nid989 commited on
Commit
d000461
1 Parent(s): 02349b9

Upload trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +1534 -0
trainer_state.json ADDED
@@ -0,0 +1,1534 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9976812782939256,
5
+ "global_step": 126500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 1.9921131914759374e-05,
13
+ "loss": 1.2922,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 1.984226382951875e-05,
19
+ "loss": 1.3127,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 1.976339574427812e-05,
25
+ "loss": 1.293,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 1.9684527659037496e-05,
31
+ "loss": 1.2856,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.02,
36
+ "learning_rate": 1.9605659573796868e-05,
37
+ "loss": 1.3064,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.02,
42
+ "learning_rate": 1.9526791488556244e-05,
43
+ "loss": 1.3068,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.03,
48
+ "learning_rate": 1.9447923403315616e-05,
49
+ "loss": 1.2982,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.03,
54
+ "learning_rate": 1.936905531807499e-05,
55
+ "loss": 1.303,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 1.9290187232834363e-05,
61
+ "loss": 1.3007,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.04,
66
+ "learning_rate": 1.9211319147593738e-05,
67
+ "loss": 1.3172,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.04,
72
+ "learning_rate": 1.913245106235311e-05,
73
+ "loss": 1.325,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.05,
78
+ "learning_rate": 1.9053582977112482e-05,
79
+ "loss": 1.2849,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.05,
84
+ "learning_rate": 1.8974714891871857e-05,
85
+ "loss": 1.3189,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.06,
90
+ "learning_rate": 1.889584680663123e-05,
91
+ "loss": 1.2955,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.06,
96
+ "learning_rate": 1.8816978721390605e-05,
97
+ "loss": 1.3496,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.06,
102
+ "learning_rate": 1.8738110636149977e-05,
103
+ "loss": 1.3103,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.07,
108
+ "learning_rate": 1.8659242550909352e-05,
109
+ "loss": 1.2921,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.07,
114
+ "learning_rate": 1.8580374465668724e-05,
115
+ "loss": 1.303,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.07,
120
+ "learning_rate": 1.85015063804281e-05,
121
+ "loss": 1.2997,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.08,
126
+ "learning_rate": 1.842263829518747e-05,
127
+ "loss": 1.2692,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.08,
132
+ "learning_rate": 1.8343770209946847e-05,
133
+ "loss": 1.2915,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.09,
138
+ "learning_rate": 1.826490212470622e-05,
139
+ "loss": 1.2907,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.09,
144
+ "learning_rate": 1.818603403946559e-05,
145
+ "loss": 1.3233,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.09,
150
+ "learning_rate": 1.8107165954224966e-05,
151
+ "loss": 1.3047,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.1,
156
+ "learning_rate": 1.8028297868984338e-05,
157
+ "loss": 1.273,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.1,
162
+ "learning_rate": 1.7949429783743713e-05,
163
+ "loss": 1.3118,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.11,
168
+ "learning_rate": 1.7870561698503085e-05,
169
+ "loss": 1.3183,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.11,
174
+ "learning_rate": 1.779169361326246e-05,
175
+ "loss": 1.304,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.11,
180
+ "learning_rate": 1.7712825528021833e-05,
181
+ "loss": 1.3409,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.12,
186
+ "learning_rate": 1.7633957442781208e-05,
187
+ "loss": 1.3068,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.12,
192
+ "learning_rate": 1.755508935754058e-05,
193
+ "loss": 1.3038,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.13,
198
+ "learning_rate": 1.7476221272299955e-05,
199
+ "loss": 1.3331,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.13,
204
+ "learning_rate": 1.7397353187059327e-05,
205
+ "loss": 1.2997,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.13,
210
+ "learning_rate": 1.73184851018187e-05,
211
+ "loss": 1.2833,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.14,
216
+ "learning_rate": 1.723961701657807e-05,
217
+ "loss": 1.3178,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.14,
222
+ "learning_rate": 1.7160748931337447e-05,
223
+ "loss": 1.3263,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.15,
228
+ "learning_rate": 1.708188084609682e-05,
229
+ "loss": 1.2907,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.15,
234
+ "learning_rate": 1.7003012760856194e-05,
235
+ "loss": 1.3328,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.15,
240
+ "learning_rate": 1.6924144675615566e-05,
241
+ "loss": 1.3268,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.16,
246
+ "learning_rate": 1.6845276590374938e-05,
247
+ "loss": 1.3161,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.16,
252
+ "learning_rate": 1.6766408505134313e-05,
253
+ "loss": 1.3026,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.17,
258
+ "learning_rate": 1.6687540419893685e-05,
259
+ "loss": 1.3163,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.17,
264
+ "learning_rate": 1.660867233465306e-05,
265
+ "loss": 1.3202,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.17,
270
+ "learning_rate": 1.6529804249412432e-05,
271
+ "loss": 1.2872,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.18,
276
+ "learning_rate": 1.6450936164171808e-05,
277
+ "loss": 1.3209,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.18,
282
+ "learning_rate": 1.637206807893118e-05,
283
+ "loss": 1.3165,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.19,
288
+ "learning_rate": 1.6293199993690555e-05,
289
+ "loss": 1.3026,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.19,
294
+ "learning_rate": 1.6214331908449927e-05,
295
+ "loss": 1.3313,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.19,
300
+ "learning_rate": 1.6135463823209302e-05,
301
+ "loss": 1.3123,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.2,
306
+ "learning_rate": 1.6056595737968674e-05,
307
+ "loss": 1.3124,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.2,
312
+ "learning_rate": 1.5977727652728046e-05,
313
+ "loss": 1.3227,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.21,
318
+ "learning_rate": 1.589885956748742e-05,
319
+ "loss": 1.316,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.21,
324
+ "learning_rate": 1.5819991482246794e-05,
325
+ "loss": 1.2764,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.21,
330
+ "learning_rate": 1.574112339700617e-05,
331
+ "loss": 1.3096,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.22,
336
+ "learning_rate": 1.566225531176554e-05,
337
+ "loss": 1.3499,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.22,
342
+ "learning_rate": 1.5583387226524916e-05,
343
+ "loss": 1.3339,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.22,
348
+ "learning_rate": 1.5504519141284288e-05,
349
+ "loss": 1.3423,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.23,
354
+ "learning_rate": 1.5425651056043664e-05,
355
+ "loss": 1.3295,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.23,
360
+ "learning_rate": 1.5346782970803036e-05,
361
+ "loss": 1.3226,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.24,
366
+ "learning_rate": 1.526791488556241e-05,
367
+ "loss": 1.2861,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.24,
372
+ "learning_rate": 1.5189046800321783e-05,
373
+ "loss": 1.3504,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.24,
378
+ "learning_rate": 1.5110178715081157e-05,
379
+ "loss": 1.3619,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.25,
384
+ "learning_rate": 1.503131062984053e-05,
385
+ "loss": 1.329,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.25,
390
+ "learning_rate": 1.4952442544599904e-05,
391
+ "loss": 1.2826,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.26,
396
+ "learning_rate": 1.4873574459359277e-05,
397
+ "loss": 1.3075,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.26,
402
+ "learning_rate": 1.4794706374118651e-05,
403
+ "loss": 1.3293,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.26,
408
+ "learning_rate": 1.4715838288878023e-05,
409
+ "loss": 1.2638,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.27,
414
+ "learning_rate": 1.4636970203637397e-05,
415
+ "loss": 1.3282,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.27,
420
+ "learning_rate": 1.455810211839677e-05,
421
+ "loss": 1.3562,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.28,
426
+ "learning_rate": 1.4479234033156144e-05,
427
+ "loss": 1.3446,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.28,
432
+ "learning_rate": 1.4400365947915518e-05,
433
+ "loss": 1.3339,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.28,
438
+ "learning_rate": 1.4321497862674891e-05,
439
+ "loss": 1.31,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.29,
444
+ "learning_rate": 1.4242629777434265e-05,
445
+ "loss": 1.3224,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.29,
450
+ "learning_rate": 1.4163761692193639e-05,
451
+ "loss": 1.3715,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.3,
456
+ "learning_rate": 1.4084893606953012e-05,
457
+ "loss": 1.3081,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 0.3,
462
+ "learning_rate": 1.4006025521712386e-05,
463
+ "loss": 1.2892,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 0.3,
468
+ "learning_rate": 1.3927157436471758e-05,
469
+ "loss": 1.3046,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 0.31,
474
+ "learning_rate": 1.3848289351231132e-05,
475
+ "loss": 1.3344,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 0.31,
480
+ "learning_rate": 1.3769421265990505e-05,
481
+ "loss": 1.3303,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 0.32,
486
+ "learning_rate": 1.3690553180749879e-05,
487
+ "loss": 1.3237,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 0.32,
492
+ "learning_rate": 1.3611685095509253e-05,
493
+ "loss": 1.3162,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 0.32,
498
+ "learning_rate": 1.3532817010268626e-05,
499
+ "loss": 1.3496,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 0.33,
504
+ "learning_rate": 1.3453948925028e-05,
505
+ "loss": 1.355,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 0.33,
510
+ "learning_rate": 1.3375080839787374e-05,
511
+ "loss": 1.348,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 0.34,
516
+ "learning_rate": 1.3296212754546747e-05,
517
+ "loss": 1.344,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 0.34,
522
+ "learning_rate": 1.3217344669306121e-05,
523
+ "loss": 1.3256,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 0.34,
528
+ "learning_rate": 1.3138476584065495e-05,
529
+ "loss": 1.3724,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 0.35,
534
+ "learning_rate": 1.3059608498824867e-05,
535
+ "loss": 1.2848,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 0.35,
540
+ "learning_rate": 1.298074041358424e-05,
541
+ "loss": 1.3318,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 0.35,
546
+ "learning_rate": 1.2901872328343614e-05,
547
+ "loss": 1.3025,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 0.36,
552
+ "learning_rate": 1.2823004243102987e-05,
553
+ "loss": 1.2953,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 0.36,
558
+ "learning_rate": 1.2744136157862361e-05,
559
+ "loss": 1.3388,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 0.37,
564
+ "learning_rate": 1.2665268072621735e-05,
565
+ "loss": 1.3444,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 0.37,
570
+ "learning_rate": 1.2586399987381108e-05,
571
+ "loss": 1.3239,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 0.37,
576
+ "learning_rate": 1.2507531902140482e-05,
577
+ "loss": 1.296,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 0.38,
582
+ "learning_rate": 1.2428663816899856e-05,
583
+ "loss": 1.3024,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 0.38,
588
+ "learning_rate": 1.234979573165923e-05,
589
+ "loss": 1.33,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 0.39,
594
+ "learning_rate": 1.2270927646418603e-05,
595
+ "loss": 1.3103,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 0.39,
600
+ "learning_rate": 1.2192059561177975e-05,
601
+ "loss": 1.3585,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 0.39,
606
+ "learning_rate": 1.2113191475937349e-05,
607
+ "loss": 1.3155,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 0.4,
612
+ "learning_rate": 1.2034323390696722e-05,
613
+ "loss": 1.3307,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 0.4,
618
+ "learning_rate": 1.1955455305456094e-05,
619
+ "loss": 1.3344,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 0.41,
624
+ "learning_rate": 1.1876587220215468e-05,
625
+ "loss": 1.3154,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 0.41,
630
+ "learning_rate": 1.1797719134974842e-05,
631
+ "loss": 1.3233,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 0.41,
636
+ "learning_rate": 1.1718851049734215e-05,
637
+ "loss": 1.2995,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 0.42,
642
+ "learning_rate": 1.1639982964493587e-05,
643
+ "loss": 1.3273,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 0.42,
648
+ "learning_rate": 1.1561114879252961e-05,
649
+ "loss": 1.3384,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 0.43,
654
+ "learning_rate": 1.1482246794012335e-05,
655
+ "loss": 1.3127,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 0.43,
660
+ "learning_rate": 1.1403378708771708e-05,
661
+ "loss": 1.3681,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 0.43,
666
+ "learning_rate": 1.1324510623531082e-05,
667
+ "loss": 1.3499,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 0.44,
672
+ "learning_rate": 1.1245642538290456e-05,
673
+ "loss": 1.3132,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 0.44,
678
+ "learning_rate": 1.116677445304983e-05,
679
+ "loss": 1.3173,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 0.45,
684
+ "learning_rate": 1.1087906367809203e-05,
685
+ "loss": 1.3366,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 0.45,
690
+ "learning_rate": 1.1009038282568577e-05,
691
+ "loss": 1.3433,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 0.45,
696
+ "learning_rate": 1.093017019732795e-05,
697
+ "loss": 1.3168,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 0.46,
702
+ "learning_rate": 1.0851302112087324e-05,
703
+ "loss": 1.3489,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 0.46,
708
+ "learning_rate": 1.0772434026846696e-05,
709
+ "loss": 1.3441,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 0.47,
714
+ "learning_rate": 1.069356594160607e-05,
715
+ "loss": 1.2982,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 0.47,
720
+ "learning_rate": 1.0614697856365443e-05,
721
+ "loss": 1.3138,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 0.47,
726
+ "learning_rate": 1.0535829771124817e-05,
727
+ "loss": 1.3417,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 0.48,
732
+ "learning_rate": 1.045696168588419e-05,
733
+ "loss": 1.3107,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 0.48,
738
+ "learning_rate": 1.0378093600643564e-05,
739
+ "loss": 1.3524,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 0.49,
744
+ "learning_rate": 1.0299225515402938e-05,
745
+ "loss": 1.3509,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 0.49,
750
+ "learning_rate": 1.0220357430162311e-05,
751
+ "loss": 1.3469,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 0.49,
756
+ "learning_rate": 1.0141489344921685e-05,
757
+ "loss": 1.3231,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 0.5,
762
+ "learning_rate": 1.0062621259681059e-05,
763
+ "loss": 1.3218,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 0.5,
768
+ "learning_rate": 9.98375317444043e-06,
769
+ "loss": 1.3203,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 0.5,
774
+ "learning_rate": 9.904885089199804e-06,
775
+ "loss": 1.286,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 0.51,
780
+ "learning_rate": 9.826017003959178e-06,
781
+ "loss": 1.2927,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 0.51,
786
+ "learning_rate": 9.747148918718552e-06,
787
+ "loss": 1.3631,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 0.52,
792
+ "learning_rate": 9.668280833477925e-06,
793
+ "loss": 1.3556,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 0.52,
798
+ "learning_rate": 9.589412748237299e-06,
799
+ "loss": 1.3153,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 0.52,
804
+ "learning_rate": 9.510544662996673e-06,
805
+ "loss": 1.3567,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 0.53,
810
+ "learning_rate": 9.431676577756046e-06,
811
+ "loss": 1.2935,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 0.53,
816
+ "learning_rate": 9.35280849251542e-06,
817
+ "loss": 1.3497,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 0.54,
822
+ "learning_rate": 9.273940407274794e-06,
823
+ "loss": 1.3418,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 0.54,
828
+ "learning_rate": 9.195072322034167e-06,
829
+ "loss": 1.2812,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 0.54,
834
+ "learning_rate": 9.11620423679354e-06,
835
+ "loss": 1.3112,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 0.55,
840
+ "learning_rate": 9.037336151552913e-06,
841
+ "loss": 1.3378,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 0.55,
846
+ "learning_rate": 8.958468066312287e-06,
847
+ "loss": 1.3173,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 0.56,
852
+ "learning_rate": 8.87959998107166e-06,
853
+ "loss": 1.3013,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 0.56,
858
+ "learning_rate": 8.800731895831034e-06,
859
+ "loss": 1.3323,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 0.56,
864
+ "learning_rate": 8.721863810590407e-06,
865
+ "loss": 1.2787,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 0.57,
870
+ "learning_rate": 8.642995725349781e-06,
871
+ "loss": 1.3173,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 0.57,
876
+ "learning_rate": 8.564127640109155e-06,
877
+ "loss": 1.3196,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 0.58,
882
+ "learning_rate": 8.485259554868528e-06,
883
+ "loss": 1.2903,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 0.58,
888
+ "learning_rate": 8.406391469627902e-06,
889
+ "loss": 1.3292,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 0.58,
894
+ "learning_rate": 8.327523384387276e-06,
895
+ "loss": 1.3402,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 0.59,
900
+ "learning_rate": 8.248655299146648e-06,
901
+ "loss": 1.3504,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 0.59,
906
+ "learning_rate": 8.169787213906021e-06,
907
+ "loss": 1.318,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 0.6,
912
+ "learning_rate": 8.090919128665395e-06,
913
+ "loss": 1.3544,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 0.6,
918
+ "learning_rate": 8.012051043424769e-06,
919
+ "loss": 1.2946,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 0.6,
924
+ "learning_rate": 7.93318295818414e-06,
925
+ "loss": 1.357,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 0.61,
930
+ "learning_rate": 7.854314872943514e-06,
931
+ "loss": 1.3001,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 0.61,
936
+ "learning_rate": 7.775446787702888e-06,
937
+ "loss": 1.3008,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 0.62,
942
+ "learning_rate": 7.696578702462262e-06,
943
+ "loss": 1.3285,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 0.62,
948
+ "learning_rate": 7.617710617221635e-06,
949
+ "loss": 1.3149,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 0.62,
954
+ "learning_rate": 7.538842531981009e-06,
955
+ "loss": 1.2988,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 0.63,
960
+ "learning_rate": 7.459974446740383e-06,
961
+ "loss": 1.3254,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 0.63,
966
+ "learning_rate": 7.381106361499755e-06,
967
+ "loss": 1.3198,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 0.63,
972
+ "learning_rate": 7.302238276259129e-06,
973
+ "loss": 1.2625,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 0.64,
978
+ "learning_rate": 7.223370191018503e-06,
979
+ "loss": 1.3755,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 0.64,
984
+ "learning_rate": 7.144502105777876e-06,
985
+ "loss": 1.3148,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 0.65,
990
+ "learning_rate": 7.06563402053725e-06,
991
+ "loss": 1.2799,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 0.65,
996
+ "learning_rate": 6.986765935296624e-06,
997
+ "loss": 1.3211,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 0.65,
1002
+ "learning_rate": 6.9078978500559965e-06,
1003
+ "loss": 1.3325,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 0.66,
1008
+ "learning_rate": 6.82902976481537e-06,
1009
+ "loss": 1.2948,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 0.66,
1014
+ "learning_rate": 6.750161679574744e-06,
1015
+ "loss": 1.314,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 0.67,
1020
+ "learning_rate": 6.6712935943341175e-06,
1021
+ "loss": 1.2739,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 0.67,
1026
+ "learning_rate": 6.592425509093491e-06,
1027
+ "loss": 1.3231,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 0.67,
1032
+ "learning_rate": 6.513557423852864e-06,
1033
+ "loss": 1.3044,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 0.68,
1038
+ "learning_rate": 6.434689338612238e-06,
1039
+ "loss": 1.3177,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 0.68,
1044
+ "learning_rate": 6.355821253371611e-06,
1045
+ "loss": 1.3457,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 0.69,
1050
+ "learning_rate": 6.276953168130985e-06,
1051
+ "loss": 1.3338,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 0.69,
1056
+ "learning_rate": 6.1980850828903586e-06,
1057
+ "loss": 1.3145,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 0.69,
1062
+ "learning_rate": 6.119216997649731e-06,
1063
+ "loss": 1.2863,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 0.7,
1068
+ "learning_rate": 6.040348912409105e-06,
1069
+ "loss": 1.3334,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 0.7,
1074
+ "learning_rate": 5.961480827168479e-06,
1075
+ "loss": 1.2809,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 0.71,
1080
+ "learning_rate": 5.882612741927852e-06,
1081
+ "loss": 1.3164,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 0.71,
1086
+ "learning_rate": 5.803744656687226e-06,
1087
+ "loss": 1.2591,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 0.71,
1092
+ "learning_rate": 5.7248765714466e-06,
1093
+ "loss": 1.3127,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 0.72,
1098
+ "learning_rate": 5.6460084862059725e-06,
1099
+ "loss": 1.2838,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 0.72,
1104
+ "learning_rate": 5.567140400965346e-06,
1105
+ "loss": 1.2783,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 0.73,
1110
+ "learning_rate": 5.48827231572472e-06,
1111
+ "loss": 1.2546,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 0.73,
1116
+ "learning_rate": 5.4094042304840934e-06,
1117
+ "loss": 1.3429,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 0.73,
1122
+ "learning_rate": 5.330536145243467e-06,
1123
+ "loss": 1.3223,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 0.74,
1128
+ "learning_rate": 5.251668060002839e-06,
1129
+ "loss": 1.2925,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 0.74,
1134
+ "learning_rate": 5.172799974762213e-06,
1135
+ "loss": 1.2802,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 0.75,
1140
+ "learning_rate": 5.093931889521586e-06,
1141
+ "loss": 1.2491,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 0.75,
1146
+ "learning_rate": 5.01506380428096e-06,
1147
+ "loss": 1.3079,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 0.75,
1152
+ "learning_rate": 4.936195719040334e-06,
1153
+ "loss": 1.2788,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 0.76,
1158
+ "learning_rate": 4.857327633799707e-06,
1159
+ "loss": 1.2728,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 0.76,
1164
+ "learning_rate": 4.778459548559081e-06,
1165
+ "loss": 1.3208,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 0.77,
1170
+ "learning_rate": 4.699591463318454e-06,
1171
+ "loss": 1.3279,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 0.77,
1176
+ "learning_rate": 4.6207233780778275e-06,
1177
+ "loss": 1.3523,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 0.77,
1182
+ "learning_rate": 4.5418552928372e-06,
1183
+ "loss": 1.3168,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 0.78,
1188
+ "learning_rate": 4.462987207596574e-06,
1189
+ "loss": 1.2678,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 0.78,
1194
+ "learning_rate": 4.384119122355948e-06,
1195
+ "loss": 1.2477,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 0.78,
1200
+ "learning_rate": 4.305251037115321e-06,
1201
+ "loss": 1.3298,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 0.79,
1206
+ "learning_rate": 4.226382951874695e-06,
1207
+ "loss": 1.2886,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 0.79,
1212
+ "learning_rate": 4.147514866634068e-06,
1213
+ "loss": 1.3194,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 0.8,
1218
+ "learning_rate": 4.068646781393441e-06,
1219
+ "loss": 1.3336,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 0.8,
1224
+ "learning_rate": 3.989778696152815e-06,
1225
+ "loss": 1.28,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 0.8,
1230
+ "learning_rate": 3.910910610912189e-06,
1231
+ "loss": 1.2774,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 0.81,
1236
+ "learning_rate": 3.832042525671562e-06,
1237
+ "loss": 1.3041,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 0.81,
1242
+ "learning_rate": 3.7531744404309356e-06,
1243
+ "loss": 1.3761,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 0.82,
1248
+ "learning_rate": 3.6743063551903092e-06,
1249
+ "loss": 1.37,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 0.82,
1254
+ "learning_rate": 3.5954382699496825e-06,
1255
+ "loss": 1.3055,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 0.82,
1260
+ "learning_rate": 3.516570184709056e-06,
1261
+ "loss": 1.3361,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 0.83,
1266
+ "learning_rate": 3.4377020994684298e-06,
1267
+ "loss": 1.3031,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 0.83,
1272
+ "learning_rate": 3.358834014227803e-06,
1273
+ "loss": 1.3137,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 0.84,
1278
+ "learning_rate": 3.2799659289871767e-06,
1279
+ "loss": 1.2919,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 0.84,
1284
+ "learning_rate": 3.2010978437465495e-06,
1285
+ "loss": 1.2852,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 0.84,
1290
+ "learning_rate": 3.122229758505923e-06,
1291
+ "loss": 1.2751,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 0.85,
1296
+ "learning_rate": 3.0433616732652964e-06,
1297
+ "loss": 1.2906,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 0.85,
1302
+ "learning_rate": 2.96449358802467e-06,
1303
+ "loss": 1.3119,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 0.86,
1308
+ "learning_rate": 2.8856255027840437e-06,
1309
+ "loss": 1.2621,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 0.86,
1314
+ "learning_rate": 2.806757417543417e-06,
1315
+ "loss": 1.2994,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 0.86,
1320
+ "learning_rate": 2.7278893323027906e-06,
1321
+ "loss": 1.2909,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 0.87,
1326
+ "learning_rate": 2.649021247062164e-06,
1327
+ "loss": 1.3306,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 0.87,
1332
+ "learning_rate": 2.5701531618215375e-06,
1333
+ "loss": 1.2881,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 0.88,
1338
+ "learning_rate": 2.491285076580911e-06,
1339
+ "loss": 1.306,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 0.88,
1344
+ "learning_rate": 2.4124169913402844e-06,
1345
+ "loss": 1.3357,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 0.88,
1350
+ "learning_rate": 2.333548906099658e-06,
1351
+ "loss": 1.3152,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 0.89,
1356
+ "learning_rate": 2.2546808208590317e-06,
1357
+ "loss": 1.3011,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 0.89,
1362
+ "learning_rate": 2.175812735618405e-06,
1363
+ "loss": 1.2949,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 0.9,
1368
+ "learning_rate": 2.096944650377778e-06,
1369
+ "loss": 1.2997,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 0.9,
1374
+ "learning_rate": 2.018076565137152e-06,
1375
+ "loss": 1.3148,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 0.9,
1380
+ "learning_rate": 1.939208479896525e-06,
1381
+ "loss": 1.3464,
1382
+ "step": 114500
1383
+ },
1384
+ {
1385
+ "epoch": 0.91,
1386
+ "learning_rate": 1.8603403946558987e-06,
1387
+ "loss": 1.2722,
1388
+ "step": 115000
1389
+ },
1390
+ {
1391
+ "epoch": 0.91,
1392
+ "learning_rate": 1.7814723094152721e-06,
1393
+ "loss": 1.2361,
1394
+ "step": 115500
1395
+ },
1396
+ {
1397
+ "epoch": 0.91,
1398
+ "learning_rate": 1.7026042241746458e-06,
1399
+ "loss": 1.2932,
1400
+ "step": 116000
1401
+ },
1402
+ {
1403
+ "epoch": 0.92,
1404
+ "learning_rate": 1.6237361389340192e-06,
1405
+ "loss": 1.2929,
1406
+ "step": 116500
1407
+ },
1408
+ {
1409
+ "epoch": 0.92,
1410
+ "learning_rate": 1.5448680536933925e-06,
1411
+ "loss": 1.2514,
1412
+ "step": 117000
1413
+ },
1414
+ {
1415
+ "epoch": 0.93,
1416
+ "learning_rate": 1.465999968452766e-06,
1417
+ "loss": 1.3133,
1418
+ "step": 117500
1419
+ },
1420
+ {
1421
+ "epoch": 0.93,
1422
+ "learning_rate": 1.3871318832121394e-06,
1423
+ "loss": 1.2766,
1424
+ "step": 118000
1425
+ },
1426
+ {
1427
+ "epoch": 0.93,
1428
+ "learning_rate": 1.308263797971513e-06,
1429
+ "loss": 1.2892,
1430
+ "step": 118500
1431
+ },
1432
+ {
1433
+ "epoch": 0.94,
1434
+ "learning_rate": 1.2293957127308865e-06,
1435
+ "loss": 1.3185,
1436
+ "step": 119000
1437
+ },
1438
+ {
1439
+ "epoch": 0.94,
1440
+ "learning_rate": 1.15052762749026e-06,
1441
+ "loss": 1.299,
1442
+ "step": 119500
1443
+ },
1444
+ {
1445
+ "epoch": 0.95,
1446
+ "learning_rate": 1.0716595422496333e-06,
1447
+ "loss": 1.2878,
1448
+ "step": 120000
1449
+ },
1450
+ {
1451
+ "epoch": 0.95,
1452
+ "learning_rate": 9.927914570090068e-07,
1453
+ "loss": 1.2827,
1454
+ "step": 120500
1455
+ },
1456
+ {
1457
+ "epoch": 0.95,
1458
+ "learning_rate": 9.139233717683802e-07,
1459
+ "loss": 1.3103,
1460
+ "step": 121000
1461
+ },
1462
+ {
1463
+ "epoch": 0.96,
1464
+ "learning_rate": 8.350552865277538e-07,
1465
+ "loss": 1.3186,
1466
+ "step": 121500
1467
+ },
1468
+ {
1469
+ "epoch": 0.96,
1470
+ "learning_rate": 7.561872012871272e-07,
1471
+ "loss": 1.3046,
1472
+ "step": 122000
1473
+ },
1474
+ {
1475
+ "epoch": 0.97,
1476
+ "learning_rate": 6.773191160465007e-07,
1477
+ "loss": 1.2945,
1478
+ "step": 122500
1479
+ },
1480
+ {
1481
+ "epoch": 0.97,
1482
+ "learning_rate": 5.984510308058741e-07,
1483
+ "loss": 1.3193,
1484
+ "step": 123000
1485
+ },
1486
+ {
1487
+ "epoch": 0.97,
1488
+ "learning_rate": 5.195829455652477e-07,
1489
+ "loss": 1.3048,
1490
+ "step": 123500
1491
+ },
1492
+ {
1493
+ "epoch": 0.98,
1494
+ "learning_rate": 4.4071486032462107e-07,
1495
+ "loss": 1.2599,
1496
+ "step": 124000
1497
+ },
1498
+ {
1499
+ "epoch": 0.98,
1500
+ "learning_rate": 3.6184677508399457e-07,
1501
+ "loss": 1.2585,
1502
+ "step": 124500
1503
+ },
1504
+ {
1505
+ "epoch": 0.99,
1506
+ "learning_rate": 2.82978689843368e-07,
1507
+ "loss": 1.2506,
1508
+ "step": 125000
1509
+ },
1510
+ {
1511
+ "epoch": 0.99,
1512
+ "learning_rate": 2.0411060460274149e-07,
1513
+ "loss": 1.2928,
1514
+ "step": 125500
1515
+ },
1516
+ {
1517
+ "epoch": 0.99,
1518
+ "learning_rate": 1.2524251936211493e-07,
1519
+ "loss": 1.2959,
1520
+ "step": 126000
1521
+ },
1522
+ {
1523
+ "epoch": 1.0,
1524
+ "learning_rate": 4.63744341214884e-08,
1525
+ "loss": 1.2957,
1526
+ "step": 126500
1527
+ }
1528
+ ],
1529
+ "max_steps": 126794,
1530
+ "num_train_epochs": 1,
1531
+ "total_flos": 4757862742622208.0,
1532
+ "trial_name": null,
1533
+ "trial_params": null
1534
+ }