Davlan commited on
Commit
f30098f
1 Parent(s): e2e5dac

Upload trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +1393 -0
trainer_state.json ADDED
@@ -0,0 +1,1393 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 114399,
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": 4.9781466621211725e-05,
13
+ "loss": 2.1728,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 4.956293324242345e-05,
19
+ "loss": 2.0108,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 4.934439986363518e-05,
25
+ "loss": 1.9521,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 4.91258664848469e-05,
31
+ "loss": 1.8771,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.07,
36
+ "learning_rate": 4.8907333106058624e-05,
37
+ "loss": 1.8337,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 4.868879972727035e-05,
43
+ "loss": 1.8045,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.09,
48
+ "learning_rate": 4.847026634848207e-05,
49
+ "loss": 1.771,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.1,
54
+ "learning_rate": 4.825173296969379e-05,
55
+ "loss": 1.7447,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.12,
60
+ "learning_rate": 4.803319959090552e-05,
61
+ "loss": 1.7067,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.13,
66
+ "learning_rate": 4.7814666212117245e-05,
67
+ "loss": 1.6988,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.14,
72
+ "learning_rate": 4.759613283332897e-05,
73
+ "loss": 1.676,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.16,
78
+ "learning_rate": 4.737759945454069e-05,
79
+ "loss": 1.6483,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.17,
84
+ "learning_rate": 4.7159066075752414e-05,
85
+ "loss": 1.638,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.18,
90
+ "learning_rate": 4.694053269696414e-05,
91
+ "loss": 1.6235,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.2,
96
+ "learning_rate": 4.6721999318175866e-05,
97
+ "loss": 1.6096,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.21,
102
+ "learning_rate": 4.650346593938759e-05,
103
+ "loss": 1.6011,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.22,
108
+ "learning_rate": 4.628493256059931e-05,
109
+ "loss": 1.5781,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.24,
114
+ "learning_rate": 4.6066399181811035e-05,
115
+ "loss": 1.572,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.25,
120
+ "learning_rate": 4.584786580302276e-05,
121
+ "loss": 1.5612,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.26,
126
+ "learning_rate": 4.562933242423448e-05,
127
+ "loss": 1.5728,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.28,
132
+ "learning_rate": 4.5410799045446204e-05,
133
+ "loss": 1.5437,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.29,
138
+ "learning_rate": 4.519226566665793e-05,
139
+ "loss": 1.5245,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.3,
144
+ "learning_rate": 4.497373228786965e-05,
145
+ "loss": 1.5209,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.31,
150
+ "learning_rate": 4.475519890908137e-05,
151
+ "loss": 1.5211,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.33,
156
+ "learning_rate": 4.4536665530293096e-05,
157
+ "loss": 1.5035,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.34,
162
+ "learning_rate": 4.431813215150482e-05,
163
+ "loss": 1.4973,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.35,
168
+ "learning_rate": 4.409959877271655e-05,
169
+ "loss": 1.4881,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.37,
174
+ "learning_rate": 4.388106539392827e-05,
175
+ "loss": 1.474,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.38,
180
+ "learning_rate": 4.3662532015139994e-05,
181
+ "loss": 1.4652,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.39,
186
+ "learning_rate": 4.344399863635172e-05,
187
+ "loss": 1.4629,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.41,
192
+ "learning_rate": 4.322546525756344e-05,
193
+ "loss": 1.4809,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.42,
198
+ "learning_rate": 4.300693187877516e-05,
199
+ "loss": 1.4629,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.43,
204
+ "learning_rate": 4.278839849998689e-05,
205
+ "loss": 1.4502,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.45,
210
+ "learning_rate": 4.2569865121198615e-05,
211
+ "loss": 1.4361,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.46,
216
+ "learning_rate": 4.235133174241034e-05,
217
+ "loss": 1.4425,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.47,
222
+ "learning_rate": 4.213279836362206e-05,
223
+ "loss": 1.422,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.49,
228
+ "learning_rate": 4.1914264984833784e-05,
229
+ "loss": 1.4164,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.5,
234
+ "learning_rate": 4.169573160604551e-05,
235
+ "loss": 1.4208,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.51,
240
+ "learning_rate": 4.147719822725724e-05,
241
+ "loss": 1.4204,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.52,
246
+ "learning_rate": 4.125866484846896e-05,
247
+ "loss": 1.3949,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.54,
252
+ "learning_rate": 4.104013146968068e-05,
253
+ "loss": 1.4103,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.55,
258
+ "learning_rate": 4.0821598090892405e-05,
259
+ "loss": 1.4015,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.56,
264
+ "learning_rate": 4.060306471210413e-05,
265
+ "loss": 1.3844,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.58,
270
+ "learning_rate": 4.038453133331585e-05,
271
+ "loss": 1.3823,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.59,
276
+ "learning_rate": 4.016599795452758e-05,
277
+ "loss": 1.3842,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.6,
282
+ "learning_rate": 3.9947464575739304e-05,
283
+ "loss": 1.373,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.62,
288
+ "learning_rate": 3.972893119695103e-05,
289
+ "loss": 1.3756,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.63,
294
+ "learning_rate": 3.951039781816275e-05,
295
+ "loss": 1.3714,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.64,
300
+ "learning_rate": 3.929186443937447e-05,
301
+ "loss": 1.3614,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.66,
306
+ "learning_rate": 3.9073331060586195e-05,
307
+ "loss": 1.3703,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.67,
312
+ "learning_rate": 3.8854797681797925e-05,
313
+ "loss": 1.3494,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.68,
318
+ "learning_rate": 3.863626430300965e-05,
319
+ "loss": 1.3588,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.69,
324
+ "learning_rate": 3.841773092422137e-05,
325
+ "loss": 1.3514,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.71,
330
+ "learning_rate": 3.8199197545433094e-05,
331
+ "loss": 1.3591,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.72,
336
+ "learning_rate": 3.798066416664482e-05,
337
+ "loss": 1.3479,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.73,
342
+ "learning_rate": 3.776213078785654e-05,
343
+ "loss": 1.3524,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.75,
348
+ "learning_rate": 3.754359740906826e-05,
349
+ "loss": 1.3366,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.76,
354
+ "learning_rate": 3.7325064030279986e-05,
355
+ "loss": 1.3498,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.77,
360
+ "learning_rate": 3.710653065149171e-05,
361
+ "loss": 1.3305,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.79,
366
+ "learning_rate": 3.688799727270343e-05,
367
+ "loss": 1.3263,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.8,
372
+ "learning_rate": 3.6669463893915154e-05,
373
+ "loss": 1.3203,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.81,
378
+ "learning_rate": 3.645093051512688e-05,
379
+ "loss": 1.3128,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.83,
384
+ "learning_rate": 3.623239713633861e-05,
385
+ "loss": 1.3157,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.84,
390
+ "learning_rate": 3.601386375755033e-05,
391
+ "loss": 1.3167,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.85,
396
+ "learning_rate": 3.579533037876205e-05,
397
+ "loss": 1.3075,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.87,
402
+ "learning_rate": 3.5576796999973776e-05,
403
+ "loss": 1.3077,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.88,
408
+ "learning_rate": 3.53582636211855e-05,
409
+ "loss": 1.3082,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.89,
414
+ "learning_rate": 3.513973024239722e-05,
415
+ "loss": 1.3128,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.9,
420
+ "learning_rate": 3.492119686360895e-05,
421
+ "loss": 1.3122,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.92,
426
+ "learning_rate": 3.4702663484820674e-05,
427
+ "loss": 1.2951,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.93,
432
+ "learning_rate": 3.44841301060324e-05,
433
+ "loss": 1.2928,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.94,
438
+ "learning_rate": 3.426559672724412e-05,
439
+ "loss": 1.2889,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.96,
444
+ "learning_rate": 3.404706334845584e-05,
445
+ "loss": 1.2825,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.97,
450
+ "learning_rate": 3.3828529969667566e-05,
451
+ "loss": 1.2809,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.98,
456
+ "learning_rate": 3.3609996590879295e-05,
457
+ "loss": 1.2806,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 1.0,
462
+ "learning_rate": 3.339146321209102e-05,
463
+ "loss": 1.2847,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 1.01,
468
+ "learning_rate": 3.317292983330274e-05,
469
+ "loss": 1.2637,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 1.02,
474
+ "learning_rate": 3.2954396454514464e-05,
475
+ "loss": 1.2698,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 1.04,
480
+ "learning_rate": 3.273586307572619e-05,
481
+ "loss": 1.2469,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 1.05,
486
+ "learning_rate": 3.251732969693791e-05,
487
+ "loss": 1.2516,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 1.06,
492
+ "learning_rate": 3.229879631814964e-05,
493
+ "loss": 1.2525,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 1.08,
498
+ "learning_rate": 3.208026293936136e-05,
499
+ "loss": 1.2573,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 1.09,
504
+ "learning_rate": 3.1861729560573085e-05,
505
+ "loss": 1.2516,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 1.1,
510
+ "learning_rate": 3.164319618178481e-05,
511
+ "loss": 1.2425,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 1.11,
516
+ "learning_rate": 3.142466280299653e-05,
517
+ "loss": 1.2487,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 1.13,
522
+ "learning_rate": 3.1206129424208254e-05,
523
+ "loss": 1.238,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 1.14,
528
+ "learning_rate": 3.0987596045419984e-05,
529
+ "loss": 1.2293,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 1.15,
534
+ "learning_rate": 3.076906266663171e-05,
535
+ "loss": 1.2392,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 1.17,
540
+ "learning_rate": 3.055052928784343e-05,
541
+ "loss": 1.2382,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 1.18,
546
+ "learning_rate": 3.033199590905515e-05,
547
+ "loss": 1.2327,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 1.19,
552
+ "learning_rate": 3.0113462530266872e-05,
553
+ "loss": 1.2345,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 1.21,
558
+ "learning_rate": 2.9894929151478595e-05,
559
+ "loss": 1.2214,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 1.22,
564
+ "learning_rate": 2.9676395772690325e-05,
565
+ "loss": 1.2356,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 1.23,
570
+ "learning_rate": 2.9457862393902048e-05,
571
+ "loss": 1.2336,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 1.25,
576
+ "learning_rate": 2.923932901511377e-05,
577
+ "loss": 1.2354,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 1.26,
582
+ "learning_rate": 2.9020795636325493e-05,
583
+ "loss": 1.2396,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 1.27,
588
+ "learning_rate": 2.8802262257537216e-05,
589
+ "loss": 1.2234,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 1.28,
594
+ "learning_rate": 2.858372887874894e-05,
595
+ "loss": 1.2239,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 1.3,
600
+ "learning_rate": 2.8365195499960666e-05,
601
+ "loss": 1.2123,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 1.31,
606
+ "learning_rate": 2.8146662121172392e-05,
607
+ "loss": 1.2116,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 1.32,
612
+ "learning_rate": 2.7928128742384115e-05,
613
+ "loss": 1.2139,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 1.34,
618
+ "learning_rate": 2.7709595363595838e-05,
619
+ "loss": 1.2115,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 1.35,
624
+ "learning_rate": 2.749106198480756e-05,
625
+ "loss": 1.2051,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 1.36,
630
+ "learning_rate": 2.7272528606019284e-05,
631
+ "loss": 1.2152,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 1.38,
636
+ "learning_rate": 2.705399522723101e-05,
637
+ "loss": 1.2084,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 1.39,
642
+ "learning_rate": 2.6835461848442733e-05,
643
+ "loss": 1.1987,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 1.4,
648
+ "learning_rate": 2.6616928469654456e-05,
649
+ "loss": 1.1892,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 1.42,
654
+ "learning_rate": 2.639839509086618e-05,
655
+ "loss": 1.2074,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 1.43,
660
+ "learning_rate": 2.61798617120779e-05,
661
+ "loss": 1.1952,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 1.44,
666
+ "learning_rate": 2.596132833328963e-05,
667
+ "loss": 1.196,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 1.46,
672
+ "learning_rate": 2.5742794954501354e-05,
673
+ "loss": 1.182,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 1.47,
678
+ "learning_rate": 2.5524261575713077e-05,
679
+ "loss": 1.1845,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 1.48,
684
+ "learning_rate": 2.53057281969248e-05,
685
+ "loss": 1.1889,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 1.49,
690
+ "learning_rate": 2.5087194818136523e-05,
691
+ "loss": 1.177,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 1.51,
696
+ "learning_rate": 2.486866143934825e-05,
697
+ "loss": 1.1845,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 1.52,
702
+ "learning_rate": 2.4650128060559972e-05,
703
+ "loss": 1.1865,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 1.53,
708
+ "learning_rate": 2.4431594681771695e-05,
709
+ "loss": 1.1766,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 1.55,
714
+ "learning_rate": 2.421306130298342e-05,
715
+ "loss": 1.1784,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 1.56,
720
+ "learning_rate": 2.3994527924195144e-05,
721
+ "loss": 1.1728,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 1.57,
726
+ "learning_rate": 2.3775994545406867e-05,
727
+ "loss": 1.1775,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 1.59,
732
+ "learning_rate": 2.355746116661859e-05,
733
+ "loss": 1.1767,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 1.6,
738
+ "learning_rate": 2.3338927787830313e-05,
739
+ "loss": 1.1809,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 1.61,
744
+ "learning_rate": 2.3120394409042036e-05,
745
+ "loss": 1.1744,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 1.63,
750
+ "learning_rate": 2.2901861030253762e-05,
751
+ "loss": 1.1694,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 1.64,
756
+ "learning_rate": 2.2683327651465485e-05,
757
+ "loss": 1.1568,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 1.65,
762
+ "learning_rate": 2.2464794272677208e-05,
763
+ "loss": 1.158,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 1.67,
768
+ "learning_rate": 2.2246260893888934e-05,
769
+ "loss": 1.1526,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 1.68,
774
+ "learning_rate": 2.2027727515100657e-05,
775
+ "loss": 1.1663,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 1.69,
780
+ "learning_rate": 2.180919413631238e-05,
781
+ "loss": 1.1625,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 1.7,
786
+ "learning_rate": 2.1590660757524106e-05,
787
+ "loss": 1.1565,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 1.72,
792
+ "learning_rate": 2.137212737873583e-05,
793
+ "loss": 1.1625,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 1.73,
798
+ "learning_rate": 2.1153593999947552e-05,
799
+ "loss": 1.1556,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 1.74,
804
+ "learning_rate": 2.093506062115928e-05,
805
+ "loss": 1.1553,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 1.76,
810
+ "learning_rate": 2.0716527242371e-05,
811
+ "loss": 1.1529,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 1.77,
816
+ "learning_rate": 2.0497993863582724e-05,
817
+ "loss": 1.1503,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 1.78,
822
+ "learning_rate": 2.027946048479445e-05,
823
+ "loss": 1.1627,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 1.8,
828
+ "learning_rate": 2.0060927106006174e-05,
829
+ "loss": 1.1392,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 1.81,
834
+ "learning_rate": 1.9842393727217896e-05,
835
+ "loss": 1.1399,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 1.82,
840
+ "learning_rate": 1.962386034842962e-05,
841
+ "loss": 1.1443,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 1.84,
846
+ "learning_rate": 1.9405326969641342e-05,
847
+ "loss": 1.129,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 1.85,
852
+ "learning_rate": 1.9186793590853065e-05,
853
+ "loss": 1.1355,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 1.86,
858
+ "learning_rate": 1.896826021206479e-05,
859
+ "loss": 1.1401,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 1.88,
864
+ "learning_rate": 1.8749726833276514e-05,
865
+ "loss": 1.1303,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 1.89,
870
+ "learning_rate": 1.8531193454488237e-05,
871
+ "loss": 1.1327,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 1.9,
876
+ "learning_rate": 1.8312660075699964e-05,
877
+ "loss": 1.1316,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 1.91,
882
+ "learning_rate": 1.8094126696911687e-05,
883
+ "loss": 1.1488,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 1.93,
888
+ "learning_rate": 1.787559331812341e-05,
889
+ "loss": 1.1397,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 1.94,
894
+ "learning_rate": 1.7657059939335136e-05,
895
+ "loss": 1.129,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 1.95,
900
+ "learning_rate": 1.743852656054686e-05,
901
+ "loss": 1.1278,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 1.97,
906
+ "learning_rate": 1.721999318175858e-05,
907
+ "loss": 1.1146,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 1.98,
912
+ "learning_rate": 1.7001459802970308e-05,
913
+ "loss": 1.1195,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 1.99,
918
+ "learning_rate": 1.678292642418203e-05,
919
+ "loss": 1.1094,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 2.01,
924
+ "learning_rate": 1.6564393045393754e-05,
925
+ "loss": 1.1165,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 2.02,
930
+ "learning_rate": 1.634585966660548e-05,
931
+ "loss": 1.103,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 2.03,
936
+ "learning_rate": 1.6127326287817203e-05,
937
+ "loss": 1.105,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 2.05,
942
+ "learning_rate": 1.5908792909028926e-05,
943
+ "loss": 1.0976,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 2.06,
948
+ "learning_rate": 1.569025953024065e-05,
949
+ "loss": 1.1125,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 2.07,
954
+ "learning_rate": 1.547172615145237e-05,
955
+ "loss": 1.0944,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 2.08,
960
+ "learning_rate": 1.5253192772664096e-05,
961
+ "loss": 1.1034,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 2.1,
966
+ "learning_rate": 1.5034659393875823e-05,
967
+ "loss": 1.0957,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 2.11,
972
+ "learning_rate": 1.4816126015087545e-05,
973
+ "loss": 1.1147,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 2.12,
978
+ "learning_rate": 1.459759263629927e-05,
979
+ "loss": 1.0976,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 2.14,
984
+ "learning_rate": 1.4379059257510993e-05,
985
+ "loss": 1.1009,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 2.15,
990
+ "learning_rate": 1.4160525878722716e-05,
991
+ "loss": 1.0959,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 2.16,
996
+ "learning_rate": 1.3941992499934442e-05,
997
+ "loss": 1.085,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 2.18,
1002
+ "learning_rate": 1.3723459121146165e-05,
1003
+ "loss": 1.0898,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 2.19,
1008
+ "learning_rate": 1.3504925742357888e-05,
1009
+ "loss": 1.0855,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 2.2,
1014
+ "learning_rate": 1.3286392363569614e-05,
1015
+ "loss": 1.0875,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 2.22,
1020
+ "learning_rate": 1.3067858984781337e-05,
1021
+ "loss": 1.0904,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 2.23,
1026
+ "learning_rate": 1.284932560599306e-05,
1027
+ "loss": 1.0889,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 2.24,
1032
+ "learning_rate": 1.2630792227204785e-05,
1033
+ "loss": 1.0947,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 2.26,
1038
+ "learning_rate": 1.2412258848416508e-05,
1039
+ "loss": 1.0895,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 2.27,
1044
+ "learning_rate": 1.219372546962823e-05,
1045
+ "loss": 1.084,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 2.28,
1050
+ "learning_rate": 1.1975192090839955e-05,
1051
+ "loss": 1.0783,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 2.29,
1056
+ "learning_rate": 1.175665871205168e-05,
1057
+ "loss": 1.0837,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 2.31,
1062
+ "learning_rate": 1.1538125333263403e-05,
1063
+ "loss": 1.0832,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 2.32,
1068
+ "learning_rate": 1.1319591954475127e-05,
1069
+ "loss": 1.0814,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 2.33,
1074
+ "learning_rate": 1.1101058575686852e-05,
1075
+ "loss": 1.0842,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 2.35,
1080
+ "learning_rate": 1.0882525196898575e-05,
1081
+ "loss": 1.0766,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 2.36,
1086
+ "learning_rate": 1.0663991818110298e-05,
1087
+ "loss": 1.0814,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 2.37,
1092
+ "learning_rate": 1.0445458439322022e-05,
1093
+ "loss": 1.0742,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 2.39,
1098
+ "learning_rate": 1.0226925060533745e-05,
1099
+ "loss": 1.0761,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 2.4,
1104
+ "learning_rate": 1.000839168174547e-05,
1105
+ "loss": 1.0693,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 2.41,
1110
+ "learning_rate": 9.789858302957194e-06,
1111
+ "loss": 1.0651,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 2.43,
1116
+ "learning_rate": 9.571324924168917e-06,
1117
+ "loss": 1.0774,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 2.44,
1122
+ "learning_rate": 9.352791545380642e-06,
1123
+ "loss": 1.0595,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 2.45,
1128
+ "learning_rate": 9.134258166592367e-06,
1129
+ "loss": 1.0623,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 2.47,
1134
+ "learning_rate": 8.91572478780409e-06,
1135
+ "loss": 1.0658,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 2.48,
1140
+ "learning_rate": 8.697191409015812e-06,
1141
+ "loss": 1.0673,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 2.49,
1146
+ "learning_rate": 8.478658030227537e-06,
1147
+ "loss": 1.063,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 2.5,
1152
+ "learning_rate": 8.260124651439262e-06,
1153
+ "loss": 1.0639,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 2.52,
1158
+ "learning_rate": 8.041591272650985e-06,
1159
+ "loss": 1.0542,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 2.53,
1164
+ "learning_rate": 7.823057893862709e-06,
1165
+ "loss": 1.0651,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 2.54,
1170
+ "learning_rate": 7.604524515074434e-06,
1171
+ "loss": 1.0594,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 2.56,
1176
+ "learning_rate": 7.385991136286157e-06,
1177
+ "loss": 1.0646,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 2.57,
1182
+ "learning_rate": 7.16745775749788e-06,
1183
+ "loss": 1.0554,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 2.58,
1188
+ "learning_rate": 6.948924378709605e-06,
1189
+ "loss": 1.0653,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 2.6,
1194
+ "learning_rate": 6.730390999921328e-06,
1195
+ "loss": 1.0483,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 2.61,
1200
+ "learning_rate": 6.5118576211330525e-06,
1201
+ "loss": 1.063,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 2.62,
1206
+ "learning_rate": 6.293324242344776e-06,
1207
+ "loss": 1.0314,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 2.64,
1212
+ "learning_rate": 6.074790863556499e-06,
1213
+ "loss": 1.0514,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 2.65,
1218
+ "learning_rate": 5.856257484768224e-06,
1219
+ "loss": 1.044,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 2.66,
1224
+ "learning_rate": 5.637724105979948e-06,
1225
+ "loss": 1.0417,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 2.67,
1230
+ "learning_rate": 5.419190727191672e-06,
1231
+ "loss": 1.0528,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 2.69,
1236
+ "learning_rate": 5.200657348403395e-06,
1237
+ "loss": 1.0521,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 2.7,
1242
+ "learning_rate": 4.982123969615119e-06,
1243
+ "loss": 1.0451,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 2.71,
1248
+ "learning_rate": 4.7635905908268435e-06,
1249
+ "loss": 1.0449,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 2.73,
1254
+ "learning_rate": 4.545057212038567e-06,
1255
+ "loss": 1.0481,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 2.74,
1260
+ "learning_rate": 4.326523833250291e-06,
1261
+ "loss": 1.044,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 2.75,
1266
+ "learning_rate": 4.107990454462015e-06,
1267
+ "loss": 1.0465,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 2.77,
1272
+ "learning_rate": 3.8894570756737385e-06,
1273
+ "loss": 1.0466,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 2.78,
1278
+ "learning_rate": 3.6709236968854623e-06,
1279
+ "loss": 1.0337,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 2.79,
1284
+ "learning_rate": 3.4523903180971864e-06,
1285
+ "loss": 1.0418,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 2.81,
1290
+ "learning_rate": 3.23385693930891e-06,
1291
+ "loss": 1.0527,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 2.82,
1296
+ "learning_rate": 3.0153235605206344e-06,
1297
+ "loss": 1.0388,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 2.83,
1302
+ "learning_rate": 2.7967901817323577e-06,
1303
+ "loss": 1.0464,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 2.85,
1308
+ "learning_rate": 2.578256802944082e-06,
1309
+ "loss": 1.0339,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 2.86,
1314
+ "learning_rate": 2.3597234241558057e-06,
1315
+ "loss": 1.0395,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 2.87,
1320
+ "learning_rate": 2.1411900453675294e-06,
1321
+ "loss": 1.0362,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 2.88,
1326
+ "learning_rate": 1.922656666579253e-06,
1327
+ "loss": 1.0408,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 2.9,
1332
+ "learning_rate": 1.7041232877909774e-06,
1333
+ "loss": 1.0354,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 2.91,
1338
+ "learning_rate": 1.4855899090027011e-06,
1339
+ "loss": 1.0315,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 2.92,
1344
+ "learning_rate": 1.2670565302144249e-06,
1345
+ "loss": 1.0385,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 2.94,
1350
+ "learning_rate": 1.0485231514261489e-06,
1351
+ "loss": 1.0426,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 2.95,
1356
+ "learning_rate": 8.299897726378727e-07,
1357
+ "loss": 1.0348,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 2.96,
1362
+ "learning_rate": 6.114563938495966e-07,
1363
+ "loss": 1.0484,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 2.98,
1368
+ "learning_rate": 3.929230150613205e-07,
1369
+ "loss": 1.0376,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 2.99,
1374
+ "learning_rate": 1.7438963627304437e-07,
1375
+ "loss": 1.037,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 3.0,
1380
+ "step": 114399,
1381
+ "total_flos": 1.0669987784955003e+18,
1382
+ "train_loss": 1.2423339333696508,
1383
+ "train_runtime": 126615.6451,
1384
+ "train_samples_per_second": 9.035,
1385
+ "train_steps_per_second": 0.904
1386
+ }
1387
+ ],
1388
+ "max_steps": 114399,
1389
+ "num_train_epochs": 3,
1390
+ "total_flos": 1.0669987784955003e+18,
1391
+ "trial_name": null,
1392
+ "trial_params": null
1393
+ }