Tsubasaz commited on
Commit
564ba91
1 Parent(s): 0d51cef

Upload trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +736 -0
trainer_state.json ADDED
@@ -0,0 +1,736 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 4.919645785503444,
5
+ "global_step": 60000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.04,
12
+ "learning_rate": 5e-06,
13
+ "loss": 8.6674,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.08,
18
+ "learning_rate": 1e-05,
19
+ "loss": 6.0699,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.12,
24
+ "learning_rate": 1.5e-05,
25
+ "loss": 5.3064,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.16,
30
+ "learning_rate": 2e-05,
31
+ "loss": 4.6019,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.2,
36
+ "learning_rate": 2.5e-05,
37
+ "loss": 3.9841,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.25,
42
+ "learning_rate": 3e-05,
43
+ "loss": 3.5033,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.29,
48
+ "learning_rate": 3.5e-05,
49
+ "loss": 3.1509,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.33,
54
+ "learning_rate": 4e-05,
55
+ "loss": 2.8726,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.37,
60
+ "learning_rate": 4.5e-05,
61
+ "loss": 2.6501,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.41,
66
+ "learning_rate": 5e-05,
67
+ "loss": 2.4705,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.45,
72
+ "learning_rate": 4.955341193283315e-05,
73
+ "loss": 2.329,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.49,
78
+ "learning_rate": 4.9106823865666315e-05,
79
+ "loss": 2.2016,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.53,
84
+ "learning_rate": 4.8660235798499465e-05,
85
+ "loss": 2.108,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.57,
90
+ "learning_rate": 4.821364773133262e-05,
91
+ "loss": 2.0367,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.61,
96
+ "learning_rate": 4.776705966416578e-05,
97
+ "loss": 1.9619,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.66,
102
+ "learning_rate": 4.7320471596998934e-05,
103
+ "loss": 1.905,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.7,
108
+ "learning_rate": 4.6873883529832084e-05,
109
+ "loss": 1.8638,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.74,
114
+ "learning_rate": 4.642729546266524e-05,
115
+ "loss": 1.8176,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.78,
120
+ "learning_rate": 4.59807073954984e-05,
121
+ "loss": 1.7665,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.82,
126
+ "learning_rate": 4.5534119328331546e-05,
127
+ "loss": 1.7427,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.86,
132
+ "learning_rate": 4.50875312611647e-05,
133
+ "loss": 1.7148,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.9,
138
+ "learning_rate": 4.464094319399786e-05,
139
+ "loss": 1.686,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.94,
144
+ "learning_rate": 4.419435512683101e-05,
145
+ "loss": 1.6532,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.98,
150
+ "learning_rate": 4.374776705966417e-05,
151
+ "loss": 1.6338,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 1.02,
156
+ "learning_rate": 4.330117899249732e-05,
157
+ "loss": 1.6049,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 1.07,
162
+ "learning_rate": 4.285459092533048e-05,
163
+ "loss": 1.5844,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 1.11,
168
+ "learning_rate": 4.2408002858163634e-05,
169
+ "loss": 1.5617,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 1.15,
174
+ "learning_rate": 4.196141479099679e-05,
175
+ "loss": 1.5475,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 1.19,
180
+ "learning_rate": 4.151482672382994e-05,
181
+ "loss": 1.533,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 1.23,
186
+ "learning_rate": 4.10682386566631e-05,
187
+ "loss": 1.5207,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 1.27,
192
+ "learning_rate": 4.062165058949625e-05,
193
+ "loss": 1.5027,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 1.31,
198
+ "learning_rate": 4.01750625223294e-05,
199
+ "loss": 1.4911,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 1.35,
204
+ "learning_rate": 3.972847445516256e-05,
205
+ "loss": 1.4758,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 1.39,
210
+ "learning_rate": 3.9281886387995716e-05,
211
+ "loss": 1.4632,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 1.43,
216
+ "learning_rate": 3.8835298320828865e-05,
217
+ "loss": 1.4508,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 1.48,
222
+ "learning_rate": 3.838871025366202e-05,
223
+ "loss": 1.4448,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 1.52,
228
+ "learning_rate": 3.794212218649518e-05,
229
+ "loss": 1.4325,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 1.56,
234
+ "learning_rate": 3.7495534119328335e-05,
235
+ "loss": 1.4219,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 1.6,
240
+ "learning_rate": 3.7048946052161484e-05,
241
+ "loss": 1.4095,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 1.64,
246
+ "learning_rate": 3.660235798499465e-05,
247
+ "loss": 1.4042,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 1.68,
252
+ "learning_rate": 3.61557699178278e-05,
253
+ "loss": 1.3934,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 1.72,
258
+ "learning_rate": 3.570918185066095e-05,
259
+ "loss": 1.3905,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 1.76,
264
+ "learning_rate": 3.526259378349411e-05,
265
+ "loss": 1.3783,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 1.8,
270
+ "learning_rate": 3.481600571632726e-05,
271
+ "loss": 1.3627,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 1.84,
276
+ "learning_rate": 3.4369417649160416e-05,
277
+ "loss": 1.3597,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 1.89,
282
+ "learning_rate": 3.392282958199357e-05,
283
+ "loss": 1.3594,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 1.93,
288
+ "learning_rate": 3.347624151482672e-05,
289
+ "loss": 1.3407,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 1.97,
294
+ "learning_rate": 3.302965344765988e-05,
295
+ "loss": 1.3396,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 2.01,
300
+ "learning_rate": 3.2583065380493035e-05,
301
+ "loss": 1.3309,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 2.05,
306
+ "learning_rate": 3.213647731332619e-05,
307
+ "loss": 1.3214,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 2.09,
312
+ "learning_rate": 3.168988924615934e-05,
313
+ "loss": 1.3113,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 2.13,
318
+ "learning_rate": 3.1243301178992504e-05,
319
+ "loss": 1.3082,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 2.17,
324
+ "learning_rate": 3.0796713111825654e-05,
325
+ "loss": 1.3057,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 2.21,
330
+ "learning_rate": 3.0350125044658807e-05,
331
+ "loss": 1.2962,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 2.25,
336
+ "learning_rate": 2.9903536977491963e-05,
337
+ "loss": 1.2969,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 2.3,
342
+ "learning_rate": 2.9456948910325116e-05,
343
+ "loss": 1.2909,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 2.34,
348
+ "learning_rate": 2.9010360843158273e-05,
349
+ "loss": 1.2803,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 2.38,
354
+ "learning_rate": 2.8563772775991426e-05,
355
+ "loss": 1.2762,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 2.42,
360
+ "learning_rate": 2.811718470882458e-05,
361
+ "loss": 1.2691,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 2.46,
366
+ "learning_rate": 2.7670596641657735e-05,
367
+ "loss": 1.2698,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 2.5,
372
+ "learning_rate": 2.722400857449089e-05,
373
+ "loss": 1.2646,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 2.54,
378
+ "learning_rate": 2.6777420507324048e-05,
379
+ "loss": 1.2598,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 2.58,
384
+ "learning_rate": 2.6330832440157198e-05,
385
+ "loss": 1.2598,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 2.62,
390
+ "learning_rate": 2.5884244372990358e-05,
391
+ "loss": 1.2499,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 2.66,
396
+ "learning_rate": 2.543765630582351e-05,
397
+ "loss": 1.2492,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 2.71,
402
+ "learning_rate": 2.4991068238656664e-05,
403
+ "loss": 1.2417,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 2.75,
408
+ "learning_rate": 2.454448017148982e-05,
409
+ "loss": 1.2458,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 2.79,
414
+ "learning_rate": 2.4097892104322976e-05,
415
+ "loss": 1.2428,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 2.83,
420
+ "learning_rate": 2.3651304037156126e-05,
421
+ "loss": 1.2319,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 2.87,
426
+ "learning_rate": 2.3204715969989283e-05,
427
+ "loss": 1.2334,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 2.91,
432
+ "learning_rate": 2.275812790282244e-05,
433
+ "loss": 1.2275,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 2.95,
438
+ "learning_rate": 2.2311539835655592e-05,
439
+ "loss": 1.2234,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 2.99,
444
+ "learning_rate": 2.186495176848875e-05,
445
+ "loss": 1.2297,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 3.03,
450
+ "learning_rate": 2.14183637013219e-05,
451
+ "loss": 1.2149,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 3.07,
456
+ "learning_rate": 2.0971775634155054e-05,
457
+ "loss": 1.2124,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 3.12,
462
+ "learning_rate": 2.052518756698821e-05,
463
+ "loss": 1.2099,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 3.16,
468
+ "learning_rate": 2.0078599499821364e-05,
469
+ "loss": 1.1997,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 3.2,
474
+ "learning_rate": 1.963201143265452e-05,
475
+ "loss": 1.2052,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 3.24,
480
+ "learning_rate": 1.9185423365487677e-05,
481
+ "loss": 1.2032,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 3.28,
486
+ "learning_rate": 1.873883529832083e-05,
487
+ "loss": 1.1958,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 3.32,
492
+ "learning_rate": 1.8292247231153983e-05,
493
+ "loss": 1.1938,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 3.36,
498
+ "learning_rate": 1.784565916398714e-05,
499
+ "loss": 1.1927,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 3.4,
504
+ "learning_rate": 1.7399071096820292e-05,
505
+ "loss": 1.1898,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 3.44,
510
+ "learning_rate": 1.695248302965345e-05,
511
+ "loss": 1.1871,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 3.48,
516
+ "learning_rate": 1.6505894962486605e-05,
517
+ "loss": 1.1847,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 3.53,
522
+ "learning_rate": 1.6059306895319758e-05,
523
+ "loss": 1.1805,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 3.57,
528
+ "learning_rate": 1.561271882815291e-05,
529
+ "loss": 1.182,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 3.61,
534
+ "learning_rate": 1.5166130760986066e-05,
535
+ "loss": 1.1766,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 3.65,
540
+ "learning_rate": 1.4719542693819222e-05,
541
+ "loss": 1.177,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 3.69,
546
+ "learning_rate": 1.4272954626652377e-05,
547
+ "loss": 1.1756,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 3.73,
552
+ "learning_rate": 1.3826366559485532e-05,
553
+ "loss": 1.1744,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 3.77,
558
+ "learning_rate": 1.3379778492318687e-05,
559
+ "loss": 1.172,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 3.81,
564
+ "learning_rate": 1.293319042515184e-05,
565
+ "loss": 1.1656,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 3.85,
570
+ "learning_rate": 1.2486602357984996e-05,
571
+ "loss": 1.1658,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 3.89,
576
+ "learning_rate": 1.2040014290818149e-05,
577
+ "loss": 1.1656,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 3.94,
582
+ "learning_rate": 1.1593426223651305e-05,
583
+ "loss": 1.1576,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 3.98,
588
+ "learning_rate": 1.114683815648446e-05,
589
+ "loss": 1.1572,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 4.02,
594
+ "learning_rate": 1.0700250089317613e-05,
595
+ "loss": 1.1591,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 4.06,
600
+ "learning_rate": 1.0253662022150768e-05,
601
+ "loss": 1.1567,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 4.1,
606
+ "learning_rate": 9.807073954983924e-06,
607
+ "loss": 1.153,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 4.14,
612
+ "learning_rate": 9.360485887817077e-06,
613
+ "loss": 1.1554,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 4.18,
618
+ "learning_rate": 8.913897820650232e-06,
619
+ "loss": 1.1463,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 4.22,
624
+ "learning_rate": 8.467309753483389e-06,
625
+ "loss": 1.1527,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 4.26,
630
+ "learning_rate": 8.020721686316542e-06,
631
+ "loss": 1.15,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 4.3,
636
+ "learning_rate": 7.574133619149697e-06,
637
+ "loss": 1.1503,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 4.35,
642
+ "learning_rate": 7.127545551982852e-06,
643
+ "loss": 1.1455,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 4.39,
648
+ "learning_rate": 6.680957484816006e-06,
649
+ "loss": 1.1473,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 4.43,
654
+ "learning_rate": 6.2343694176491604e-06,
655
+ "loss": 1.1453,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 4.47,
660
+ "learning_rate": 5.787781350482315e-06,
661
+ "loss": 1.1441,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 4.51,
666
+ "learning_rate": 5.34119328331547e-06,
667
+ "loss": 1.1399,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 4.55,
672
+ "learning_rate": 4.894605216148625e-06,
673
+ "loss": 1.1354,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 4.59,
678
+ "learning_rate": 4.448017148981779e-06,
679
+ "loss": 1.1371,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 4.63,
684
+ "learning_rate": 4.001429081814934e-06,
685
+ "loss": 1.1399,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 4.67,
690
+ "learning_rate": 3.554841014648089e-06,
691
+ "loss": 1.1317,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 4.71,
696
+ "learning_rate": 3.1082529474812435e-06,
697
+ "loss": 1.1412,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 4.76,
702
+ "learning_rate": 2.6616648803143982e-06,
703
+ "loss": 1.1368,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 4.8,
708
+ "learning_rate": 2.2150768131475525e-06,
709
+ "loss": 1.1345,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 4.84,
714
+ "learning_rate": 1.7684887459807075e-06,
715
+ "loss": 1.1316,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 4.88,
720
+ "learning_rate": 1.3219006788138622e-06,
721
+ "loss": 1.1362,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 4.92,
726
+ "learning_rate": 8.753126116470169e-07,
727
+ "loss": 1.1358,
728
+ "step": 60000
729
+ }
730
+ ],
731
+ "max_steps": 60980,
732
+ "num_train_epochs": 5,
733
+ "total_flos": 5.053516613471232e+17,
734
+ "trial_name": null,
735
+ "trial_params": null
736
+ }