60000
274
- },
275
- {
276
- "epoch": 0.93,
277
- "eval_bleu": 0.9378929040812776,
278
- "eval_brevity_penalty": 1.0,
279
- "eval_length_ratio": 1.0159213757787235,
280
- "eval_loss": 0.06079160049557686,
281
- "eval_reference_length": 688948,
282
- "eval_rouge1": 0.9552455374973553,
283
- "eval_rouge2": 0.9234403473524968,
284
- "eval_rougeL": 0.9552044295582562,
285
- "eval_rougeLsum": 0.9551288128727483,
286
- "eval_runtime": 353.7894,
287
- "eval_samples_per_second": 121.177,
288
- "eval_steps_per_second": 10.099,
289
- "eval_translation_length": 699917,
290
- "step": 60000
291
- },
292
- {
293
- "epoch": 0.96,
294
- "learning_rate": 8.853142857142858e-06,
295
- "loss": 0.0839,
296
- "step": 62000
297
- },
298
- {
299
- "epoch": 1.0,
300
- "learning_rate": 9.138571428571429e-06,
301
- "loss": 0.0815,
302
- "step": 64000
303
- },
304
- {
305
- "epoch": 1.03,
306
- "learning_rate": 9.424285714285715e-06,
307
- "loss": 0.0742,
308
- "step": 66000
309
- },
310
- {
311
- "epoch": 1.06,
312
- "learning_rate": 9.71e-06,
313
- "loss": 0.0683,
314
- "step": 68000
315
- },
316
- {
317
- "epoch": 1.09,
318
- "learning_rate": 9.995714285714286e-06,
319
- "loss": 0.067,
320
- "step": 70000
321
- },
322
- {
323
- "epoch": 1.09,
324
- "eval_bleu": 0.9320243750023054,
325
- "eval_brevity_penalty": 1.0,
326
- "eval_length_ratio": 1.0232078473266488,
327
- "eval_loss": 0.058820515871047974,
328
- "eval_reference_length": 688948,
329
- "eval_rouge1": 0.9532786206216067,
330
- "eval_rouge2": 0.9224384838543223,
331
- "eval_rougeL": 0.9531782599798171,
332
- "eval_rougeLsum": 0.9531407362403285,
333
- "eval_runtime": 346.6656,
334
- "eval_samples_per_second": 123.667,
335
- "eval_steps_per_second": 10.307,
336
- "eval_translation_length": 704937,
337
- "step": 70000
338
- },
339
- {
340
- "epoch": 1.12,
341
- "learning_rate": 9.998587848362676e-06,
342
- "loss": 0.0686,
343
- "step": 72000
344
- },
345
- {
346
- "epoch": 1.15,
347
- "learning_rate": 9.994257059669239e-06,
348
- "loss": 0.0718,
349
- "step": 74000
350
- },
351
- {
352
- "epoch": 1.18,
353
- "learning_rate": 9.987009765493164e-06,
354
- "loss": 0.0687,
355
- "step": 76000
356
- },
357
- {
358
- "epoch": 1.21,
359
- "learning_rate": 9.976861810464927e-06,
360
- "loss": 0.0691,
361
- "step": 78000
362
- },
363
- {
364
- "epoch": 1.24,
365
- "learning_rate": 9.963798805467935e-06,
366
- "loss": 0.0726,
367
- "step": 80000
368
- },
369
- {
370
- "epoch": 1.24,
371
- "eval_bleu": 0.931299618108372,
372
- "eval_brevity_penalty": 1.0,
373
- "eval_length_ratio": 1.02620517078212,
374
- "eval_loss": 0.05646243691444397,
375
- "eval_reference_length": 688948,
376
- "eval_rouge1": 0.9506828525475279,
377
- "eval_rouge2": 0.9195555913617979,
378
- "eval_rougeL": 0.9505222105099121,
379
- "eval_rougeLsum": 0.9505025253164432,
380
- "eval_runtime": 358.4748,
381
- "eval_samples_per_second": 119.593,
382
- "eval_steps_per_second": 9.967,
383
- "eval_translation_length": 707002,
384
- "step": 80000
385
- },
386
- {
387
- "epoch": 1.28,
388
- "learning_rate": 9.947845785448258e-06,
389
- "loss": 0.0672,
390
- "step": 82000
391
- },
392
- {
393
- "epoch": 1.31,
394
- "learning_rate": 9.929006251362937e-06,
395
- "loss": 0.0692,
396
- "step": 84000
397
- },
398
- {
399
- "epoch": 1.34,
400
- "learning_rate": 9.907279613454706e-06,
401
- "loss": 0.0682,
402
- "step": 86000
403
- },
404
- {
405
- "epoch": 1.37,
406
- "learning_rate": 9.882700272366437e-06,
407
- "loss": 0.065,
408
- "step": 88000
409
- },
410
- {
411
- "epoch": 1.4,
412
- "learning_rate": 9.855257991301662e-06,
413
- "loss": 0.0648,
414
- "step": 90000
415
- },
416
- {
417
- "epoch": 1.4,
418
- "eval_bleu": 0.9337440022630773,
419
- "eval_brevity_penalty": 1.0,
420
- "eval_length_ratio": 1.0237202227163733,
421
- "eval_loss": 0.054729994386434555,
422
- "eval_reference_length": 688948,
423
- "eval_rouge1": 0.9562602670349256,
424
- "eval_rouge2": 0.926727457591136,
425
- "eval_rougeL": 0.9561080754823015,
426
- "eval_rougeLsum": 0.9560774432065933,
427
- "eval_runtime": 347.9396,
428
- "eval_samples_per_second": 123.214,
429
- "eval_steps_per_second": 10.269,
430
- "eval_translation_length": 705290,
431
- "step": 90000
432
- },
433
- {
434
- "epoch": 1.43,
435
- "learning_rate": 9.824996220706527e-06,
436
- "loss": 0.0649,
437
- "step": 92000
438
- },
439
- {
440
- "epoch": 1.46,
441
- "learning_rate": 9.79190235716806e-06,
442
- "loss": 0.067,
443
- "step": 94000
444
- },
445
- {
446
- "epoch": 1.49,
447
- "learning_rate": 9.756028799505886e-06,
448
- "loss": 0.068,
449
- "step": 96000
450
- },
451
- {
452
- "epoch": 1.52,
453
- "learning_rate": 9.717380631513947e-06,
454
- "loss": 0.0609,
455
- "step": 98000
456
- },
457
- {
458
- "epoch": 1.56,
459
- "learning_rate": 9.675980400071303e-06,
460
- "loss": 0.0641,
461
- "step": 100000
462
- },
463
- {
464
- "epoch": 1.56,
465
- "eval_bleu": 0.9437228795566622,
466
- "eval_brevity_penalty": 1.0,
467
- "eval_length_ratio": 1.0158429954074908,
468
- "eval_loss": 0.05343015119433403,
469
- "eval_reference_length": 688948,
470
- "eval_rouge1": 0.9622428608075103,
471
- "eval_rouge2": 0.9342448516446199,
472
- "eval_rougeL": 0.9621362437715651,
473
- "eval_rougeLsum": 0.9621409491590999,
474
- "eval_runtime": 354.0159,
475
- "eval_samples_per_second": 121.099,
476
- "eval_steps_per_second": 10.093,
477
- "eval_translation_length": 699863,
478
- "step": 100000
479
- },
480
- {
481
- "epoch": 1.59,
482
- "learning_rate": 9.63182950403888e-06,
483
- "loss": 0.0625,
484
- "step": 102000
485
- },
486
- {
487
- "epoch": 1.62,
488
- "learning_rate": 9.58502194786445e-06,
489
- "loss": 0.0617,
490
- "step": 104000
491
- },
492
- {
493
- "epoch": 1.65,
494
- "learning_rate": 9.535491361586628e-06,
495
- "loss": 0.0637,
496
- "step": 106000
497
- },
498
- {
499
- "epoch": 1.68,
500
- "learning_rate": 9.483312176074826e-06,
501
- "loss": 0.0615,
502
- "step": 108000
503
- },
504
- {
505
- "epoch": 1.71,
506
- "learning_rate": 9.428542910099412e-06,
507
- "loss": 0.061,
508
- "step": 110000
509
- },
510
- {
511
- "epoch": 1.71,
512
- "eval_bleu": 0.9451141494751257,
513
- "eval_brevity_penalty": 1.0,
514
- "eval_length_ratio": 1.016459877958859,
515
- "eval_loss": 0.0512821264564991,
516
- "eval_reference_length": 688948,
517
- "eval_rouge1": 0.96485183292355,
518
- "eval_rouge2": 0.937858740593392,
519
- "eval_rougeL": 0.964753163918546,
520
- "eval_rougeLsum": 0.9647545868693599,
521
- "eval_runtime": 350.3148,
522
- "eval_samples_per_second": 122.379,
523
- "eval_steps_per_second": 10.199,
524
- "eval_translation_length": 700288,
525
- "step": 110000
526
- },
527
- {
528
- "epoch": 1.74,
529
- "learning_rate": 9.371219416280765e-06,
530
- "loss": 0.0647,
531
- "step": 112000
532
- },
533
- {
534
- "epoch": 1.77,
535
- "learning_rate": 9.311287160167118e-06,
536
- "loss": 0.0619,
537
- "step": 114000
538
- },
539
- {
540
- "epoch": 1.8,
541
- "learning_rate": 9.24886908181262e-06,
542
- "loss": 0.0634,
543
- "step": 116000
544
- },
545
- {
546
- "epoch": 1.83,
547
- "learning_rate": 9.18393918535506e-06,
548
- "loss": 0.058,
549
- "step": 118000
550
- },
551
- {
552
- "epoch": 1.87,
553
- "learning_rate": 9.116565986234595e-06,
554
- "loss": 0.0616,
555
- "step": 120000
556
- },
557
- {
558
- "epoch": 1.87,
559
- "eval_bleu": 0.9353441526892242,
560
- "eval_brevity_penalty": 1.0,
561
- "eval_length_ratio": 1.0252529944204787,
562
- "eval_loss": 0.04904291033744812,
563
- "eval_reference_length": 688948,
564
- "eval_rouge1": 0.9578546390239036,
565
- "eval_rouge2": 0.9297894756880962,
566
- "eval_rougeL": 0.9577587648786969,
567
- "eval_rougeLsum": 0.957745334663435,
568
- "eval_runtime": 360.6437,
569
- "eval_samples_per_second": 118.874,
570
- "eval_steps_per_second": 9.907,
571
- "eval_translation_length": 706346,
572
- "step": 120000
573
- },
574
- {
575
- "epoch": 1.9,
576
- "learning_rate": 9.04689527434882e-06,
577
- "loss": 0.0588,
578
- "step": 122000
579
- },
580
- {
581
- "epoch": 1.93,
582
- "learning_rate": 8.974795071500655e-06,
583
- "loss": 0.0581,
584
- "step": 124000
585
- },
586
- {
587
- "epoch": 1.96,
588
- "learning_rate": 8.90033821949533e-06,
589
- "loss": 0.0571,
590
- "step": 126000
591
- },
592
- {
593
- "epoch": 1.99,
594
- "learning_rate": 8.823603679804848e-06,
595
- "loss": 0.0575,
596
- "step": 128000
597
- },
598
- {
599
- "epoch": 2.02,
600
- "learning_rate": 8.744676297261531e-06,
601
- "loss": 0.0501,
602
- "step": 130000
603
- },
604
- {
605
- "epoch": 2.02,
606
- "eval_bleu": 0.9391629851344814,
607
- "eval_brevity_penalty": 1.0,
608
- "eval_length_ratio": 1.0246375633574667,
609
- "eval_loss": 0.04428655281662941,
610
- "eval_reference_length": 688948,
611
- "eval_rouge1": 0.9620654622397073,
612
- "eval_rouge2": 0.9358278388686297,
613
- "eval_rougeL": 0.9619903188380114,
614
- "eval_rougeLsum": 0.9620056201386562,
615
- "eval_runtime": 350.9058,
616
- "eval_samples_per_second": 122.172,
617
- "eval_steps_per_second": 10.182,
618
- "eval_translation_length": 705922,
619
- "step": 130000
620
- },
621
- {
622
- "epoch": 2.05,
623
- "learning_rate": 8.663523200213197e-06,
624
- "loss": 0.0427,
625
- "step": 132000
626
- },
627
- {
628
- "epoch": 2.08,
629
- "learning_rate": 8.580230708782164e-06,
630
- "loss": 0.0437,
631
- "step": 134000
632
- },
633
- {
634
- "epoch": 2.11,
635
- "learning_rate": 8.494890669233825e-06,
636
- "loss": 0.0425,
637
- "step": 136000
638
- },
639
- {
640
- "epoch": 2.15,
641
- "learning_rate": 8.407555964848785e-06,
642
- "loss": 0.0419,
643
- "step": 138000
644
- },
645
- {
646
- "epoch": 2.18,
647
- "learning_rate": 8.318144933677256e-06,
648
- "loss": 0.0442,
649
- "step": 140000
650
- },
651
- {
652
- "epoch": 2.18,
653
- "eval_bleu": 0.9410527501908049,
654
- "eval_brevity_penalty": 1.0,
655
- "eval_length_ratio": 1.0222832492437746,
656
- "eval_loss": 0.04393425211310387,
657
- "eval_reference_length": 688948,
658
- "eval_rouge1": 0.9622283105302971,
659
- "eval_rouge2": 0.9363579229073962,
660
- "eval_rougeL": 0.9621524453412997,
661
- "eval_rougeLsum": 0.9621713486883273,
662
- "eval_runtime": 352.9817,
663
- "eval_samples_per_second": 121.454,
664
- "eval_steps_per_second": 10.122,
665
- "eval_translation_length": 704300,
666
- "step": 140000
667
- },
668
- {
669
- "epoch": 2.21,
670
- "learning_rate": 8.226796199304702e-06,
671
- "loss": 0.0445,
672
- "step": 142000
673
- },
674
- {
675
- "epoch": 2.24,
676
- "learning_rate": 8.13361018519741e-06,
677
- "loss": 0.0445,
678
- "step": 144000
679
- },
680
- {
681
- "epoch": 2.27,
682
- "learning_rate": 8.0386440390423e-06,
683
- "loss": 0.0421,
684
- "step": 146000
685
- },
686
- {
687
- "epoch": 2.3,
688
- "learning_rate": 7.941809254954647e-06,
689
- "loss": 0.0434,
690
- "step": 148000
691
- },
692
- {
693
- "epoch": 2.33,
694
- "learning_rate": 7.843256537104586e-06,
695
- "loss": 0.0418,
696
- "step": 150000
697
- },
698
- {
699
- "epoch": 2.33,
700
- "eval_bleu": 0.9452514516559248,
701
- "eval_brevity_penalty": 1.0,
702
- "eval_length_ratio": 1.0194572014143302,
703
- "eval_loss": 0.04334454983472824,
704
- "eval_reference_length": 688948,
705
- "eval_rouge1": 0.9667034100796513,
706
- "eval_rouge2": 0.9415234249395421,
707
- "eval_rougeL": 0.966596537202159,
708
- "eval_rougeLsum": 0.9665995305639754,
709
- "eval_runtime": 350.4381,
710
- "eval_samples_per_second": 122.335,
711
- "eval_steps_per_second": 10.196,
712
- "eval_translation_length": 702353,
713
- "step": 150000
714
- },
715
- {
716
- "epoch": 2.36,
717
- "learning_rate": 7.743043437504057e-06,
718
- "loss": 0.0437,
719
- "step": 152000
720
- },
721
- {
722
- "epoch": 2.39,
723
- "learning_rate": 7.641279775661868e-06,
724
- "loss": 0.0426,
725
- "step": 154000
726
- },
727
- {
728
- "epoch": 2.43,
729
- "learning_rate": 7.538027276115405e-06,
730
- "loss": 0.0412,
731
- "step": 156000
732
- },
733
- {
734
- "epoch": 2.46,
735
- "learning_rate": 7.433190045507044e-06,
736
- "loss": 0.0409,
737
- "step": 158000
738
- },
739
- {
740
- "epoch": 2.49,
741
- "learning_rate": 7.326931900431675e-06,
742
- "loss": 0.0419,
743
- "step": 160000
744
- },
745
- {
746
- "epoch": 2.49,
747
- "eval_bleu": 0.9542889216747796,
748
- "eval_brevity_penalty": 1.0,
749
- "eval_length_ratio": 1.010262022678054,
750
- "eval_loss": 0.04232573136687279,
751
- "eval_reference_length": 688948,
752
- "eval_rouge1": 0.9704801328109232,
753
- "eval_rouge2": 0.9456205175624834,
754
- "eval_rougeL": 0.970368187542237,
755
- "eval_rougeLsum": 0.9703795940864277,
756
- "eval_runtime": 357.9392,
757
- "eval_samples_per_second": 119.772,
758
- "eval_steps_per_second": 9.982,
759
- "eval_translation_length": 696018,
760
- "step": 160000
761
- },
762
- {
763
- "epoch": 2.52,
764
- "learning_rate": 7.219369030269242e-06,
765
- "loss": 0.0421,
766
- "step": 162000
767
- },
768
- {
769
- "epoch": 2.55,
770
- "learning_rate": 7.110456637268422e-06,
771
- "loss": 0.0428,
772
- "step": 164000
773
- },
774
- {
775
- "epoch": 2.58,
776
- "learning_rate": 7.000367166717425e-06,
777
- "loss": 0.0401,
778
- "step": 166000
779
- },
780
- {
781
- "epoch": 2.61,
782
- "learning_rate": 6.889110705881452e-06,
783
- "loss": 0.0417,
784
- "step": 168000
785
- },
786
- {
787
- "epoch": 2.64,
788
- "learning_rate": 6.776752160449367e-06,
789
- "loss": 0.042,
790
- "step": 170000
791
- },
792
- {
793
- "epoch": 2.64,
794
- "eval_bleu": 0.9447481654515599,
795
- "eval_brevity_penalty": 1.0,
796
- "eval_length_ratio": 1.0186748491903599,
797
- "eval_loss": 0.04124660789966583,
798
- "eval_reference_length": 688948,
799
- "eval_rouge1": 0.9616790997827227,
800
- "eval_rouge2": 0.9356205732284686,
801
- "eval_rougeL": 0.961519978707468,
802
- "eval_rougeLsum": 0.9615393434549021,
803
- "eval_runtime": 350.4802,
804
- "eval_samples_per_second": 122.321,
805
- "eval_steps_per_second": 10.195,
806
- "eval_translation_length": 701814,
807
- "step": 170000
808
- },
809
- {
810
- "epoch": 2.67,
811
- "learning_rate": 6.663300104771134e-06,
812
- "loss": 0.0386,
813
- "step": 172000
814
- },
815
- {
816
- "epoch": 2.71,
817
- "learning_rate": 6.548876728670889e-06,
818
- "loss": 0.0442,
819
- "step": 174000
820
- },
821
- {
822
- "epoch": 2.74,
823
- "learning_rate": 6.433664609478793e-06,
824
- "loss": 0.0417,
825
- "step": 176000
826
- },
827
- {
828
- "epoch": 2.77,
829
- "learning_rate": 6.3175003841713e-06,
830
- "loss": 0.0418,
831
- "step": 178000
832
- },
833
- {
834
- "epoch": 2.8,
835
- "learning_rate": 6.200566775740774e-06,
836
- "loss": 0.0406,
837
- "step": 180000
838
- },
839
- {
840
- "epoch": 2.8,
841
- "eval_bleu": 0.9487462287220352,
842
- "eval_brevity_penalty": 1.0,
843
- "eval_length_ratio": 1.0163989154479003,
844
- "eval_loss": 0.04077766090631485,
845
- "eval_reference_length": 688948,
846
- "eval_rouge1": 0.9665622993166698,
847
- "eval_rouge2": 0.9412909286045468,
848
- "eval_rougeL": 0.9663690505539053,
849
- "eval_rougeLsum": 0.9663768593347282,
850
- "eval_runtime": 354.3716,
851
- "eval_samples_per_second": 120.978,
852
- "eval_steps_per_second": 10.083,
853
- "eval_translation_length": 700246,
854
- "step": 180000
855
- },
856
- {
857
- "epoch": 2.83,
858
- "learning_rate": 6.083109012991928e-06,
859
- "loss": 0.0416,
860
- "step": 182000
861
- },
862
- {
863
- "epoch": 2.86,
864
- "learning_rate": 5.964842802508876e-06,
865
- "loss": 0.0404,
866
- "step": 184000
867
- },
868
- {
869
- "epoch": 2.89,
870
- "learning_rate": 5.846013150999504e-06,
871
- "loss": 0.0379,
872
- "step": 186000
873
- },
874
- {
875
- "epoch": 2.92,
876
- "learning_rate": 5.726749225339994e-06,
877
- "loss": 0.0432,
878
- "step": 188000
879
- },
880
- {
881
- "epoch": 2.95,
882
- "learning_rate": 5.607001354527006e-06,
883
- "loss": 0.0393,
884
- "step": 190000
885
- },
886
- {
887
- "epoch": 2.95,
888
- "eval_bleu": 0.9547919266733801,
889
- "eval_brevity_penalty": 1.0,
890
- "eval_length_ratio": 1.0112867734575033,
891
- "eval_loss": 0.04095698148012161,
892
- "eval_reference_length": 688948,
893
- "eval_rouge1": 0.9703758390841425,
894
- "eval_rouge2": 0.9461704701896383,
895
- "eval_rougeL": 0.9702899676023662,
896
- "eval_rougeLsum": 0.9703172456040432,
897
- "eval_runtime": 353.5109,
898
- "eval_samples_per_second": 121.272,
899
- "eval_steps_per_second": 10.107,
900
- "eval_translation_length": 696724,
901
- "step": 190000
902
- },
903
- {
904
- "epoch": 2.99,
905
- "learning_rate": 5.486959140088201e-06,
906
- "loss": 0.0384,
907
- "step": 192000
908
- },
909
- {
910
- "epoch": 3.02,
911
- "learning_rate": 5.366572586878771e-06,
912
- "loss": 0.0354,
913
- "step": 194000
914
- },
915
- {
916
- "epoch": 3.05,
917
- "learning_rate": 5.246032307677414e-06,
918
- "loss": 0.0308,
919
- "step": 196000
920
- },
921
- {
922
- "epoch": 3.08,
923
- "learning_rate": 5.125469288643505e-06,
924
- "loss": 0.033,
925
- "step": 198000
926
- },
927
- {
928
- "epoch": 3.11,
929
- "learning_rate": 5.004651973065896e-06,
930
- "loss": 0.0316,
931
- "step": 200000
932
- },
933
- {
934
- "epoch": 3.11,
935
- "eval_bleu": 0.9516992046244281,
936
- "eval_brevity_penalty": 1.0,
937
- "eval_length_ratio": 1.013291278877361,
938
- "eval_loss": 0.04207807779312134,
939
- "eval_reference_length": 688948,
940
- "eval_rouge1": 0.9687692703791779,
941
- "eval_rouge2": 0.9438025496007139,
942
- "eval_rougeL": 0.9686298083794034,
943
- "eval_rougeLsum": 0.9686054829671312,
944
- "eval_runtime": 358.3539,
945
- "eval_samples_per_second": 119.633,
946
- "eval_steps_per_second": 9.971,
947
- "eval_translation_length": 698105,
948
- "step": 200000
949
- },
950
- {
951
- "epoch": 3.14,
952
- "learning_rate": 4.883831940867018e-06,
953
- "loss": 0.0332,
954
- "step": 202000
955
- },
956
- {
957
- "epoch": 3.17,
958
- "learning_rate": 4.763079747543336e-06,
959
- "loss": 0.03,
960
- "step": 204000
961
- },
962
- {
963
- "epoch": 3.2,
964
- "learning_rate": 4.642526169588442e-06,
965
- "loss": 0.0313,
966
- "step": 206000
967
- },
968
- {
969
- "epoch": 3.23,
970
- "learning_rate": 4.52218113759159e-06,
971
- "loss": 0.0319,
972
- "step": 208000
973
- },
974
- {
975
- "epoch": 3.27,
976
- "learning_rate": 4.402114859405383e-06,
977
- "loss": 0.0313,
978
- "step": 210000
979
- },
980
- {
981
- "epoch": 3.27,
982
- "eval_bleu": 0.952314893533454,
983
- "eval_brevity_penalty": 1.0,
984
- "eval_length_ratio": 1.0132433797616076,
985
- "eval_loss": 0.04261790215969086,
986
- "eval_reference_length": 688948,
987
- "eval_rouge1": 0.9687049595684978,
988
- "eval_rouge2": 0.9435984713700241,
989
- "eval_rougeL": 0.9686312046722289,
990
- "eval_rougeLsum": 0.9685702469811763,
991
- "eval_runtime": 353.4407,
992
- "eval_samples_per_second": 121.296,
993
- "eval_steps_per_second": 10.109,
994
- "eval_translation_length": 698072,
995
- "step": 210000
996
- },
997
- {
998
- "epoch": 3.3,
999
- "learning_rate": 4.2823973802607795e-06,
1000
- "loss": 0.0322,
1001
- "step": 212000
1002
- },
1003
- {
1004
- "epoch": 3.33,
1005
- "learning_rate": 4.163038979034976e-06,
1006
- "loss": 0.0312,
1007
- "step": 214000
1008
- },
1009
- {
1010
- "epoch": 3.36,
1011
- "learning_rate": 4.0441693394762706e-06,
1012
- "loss": 0.0314,
1013
- "step": 216000
1014
- },
1015
- {
1016
- "epoch": 3.39,
1017
- "learning_rate": 3.925916882841615e-06,
1018
- "loss": 0.0332,
1019
- "step": 218000
1020
- },
1021
- {
1022
- "epoch": 3.42,
1023
- "learning_rate": 3.8083497076975863e-06,
1024
- "loss": 0.0302,
1025
- "step": 220000
1026
- },
1027
- {
1028
- "epoch": 3.42,
1029
- "eval_bleu": 0.9510751957840793,
1030
- "eval_brevity_penalty": 1.0,
1031
- "eval_length_ratio": 1.014170880821194,
1032
- "eval_loss": 0.04271363466978073,
1033
- "eval_reference_length": 688948,
1034
- "eval_rouge1": 0.9682283271173224,
1035
- "eval_rouge2": 0.9430416842293828,
1036
- "eval_rougeL": 0.9681608428660635,
1037
- "eval_rougeLsum": 0.9681434321637754,
1038
- "eval_runtime": 354.0647,
1039
- "eval_samples_per_second": 121.082,
1040
- "eval_steps_per_second": 10.091,
1041
- "eval_translation_length": 698711,
1042
- "step": 220000
1043
- },
1044
- {
1045
- "epoch": 3.45,
1046
- "learning_rate": 3.691418722192835e-06,
1047
- "loss": 0.0323,
1048
- "step": 222000
1049
- },
1050
- {
1051
- "epoch": 3.48,
1052
- "learning_rate": 3.5751932368863875e-06,
1053
- "loss": 0.0302,
1054
- "step": 224000
1055
- },
1056
- {
1057
- "epoch": 3.51,
1058
- "learning_rate": 3.4597997986064915e-06,
1059
- "loss": 0.0312,
1060
- "step": 226000
1061
- },
1062
- {
1063
- "epoch": 3.55,
1064
- "learning_rate": 3.3453057938715767e-06,
1065
- "loss": 0.0309,
1066
- "step": 228000
1067
- },
1068
- {
1069
- "epoch": 3.58,
1070
- "learning_rate": 3.2318345952925634e-06,
1071
- "loss": 0.0305,
1072
- "step": 230000
1073
- },
1074
- {
1075
- "epoch": 3.58,
1076
- "eval_bleu": 0.9477212328461616,
1077
- "eval_brevity_penalty": 1.0,
1078
- "eval_length_ratio": 1.017851855292417,
1079
- "eval_loss": 0.042060352861881256,
1080
- "eval_reference_length": 688948,
1081
- "eval_rouge1": 0.9686642355024724,
1082
- "eval_rouge2": 0.9435154406416524,
1083
- "eval_rougeL": 0.9685609023356937,
1084
- "eval_rougeLsum": 0.968565755092943,
1085
- "eval_runtime": 358.0746,
1086
- "eval_samples_per_second": 119.726,
1087
- "eval_steps_per_second": 9.978,
1088
- "eval_translation_length": 701247,
1089
- "step": 230000
1090
- },
1091
- {
1092
- "epoch": 3.61,
1093
- "learning_rate": 3.1194509023896597e-06,
1094
- "loss": 0.0342,
1095
- "step": 232000
1096
- },
1097
- {
1098
- "epoch": 3.64,
1099
- "learning_rate": 3.0081077877484786e-06,
1100
- "loss": 0.0297,
1101
- "step": 234000
1102
- },
1103
- {
1104
- "epoch": 3.67,
1105
- "learning_rate": 2.8978719026697843e-06,
1106
- "loss": 0.0308,
1107
- "step": 236000
1108
- },
1109
- {
1110
- "epoch": 3.7,
1111
- "learning_rate": 2.7889177879789993e-06,
1112
- "loss": 0.0317,
1113
- "step": 238000
1114
- },
1115
- {
1116
- "epoch": 3.73,
1117
- "learning_rate": 2.6812000664997108e-06,
1118
- "loss": 0.0317,
1119
- "step": 240000
1120
- },
1121
- {
1122
- "epoch": 3.73,
1123
- "eval_bleu": 0.9545413224765983,
1124
- "eval_brevity_penalty": 1.0,
1125
- "eval_length_ratio": 1.0115161086177766,
1126
- "eval_loss": 0.04142594709992409,
1127
- "eval_reference_length": 688948,
1128
- "eval_rouge1": 0.9706993144139382,
1129
- "eval_rouge2": 0.9462893219706924,
1130
- "eval_rougeL": 0.9706221623216551,
1131
- "eval_rougeLsum": 0.9705871778859747,
1132
- "eval_runtime": 352.9472,
1133
- "eval_samples_per_second": 121.466,
1134
- "eval_steps_per_second": 10.123,
1135
- "eval_translation_length": 696882,
1136
- "step": 240000
1137
- },
1138
- {
1139
- "epoch": 3.76,
1140
- "learning_rate": 2.574942125369937e-06,
1141
- "loss": 0.0298,
1142
- "step": 242000
1143
- },
1144
- {
1145
- "epoch": 3.79,
1146
- "learning_rate": 2.4699932979874153e-06,
1147
- "loss": 0.0319,
1148
- "step": 244000
1149
- },
1150
- {
1151
- "epoch": 3.83,
1152
- "learning_rate": 2.3665732796817735e-06,
1153
- "loss": 0.0314,
1154
- "step": 246000
1155
- },
1156
- {
1157
- "epoch": 3.86,
1158
- "learning_rate": 2.2646389981153643e-06,
1159
- "loss": 0.0312,
1160
- "step": 248000
1161
- },
1162
- {
1163
- "epoch": 3.89,
1164
- "learning_rate": 2.1643020903452345e-06,
1165
- "loss": 0.0298,
1166
- "step": 250000
1167
- },
1168
- {
1169
- "epoch": 3.89,
1170
- "eval_bleu": 0.9529006548919251,
1171
- "eval_brevity_penalty": 1.0,
1172
- "eval_length_ratio": 1.0129792088807863,
1173
- "eval_loss": 0.040760207921266556,
1174
- "eval_reference_length": 688948,
1175
- "eval_rouge1": 0.9693489479213169,
1176
- "eval_rouge2": 0.9444261412988928,
1177
- "eval_rougeL": 0.9692149741999472,
1178
- "eval_rougeLsum": 0.9692295166277185,
1179
- "eval_runtime": 355.5559,
1180
- "eval_samples_per_second": 120.575,
1181
- "eval_steps_per_second": 10.049,
1182
- "eval_translation_length": 697890,
1183
- "step": 250000
1184
- },
1185
- {
1186
- "epoch": 3.92,
1187
- "learning_rate": 2.0656700673466744e-06,
1188
- "loss": 0.032,
1189
- "step": 252000
1190
- },
1191
- {
1192
- "epoch": 3.95,
1193
- "learning_rate": 1.9687498973425523e-06,
1194
- "loss": 0.0288,
1195
- "step": 254000
1196
- },
1197
- {
1198
- "epoch": 3.98,
1199
- "learning_rate": 1.8735981224010946e-06,
1200
- "loss": 0.0301,
1201
- "step": 256000
1202
- },
1203
- {
1204
- "epoch": 4.01,
1205
- "learning_rate": 1.7802702529299903e-06,
1206
- "loss": 0.0285,
1207
- "step": 258000
1208
- },
1209
- {
1210
- "epoch": 4.04,
1211
- "learning_rate": 1.6888207352922886e-06,
1212
- "loss": 0.0249,
1213
- "step": 260000
1214
- },
1215
- {
1216
- "epoch": 4.04,
1217
- "eval_bleu": 0.9580735852306802,
1218
- "eval_brevity_penalty": 1.0,
1219
- "eval_length_ratio": 1.0084215354424426,
1220
- "eval_loss": 0.04249930381774902,
1221
- "eval_reference_length": 688948,
1222
- "eval_rouge1": 0.9725141407368192,
1223
- "eval_rouge2": 0.9482526619227644,
1224
- "eval_rougeL": 0.9723960508025267,
1225
- "eval_rougeLsum": 0.972420107341233,
1226
- "eval_runtime": 352.4144,
1227
- "eval_samples_per_second": 121.649,
1228
- "eval_steps_per_second": 10.139,
1229
- "eval_translation_length": 694750,
1230
- "step": 260000
1231
- },
1232
- {
1233
- "epoch": 4.07,
1234
- "learning_rate": 1.599258630916235e-06,
1235
- "loss": 0.0231,
1236
- "step": 262000
1237
- },
1238
- {
1239
- "epoch": 4.11,
1240
- "learning_rate": 1.5117690308052164e-06,
1241
- "loss": 0.0249,
1242
- "step": 264000
1243
- },
1244
- {
1245
- "epoch": 4.14,
1246
- "learning_rate": 1.4262278806001696e-06,
1247
- "loss": 0.0242,
1248
- "step": 266000
1249
- },
1250
- {
1251
- "epoch": 4.17,
1252
- "learning_rate": 1.3427737126740498e-06,
1253
- "loss": 0.0235,
1254
- "step": 268000
1255
- },
1256
- {
1257
- "epoch": 4.2,
1258
- "learning_rate": 1.261495379512549e-06,
1259
- "loss": 0.0236,
1260
- "step": 270000
1261
- },
1262
- {
1263
- "epoch": 4.2,
1264
- "eval_bleu": 0.9578225644568897,
1265
- "eval_brevity_penalty": 1.0,
1266
- "eval_length_ratio": 1.0083475095362786,
1267
- "eval_loss": 0.04301063343882561,
1268
- "eval_reference_length": 688948,
1269
- "eval_rouge1": 0.9727779850827751,
1270
- "eval_rouge2": 0.948452540502593,
1271
- "eval_rougeL": 0.9726976320174251,
1272
- "eval_rougeLsum": 0.9727011117301205,
1273
- "eval_runtime": 360.9368,
1274
- "eval_samples_per_second": 118.777,
1275
- "eval_steps_per_second": 9.899,
1276
- "eval_translation_length": 694699,
1277
- "step": 270000
1278
- },
1279
- {
1280
- "epoch": 4.23,
1281
- "learning_rate": 1.1823980452421706e-06,
1282
- "loss": 0.024,
1283
- "step": 272000
1284
- },
1285
- {
1286
- "epoch": 4.26,
1287
- "learning_rate": 1.105489964860681e-06,
1288
- "loss": 0.0228,
1289
- "step": 274000
1290
- },
1291
- {
1292
- "epoch": 4.29,
1293
- "learning_rate": 1.0308929099656145e-06,
1294
- "loss": 0.0236,
1295
- "step": 276000
1296
- },
1297
- {
1298
- "epoch": 4.32,
1299
- "learning_rate": 9.586469575615094e-07,
1300
- "loss": 0.025,
1301
- "step": 278000
1302
- },
1303
- {
1304
- "epoch": 4.35,
1305
- "learning_rate": 8.887219403201786e-07,
1306
- "loss": 0.0259,
1307
- "step": 280000
1308
- },
1309
- {
1310
- "epoch": 4.35,
1311
- "eval_bleu": 0.9543405944053881,
1312
- "eval_brevity_penalty": 1.0,
1313
- "eval_length_ratio": 1.0119820363801042,
1314
- "eval_loss": 0.04318338260054588,
1315
- "eval_reference_length": 688948,
1316
- "eval_rouge1": 0.9707381997207458,
1317
- "eval_rouge2": 0.9463171031333721,
1318
- "eval_rougeL": 0.9706594079415762,
1319
- "eval_rougeLsum": 0.9706575791884534,
1320
- "eval_runtime": 353.2216,
1321
- "eval_samples_per_second": 121.371,
1322
- "eval_steps_per_second": 10.115,
1323
- "eval_translation_length": 697203,
1324
- "step": 280000
1325
- },
1326
- {
1327
- "epoch": 4.39,
1328
- "learning_rate": 8.211953935368261e-07,
1329
- "loss": 0.0236,
1330
- "step": 282000
1331
- },
1332
- {
1333
- "epoch": 4.42,
1334
- "learning_rate": 7.560747672425183e-07,
1335
- "loss": 0.0243,
1336
- "step": 284000
1337
- },
1338
- {
1339
- "epoch": 4.45,
1340
- "learning_rate": 6.934324737735693e-07,
1341
- "loss": 0.0245,
1342
- "step": 286000
1343
- },
1344
- {
1345
- "epoch": 4.48,
1346
- "learning_rate": 6.333345238313842e-07,
1347
- "loss": 0.0241,
1348
- "step": 288000
1349
- },
1350
- {
1351
- "epoch": 4.51,
1352
- "learning_rate": 5.757840343120758e-07,
1353
- "loss": 0.0236,
1354
- "step": 290000
1355
- },
1356
- {
1357
- "epoch": 4.51,
1358
- "eval_bleu": 0.9554825445943746,
1359
- "eval_brevity_penalty": 1.0,
1360
- "eval_length_ratio": 1.0108963230896961,
1361
- "eval_loss": 0.04305826872587204,
1362
- "eval_reference_length": 688948,
1363
- "eval_rouge1": 0.9714257759381628,
1364
- "eval_rouge2": 0.9469647669616812,
1365
- "eval_rougeL": 0.971368415701727,
1366
- "eval_rougeLsum": 0.9713609918743913,
1367
- "eval_runtime": 360.9244,
1368
- "eval_samples_per_second": 118.781,
1369
- "eval_steps_per_second": 9.9,
1370
- "eval_translation_length": 696455,
1371
- "step": 290000
1372
- },
1373
- {
1374
- "epoch": 4.54,
1375
- "learning_rate": 5.208145794830483e-07,
1376
- "loss": 0.023,
1377
- "step": 292000
1378
- },
1379
- {
1380
- "epoch": 4.57,
1381
- "learning_rate": 4.684326956861246e-07,
1382
- "loss": 0.0248,
1383
- "step": 294000
1384
- },
1385
- {
1386
- "epoch": 4.6,
1387
- "learning_rate": 4.186971195842365e-07,
1388
- "loss": 0.0234,
1389
- "step": 296000
1390
- },
1391
- {
1392
- "epoch": 4.63,
1393
- "learning_rate": 3.716597523194587e-07,
1394
- "loss": 0.023,
1395
- "step": 298000
1396
- },
1397
- {
1398
- "epoch": 4.67,
1399
- "learning_rate": 3.27322503410189e-07,
1400
- "loss": 0.0251,
1401
- "step": 300000
1402
- },
1403
- {
1404
- "epoch": 4.67,
1405
- "eval_bleu": 0.9555293197464279,
1406
- "eval_brevity_penalty": 1.0,
1407
- "eval_length_ratio": 1.010778752532847,
1408
- "eval_loss": 0.04303622618317604,
1409
- "eval_reference_length": 688948,
1410
- "eval_rouge1": 0.9711184557484207,
1411
- "eval_rouge2": 0.9465872069219559,
1412
- "eval_rougeL": 0.971016190214522,
1413
- "eval_rougeLsum": 0.9710369956602919,
1414
- "eval_runtime": 356.11,
1415
- "eval_samples_per_second": 120.387,
1416
- "eval_steps_per_second": 10.033,
1417
- "eval_translation_length": 696374,
1418
- "step": 300000
1419
- },
1420
- {
1421
- "epoch": 4.7,
1422
- "learning_rate": 2.857112386772626e-07,
1423
- "loss": 0.025,
1424
- "step": 302000
1425
- },
1426
- {
1427
- "epoch": 4.73,
1428
- "learning_rate": 2.468314855157933e-07,
1429
- "loss": 0.0246,
1430
- "step": 304000
1431
- },
1432
- {
1433
- "epoch": 4.76,
1434
- "learning_rate": 2.1072744891572238e-07,
1435
- "loss": 0.0244,
1436
- "step": 306000
1437
- },
1438
- {
1439
- "epoch": 4.79,
1440
- "learning_rate": 1.7745211624187464e-07,
1441
- "loss": 0.0235,
1442
- "step": 308000
1443
- },
1444
- {
1445
- "epoch": 4.82,
1446
- "learning_rate": 1.469728454791536e-07,
1447
- "loss": 0.0241,
1448
- "step": 310000
1449
- },
1450
- {
1451
- "epoch": 4.82,
1452
- "eval_bleu": 0.9564463934340444,
1453
- "eval_brevity_penalty": 1.0,
1454
- "eval_length_ratio": 1.0099470497047673,
1455
- "eval_loss": 0.04298330098390579,
1456
- "eval_reference_length": 688948,
1457
- "eval_rouge1": 0.9718286385746833,
1458
- "eval_rouge2": 0.9474516377136643,
1459
- "eval_rougeL": 0.9717457141369523,
1460
- "eval_rougeLsum": 0.9717098073885057,
1461
- "eval_runtime": 351.9107,
1462
- "eval_samples_per_second": 121.824,
1463
- "eval_steps_per_second": 10.153,
1464
- "eval_translation_length": 695801,
1465
- "step": 310000
1466
- },
1467
- {
1468
- "epoch": 4.85,
1469
- "learning_rate": 1.1931165304333803e-07,
1470
- "loss": 0.0229,
1471
- "step": 312000
1472
- },
1473
- {
1474
- "epoch": 4.88,
1475
- "learning_rate": 9.450064516007773e-08,
1476
- "loss": 0.0245,
1477
- "step": 314000
1478
- },
1479
- {
1480
- "epoch": 4.91,
1481
- "learning_rate": 7.256456591191674e-08,
1482
- "loss": 0.0241,
1483
- "step": 316000
1484
- },
1485
- {
1486
- "epoch": 4.94,
1487
- "learning_rate": 5.351190850388044e-08,
1488
- "loss": 0.0247,
1489
- "step": 318000
1490
- },
1491
- {
1492
- "epoch": 4.98,
1493
- "learning_rate": 3.7327348382793504e-08,
1494
- "loss": 0.025,
1495
- "step": 320000
1496
- },
1497
- {
1498
- "epoch": 4.98,
1499
- "eval_bleu": 0.9562524667691357,
1500
- "eval_brevity_penalty": 1.0,
1501
- "eval_length_ratio": 1.010135743191068,
1502
- "eval_loss": 0.04290972650051117,
1503
- "eval_reference_length": 688948,
1504
- "eval_rouge1": 0.9716126931407336,
1505
- "eval_rouge2": 0.9473094321074684,
1506
- "eval_rougeL": 0.9715449461188315,
1507
- "eval_rougeLsum": 0.9715191110303505,
1508
- "eval_runtime": 358.3002,
1509
- "eval_samples_per_second": 119.651,
1510
- "eval_steps_per_second": 9.972,
1511
- "eval_translation_length": 695931,
1512
- "step": 320000
1513
- },
1514
- {
1515
- "epoch": 5.01,
1516
- "learning_rate": 2.4046767073176436e-08,
1517
- "loss": 0.0231,
1518
- "step": 322000
1519
- },
1520
- {
1521
- "epoch": 5.04,
1522
- "learning_rate": 1.3664633482581291e-08,
1523
- "loss": 0.0213,
1524
- "step": 324000
1525
- },
1526
- {
1527
- "epoch": 5.07,
1528
- "learning_rate": 6.200393505542135e-09,
1529
- "loss": 0.0229,
1530
- "step": 326000
1531
- },
1532
- {
1533
- "epoch": 5.1,
1534
- "learning_rate": 1.6450126084593953e-09,
1535
- "loss": 0.0226,
1536
- "step": 328000
1537
- },
1538
- {
1539
- "epoch": 5.13,
1540
- "learning_rate": 5.612241452679357e-12,
1541
- "loss": 0.0216,
1542
- "step": 330000
1543
- },
1544
- {
1545
- "epoch": 5.13,
1546
- "eval_bleu": 0.956286137798364,
1547
- "eval_brevity_penalty": 1.0,
1548
- "eval_length_ratio": 1.0101197768191503,
1549
- "eval_loss": 0.04293430969119072,
1550
- "eval_reference_length": 688948,
1551
- "eval_rouge1": 0.9716103880266641,
1552
- "eval_rouge2": 0.9473106979225123,
1553
- "eval_rougeL": 0.9715446246919781,
1554
- "eval_rougeLsum": 0.971525437075219,
1555
- "eval_runtime": 356.663,
1556
- "eval_samples_per_second": 120.2,
1557
- "eval_steps_per_second": 10.018,
1558
- "eval_translation_length": 695920,
1559
- "step": 330000
1560
- },
1561
- {
1562
- "epoch": 5.13,
1563
- "step": 330000,
1564
- "total_flos": 3.0484200532475904e+16,
1565
- "train_loss": 0.08529333166642622,
1566
- "train_runtime": 78966.8059,
1567
- "train_samples_per_second": 25.074,
1568
- "train_steps_per_second": 4.179
1569
- }
1570
- ],
1571
- "max_steps": 330000,
1572
- "num_train_epochs": 6,
1573
- "total_flos": 3.0484200532475904e+16,
1574
- "trial_name": null,
1575
- "trial_params": null
1576
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cce437cedae8c24f6d18c98845dc6e0e47724da69c7957d1573301b1d6e6301b
3
- size 3695
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b738475cfda6e757adae542018576545744bdcce6aef2f1e26a9be48fa6be09
3
+ size 3771