kotstantinovskii commited on
Commit
b474b3f
1 Parent(s): 2098228

Upload checkpoint-15500/trainer_state.json

Browse files
Files changed (1) hide show
  1. checkpoint-15500/trainer_state.json +1008 -0
checkpoint-15500/trainer_state.json ADDED
@@ -0,0 +1,1008 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.0360217094421387,
3
+ "best_model_checkpoint": "./res_2/checkpoint-12000",
4
+ "epoch": 3.7685387794796985,
5
+ "global_step": 15500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.06,
12
+ "learning_rate": 2.5e-05,
13
+ "loss": 4.4105,
14
+ "step": 250
15
+ },
16
+ {
17
+ "epoch": 0.06,
18
+ "eval_accuracy": 0.2648573215918387,
19
+ "eval_f1_score": 0.1461888457719469,
20
+ "eval_loss": 3.3562278747558594,
21
+ "eval_runtime": 104.4517,
22
+ "eval_samples_per_second": 198.953,
23
+ "eval_steps_per_second": 2.078,
24
+ "step": 250
25
+ },
26
+ {
27
+ "epoch": 0.12,
28
+ "learning_rate": 5e-05,
29
+ "loss": 2.7264,
30
+ "step": 500
31
+ },
32
+ {
33
+ "epoch": 0.12,
34
+ "eval_accuracy": 0.4484865983350176,
35
+ "eval_f1_score": 0.36937033203465997,
36
+ "eval_loss": 2.1725332736968994,
37
+ "eval_runtime": 104.6188,
38
+ "eval_samples_per_second": 198.635,
39
+ "eval_steps_per_second": 2.074,
40
+ "step": 500
41
+ },
42
+ {
43
+ "epoch": 0.18,
44
+ "learning_rate": 4.9216399197592786e-05,
45
+ "loss": 2.0303,
46
+ "step": 750
47
+ },
48
+ {
49
+ "epoch": 0.18,
50
+ "eval_accuracy": 0.5143640825754294,
51
+ "eval_f1_score": 0.4607925258467065,
52
+ "eval_loss": 1.784591555595398,
53
+ "eval_runtime": 104.5717,
54
+ "eval_samples_per_second": 198.725,
55
+ "eval_steps_per_second": 2.075,
56
+ "step": 750
57
+ },
58
+ {
59
+ "epoch": 0.24,
60
+ "learning_rate": 4.8432798395185555e-05,
61
+ "loss": 1.777,
62
+ "step": 1000
63
+ },
64
+ {
65
+ "epoch": 0.24,
66
+ "eval_accuracy": 0.5497329291179442,
67
+ "eval_f1_score": 0.5086586604390044,
68
+ "eval_loss": 1.6212031841278076,
69
+ "eval_runtime": 104.6796,
70
+ "eval_samples_per_second": 198.52,
71
+ "eval_steps_per_second": 2.073,
72
+ "step": 1000
73
+ },
74
+ {
75
+ "epoch": 0.3,
76
+ "learning_rate": 4.764919759277834e-05,
77
+ "loss": 1.635,
78
+ "step": 1250
79
+ },
80
+ {
81
+ "epoch": 0.3,
82
+ "eval_accuracy": 0.57292719310909,
83
+ "eval_f1_score": 0.5363731366139944,
84
+ "eval_loss": 1.4894088506698608,
85
+ "eval_runtime": 104.782,
86
+ "eval_samples_per_second": 198.326,
87
+ "eval_steps_per_second": 2.071,
88
+ "step": 1250
89
+ },
90
+ {
91
+ "epoch": 0.36,
92
+ "learning_rate": 4.6865596790371115e-05,
93
+ "loss": 1.5492,
94
+ "step": 1500
95
+ },
96
+ {
97
+ "epoch": 0.36,
98
+ "eval_accuracy": 0.5871228526057456,
99
+ "eval_f1_score": 0.5565777587100311,
100
+ "eval_loss": 1.4159808158874512,
101
+ "eval_runtime": 104.9221,
102
+ "eval_samples_per_second": 198.061,
103
+ "eval_steps_per_second": 2.068,
104
+ "step": 1500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 4.608199598796389e-05,
109
+ "loss": 1.4853,
110
+ "step": 1750
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_accuracy": 0.6005004571483566,
115
+ "eval_f1_score": 0.5743468856693701,
116
+ "eval_loss": 1.3664453029632568,
117
+ "eval_runtime": 105.1103,
118
+ "eval_samples_per_second": 197.707,
119
+ "eval_steps_per_second": 2.064,
120
+ "step": 1750
121
+ },
122
+ {
123
+ "epoch": 0.49,
124
+ "learning_rate": 4.5298395185556675e-05,
125
+ "loss": 1.4178,
126
+ "step": 2000
127
+ },
128
+ {
129
+ "epoch": 0.49,
130
+ "eval_accuracy": 0.6162841056734517,
131
+ "eval_f1_score": 0.5926712498507521,
132
+ "eval_loss": 1.3030893802642822,
133
+ "eval_runtime": 104.7919,
134
+ "eval_samples_per_second": 198.307,
135
+ "eval_steps_per_second": 2.071,
136
+ "step": 2000
137
+ },
138
+ {
139
+ "epoch": 0.55,
140
+ "learning_rate": 4.451479438314945e-05,
141
+ "loss": 1.4012,
142
+ "step": 2250
143
+ },
144
+ {
145
+ "epoch": 0.55,
146
+ "eval_accuracy": 0.609932149559694,
147
+ "eval_f1_score": 0.5892368884639717,
148
+ "eval_loss": 1.3101539611816406,
149
+ "eval_runtime": 105.0064,
150
+ "eval_samples_per_second": 197.902,
151
+ "eval_steps_per_second": 2.067,
152
+ "step": 2250
153
+ },
154
+ {
155
+ "epoch": 0.61,
156
+ "learning_rate": 4.373119358074223e-05,
157
+ "loss": 1.363,
158
+ "step": 2500
159
+ },
160
+ {
161
+ "epoch": 0.61,
162
+ "eval_accuracy": 0.6175833694239931,
163
+ "eval_f1_score": 0.596718851218216,
164
+ "eval_loss": 1.2853854894638062,
165
+ "eval_runtime": 104.962,
166
+ "eval_samples_per_second": 197.986,
167
+ "eval_steps_per_second": 2.067,
168
+ "step": 2500
169
+ },
170
+ {
171
+ "epoch": 0.67,
172
+ "learning_rate": 4.2947592778335004e-05,
173
+ "loss": 1.3349,
174
+ "step": 2750
175
+ },
176
+ {
177
+ "epoch": 0.67,
178
+ "eval_accuracy": 0.6279293585486743,
179
+ "eval_f1_score": 0.6094187627596529,
180
+ "eval_loss": 1.230573058128357,
181
+ "eval_runtime": 104.847,
182
+ "eval_samples_per_second": 198.203,
183
+ "eval_steps_per_second": 2.07,
184
+ "step": 2750
185
+ },
186
+ {
187
+ "epoch": 0.73,
188
+ "learning_rate": 4.216399197592779e-05,
189
+ "loss": 1.3324,
190
+ "step": 3000
191
+ },
192
+ {
193
+ "epoch": 0.73,
194
+ "eval_accuracy": 0.6324046003560945,
195
+ "eval_f1_score": 0.6161864856783154,
196
+ "eval_loss": 1.2122113704681396,
197
+ "eval_runtime": 105.0625,
198
+ "eval_samples_per_second": 197.796,
199
+ "eval_steps_per_second": 2.065,
200
+ "step": 3000
201
+ },
202
+ {
203
+ "epoch": 0.79,
204
+ "learning_rate": 4.1380391173520564e-05,
205
+ "loss": 1.292,
206
+ "step": 3250
207
+ },
208
+ {
209
+ "epoch": 0.79,
210
+ "eval_accuracy": 0.6309128530869544,
211
+ "eval_f1_score": 0.6119500434777825,
212
+ "eval_loss": 1.20195472240448,
213
+ "eval_runtime": 105.1108,
214
+ "eval_samples_per_second": 197.706,
215
+ "eval_steps_per_second": 2.064,
216
+ "step": 3250
217
+ },
218
+ {
219
+ "epoch": 0.85,
220
+ "learning_rate": 4.059679037111334e-05,
221
+ "loss": 1.2693,
222
+ "step": 3500
223
+ },
224
+ {
225
+ "epoch": 0.85,
226
+ "eval_accuracy": 0.6373129300803618,
227
+ "eval_f1_score": 0.624172345328055,
228
+ "eval_loss": 1.1805213689804077,
229
+ "eval_runtime": 104.8144,
230
+ "eval_samples_per_second": 198.265,
231
+ "eval_steps_per_second": 2.07,
232
+ "step": 3500
233
+ },
234
+ {
235
+ "epoch": 0.91,
236
+ "learning_rate": 3.9813189568706124e-05,
237
+ "loss": 1.2729,
238
+ "step": 3750
239
+ },
240
+ {
241
+ "epoch": 0.91,
242
+ "eval_accuracy": 0.6417881718877821,
243
+ "eval_f1_score": 0.6227824246172433,
244
+ "eval_loss": 1.1760473251342773,
245
+ "eval_runtime": 104.7877,
246
+ "eval_samples_per_second": 198.315,
247
+ "eval_steps_per_second": 2.071,
248
+ "step": 3750
249
+ },
250
+ {
251
+ "epoch": 0.97,
252
+ "learning_rate": 3.90295887662989e-05,
253
+ "loss": 1.2382,
254
+ "step": 4000
255
+ },
256
+ {
257
+ "epoch": 0.97,
258
+ "eval_accuracy": 0.6395746114238968,
259
+ "eval_f1_score": 0.6232000558005116,
260
+ "eval_loss": 1.1657843589782715,
261
+ "eval_runtime": 104.9017,
262
+ "eval_samples_per_second": 198.1,
263
+ "eval_steps_per_second": 2.069,
264
+ "step": 4000
265
+ },
266
+ {
267
+ "epoch": 1.03,
268
+ "learning_rate": 3.8245987963891677e-05,
269
+ "loss": 1.175,
270
+ "step": 4250
271
+ },
272
+ {
273
+ "epoch": 1.03,
274
+ "eval_accuracy": 0.6491025455945335,
275
+ "eval_f1_score": 0.6333235301999404,
276
+ "eval_loss": 1.1377447843551636,
277
+ "eval_runtime": 104.8132,
278
+ "eval_samples_per_second": 198.267,
279
+ "eval_steps_per_second": 2.07,
280
+ "step": 4250
281
+ },
282
+ {
283
+ "epoch": 1.09,
284
+ "learning_rate": 3.746238716148446e-05,
285
+ "loss": 1.1071,
286
+ "step": 4500
287
+ },
288
+ {
289
+ "epoch": 1.09,
290
+ "eval_accuracy": 0.6498243587892787,
291
+ "eval_f1_score": 0.6338669978734944,
292
+ "eval_loss": 1.134881854057312,
293
+ "eval_runtime": 104.9571,
294
+ "eval_samples_per_second": 197.995,
295
+ "eval_steps_per_second": 2.068,
296
+ "step": 4500
297
+ },
298
+ {
299
+ "epoch": 1.15,
300
+ "learning_rate": 3.6678786359077236e-05,
301
+ "loss": 1.1257,
302
+ "step": 4750
303
+ },
304
+ {
305
+ "epoch": 1.15,
306
+ "eval_accuracy": 0.6517491939752659,
307
+ "eval_f1_score": 0.6390275803026418,
308
+ "eval_loss": 1.1246434450149536,
309
+ "eval_runtime": 118.4939,
310
+ "eval_samples_per_second": 175.376,
311
+ "eval_steps_per_second": 1.831,
312
+ "step": 4750
313
+ },
314
+ {
315
+ "epoch": 1.22,
316
+ "learning_rate": 3.589518555667001e-05,
317
+ "loss": 1.0846,
318
+ "step": 5000
319
+ },
320
+ {
321
+ "epoch": 1.22,
322
+ "eval_accuracy": 0.6530484577258072,
323
+ "eval_f1_score": 0.6374295817702883,
324
+ "eval_loss": 1.134043574333191,
325
+ "eval_runtime": 104.8506,
326
+ "eval_samples_per_second": 198.196,
327
+ "eval_steps_per_second": 2.07,
328
+ "step": 5000
329
+ },
330
+ {
331
+ "epoch": 1.28,
332
+ "learning_rate": 3.511158475426279e-05,
333
+ "loss": 1.1093,
334
+ "step": 5250
335
+ },
336
+ {
337
+ "epoch": 1.28,
338
+ "eval_accuracy": 0.6527116115682595,
339
+ "eval_f1_score": 0.6433227186941691,
340
+ "eval_loss": 1.1218096017837524,
341
+ "eval_runtime": 105.0759,
342
+ "eval_samples_per_second": 197.771,
343
+ "eval_steps_per_second": 2.065,
344
+ "step": 5250
345
+ },
346
+ {
347
+ "epoch": 1.34,
348
+ "learning_rate": 3.4327983951855566e-05,
349
+ "loss": 1.0895,
350
+ "step": 5500
351
+ },
352
+ {
353
+ "epoch": 1.34,
354
+ "eval_accuracy": 0.6540108753188009,
355
+ "eval_f1_score": 0.640048503150472,
356
+ "eval_loss": 1.120483160018921,
357
+ "eval_runtime": 105.0798,
358
+ "eval_samples_per_second": 197.764,
359
+ "eval_steps_per_second": 2.065,
360
+ "step": 5500
361
+ },
362
+ {
363
+ "epoch": 1.4,
364
+ "learning_rate": 3.354438314944834e-05,
365
+ "loss": 1.0947,
366
+ "step": 5750
367
+ },
368
+ {
369
+ "epoch": 1.4,
370
+ "eval_accuracy": 0.6575236995332274,
371
+ "eval_f1_score": 0.6418486412023349,
372
+ "eval_loss": 1.1041547060012817,
373
+ "eval_runtime": 105.0665,
374
+ "eval_samples_per_second": 197.789,
375
+ "eval_steps_per_second": 2.065,
376
+ "step": 5750
377
+ },
378
+ {
379
+ "epoch": 1.46,
380
+ "learning_rate": 3.2760782347041125e-05,
381
+ "loss": 1.0784,
382
+ "step": 6000
383
+ },
384
+ {
385
+ "epoch": 1.46,
386
+ "eval_accuracy": 0.6592079303209663,
387
+ "eval_f1_score": 0.645978450692562,
388
+ "eval_loss": 1.0955352783203125,
389
+ "eval_runtime": 104.9867,
390
+ "eval_samples_per_second": 197.939,
391
+ "eval_steps_per_second": 2.067,
392
+ "step": 6000
393
+ },
394
+ {
395
+ "epoch": 1.52,
396
+ "learning_rate": 3.19771815446339e-05,
397
+ "loss": 1.0838,
398
+ "step": 6250
399
+ },
400
+ {
401
+ "epoch": 1.52,
402
+ "eval_accuracy": 0.6625763918964439,
403
+ "eval_f1_score": 0.6483699052640116,
404
+ "eval_loss": 1.095639944076538,
405
+ "eval_runtime": 104.9633,
406
+ "eval_samples_per_second": 197.983,
407
+ "eval_steps_per_second": 2.067,
408
+ "step": 6250
409
+ },
410
+ {
411
+ "epoch": 1.58,
412
+ "learning_rate": 3.119358074222668e-05,
413
+ "loss": 1.0764,
414
+ "step": 6500
415
+ },
416
+ {
417
+ "epoch": 1.58,
418
+ "eval_accuracy": 0.6629132380539916,
419
+ "eval_f1_score": 0.6497593784369542,
420
+ "eval_loss": 1.0892637968063354,
421
+ "eval_runtime": 105.0455,
422
+ "eval_samples_per_second": 197.829,
423
+ "eval_steps_per_second": 2.066,
424
+ "step": 6500
425
+ },
426
+ {
427
+ "epoch": 1.64,
428
+ "learning_rate": 3.040997993981946e-05,
429
+ "loss": 1.0563,
430
+ "step": 6750
431
+ },
432
+ {
433
+ "epoch": 1.64,
434
+ "eval_accuracy": 0.6619508204609981,
435
+ "eval_f1_score": 0.6484351513187849,
436
+ "eval_loss": 1.089970350265503,
437
+ "eval_runtime": 105.3681,
438
+ "eval_samples_per_second": 197.223,
439
+ "eval_steps_per_second": 2.059,
440
+ "step": 6750
441
+ },
442
+ {
443
+ "epoch": 1.7,
444
+ "learning_rate": 2.9626379137412235e-05,
445
+ "loss": 1.066,
446
+ "step": 7000
447
+ },
448
+ {
449
+ "epoch": 1.7,
450
+ "eval_accuracy": 0.6659929743515711,
451
+ "eval_f1_score": 0.6533335696048064,
452
+ "eval_loss": 1.0745749473571777,
453
+ "eval_runtime": 105.1237,
454
+ "eval_samples_per_second": 197.681,
455
+ "eval_steps_per_second": 2.064,
456
+ "step": 7000
457
+ },
458
+ {
459
+ "epoch": 1.76,
460
+ "learning_rate": 2.8842778335005018e-05,
461
+ "loss": 1.0603,
462
+ "step": 7250
463
+ },
464
+ {
465
+ "epoch": 1.76,
466
+ "eval_accuracy": 0.6680140512968578,
467
+ "eval_f1_score": 0.657494747880274,
468
+ "eval_loss": 1.0701572895050049,
469
+ "eval_runtime": 105.023,
470
+ "eval_samples_per_second": 197.871,
471
+ "eval_steps_per_second": 2.066,
472
+ "step": 7250
473
+ },
474
+ {
475
+ "epoch": 1.82,
476
+ "learning_rate": 2.8059177532597798e-05,
477
+ "loss": 1.0624,
478
+ "step": 7500
479
+ },
480
+ {
481
+ "epoch": 1.82,
482
+ "eval_accuracy": 0.6685915018526538,
483
+ "eval_f1_score": 0.6554959146242652,
484
+ "eval_loss": 1.0651373863220215,
485
+ "eval_runtime": 105.2005,
486
+ "eval_samples_per_second": 197.537,
487
+ "eval_steps_per_second": 2.063,
488
+ "step": 7500
489
+ },
490
+ {
491
+ "epoch": 1.88,
492
+ "learning_rate": 2.727557673019057e-05,
493
+ "loss": 1.0477,
494
+ "step": 7750
495
+ },
496
+ {
497
+ "epoch": 1.88,
498
+ "eval_accuracy": 0.6685915018526538,
499
+ "eval_f1_score": 0.655750687605972,
500
+ "eval_loss": 1.059809684753418,
501
+ "eval_runtime": 105.0102,
502
+ "eval_samples_per_second": 197.895,
503
+ "eval_steps_per_second": 2.066,
504
+ "step": 7750
505
+ },
506
+ {
507
+ "epoch": 1.95,
508
+ "learning_rate": 2.649197592778335e-05,
509
+ "loss": 1.0455,
510
+ "step": 8000
511
+ },
512
+ {
513
+ "epoch": 1.95,
514
+ "eval_accuracy": 0.6643087435638323,
515
+ "eval_f1_score": 0.6529621516807526,
516
+ "eval_loss": 1.0628999471664429,
517
+ "eval_runtime": 104.8979,
518
+ "eval_samples_per_second": 198.107,
519
+ "eval_steps_per_second": 2.069,
520
+ "step": 8000
521
+ },
522
+ {
523
+ "epoch": 2.01,
524
+ "learning_rate": 2.5708375125376127e-05,
525
+ "loss": 1.0403,
526
+ "step": 8250
527
+ },
528
+ {
529
+ "epoch": 2.01,
530
+ "eval_accuracy": 0.6683990183340551,
531
+ "eval_f1_score": 0.655431489724863,
532
+ "eval_loss": 1.0640465021133423,
533
+ "eval_runtime": 104.8293,
534
+ "eval_samples_per_second": 198.237,
535
+ "eval_steps_per_second": 2.07,
536
+ "step": 8250
537
+ },
538
+ {
539
+ "epoch": 2.07,
540
+ "learning_rate": 2.4924774322968907e-05,
541
+ "loss": 0.9083,
542
+ "step": 8500
543
+ },
544
+ {
545
+ "epoch": 2.07,
546
+ "eval_accuracy": 0.6707569414368895,
547
+ "eval_f1_score": 0.6605996988642956,
548
+ "eval_loss": 1.0650361776351929,
549
+ "eval_runtime": 104.9496,
550
+ "eval_samples_per_second": 198.009,
551
+ "eval_steps_per_second": 2.068,
552
+ "step": 8500
553
+ },
554
+ {
555
+ "epoch": 2.13,
556
+ "learning_rate": 2.4141173520561687e-05,
557
+ "loss": 0.9045,
558
+ "step": 8750
559
+ },
560
+ {
561
+ "epoch": 2.13,
562
+ "eval_accuracy": 0.6696501612049468,
563
+ "eval_f1_score": 0.6590246628667181,
564
+ "eval_loss": 1.0609605312347412,
565
+ "eval_runtime": 105.2135,
566
+ "eval_samples_per_second": 197.513,
567
+ "eval_steps_per_second": 2.062,
568
+ "step": 8750
569
+ },
570
+ {
571
+ "epoch": 2.19,
572
+ "learning_rate": 2.3357572718154463e-05,
573
+ "loss": 0.8998,
574
+ "step": 9000
575
+ },
576
+ {
577
+ "epoch": 2.19,
578
+ "eval_accuracy": 0.6709975458351378,
579
+ "eval_f1_score": 0.6609721373546371,
580
+ "eval_loss": 1.0593194961547852,
581
+ "eval_runtime": 104.8728,
582
+ "eval_samples_per_second": 198.154,
583
+ "eval_steps_per_second": 2.069,
584
+ "step": 9000
585
+ },
586
+ {
587
+ "epoch": 2.25,
588
+ "learning_rate": 2.2573971915747243e-05,
589
+ "loss": 0.8994,
590
+ "step": 9250
591
+ },
592
+ {
593
+ "epoch": 2.25,
594
+ "eval_accuracy": 0.6718156007891825,
595
+ "eval_f1_score": 0.6593984410226574,
596
+ "eval_loss": 1.064655065536499,
597
+ "eval_runtime": 104.9299,
598
+ "eval_samples_per_second": 198.046,
599
+ "eval_steps_per_second": 2.068,
600
+ "step": 9250
601
+ },
602
+ {
603
+ "epoch": 2.31,
604
+ "learning_rate": 2.1790371113340023e-05,
605
+ "loss": 0.9093,
606
+ "step": 9500
607
+ },
608
+ {
609
+ "epoch": 2.31,
610
+ "eval_accuracy": 0.668880227130552,
611
+ "eval_f1_score": 0.6589669423489957,
612
+ "eval_loss": 1.0653289556503296,
613
+ "eval_runtime": 105.0839,
614
+ "eval_samples_per_second": 197.756,
615
+ "eval_steps_per_second": 2.065,
616
+ "step": 9500
617
+ },
618
+ {
619
+ "epoch": 2.37,
620
+ "learning_rate": 2.10067703109328e-05,
621
+ "loss": 0.8864,
622
+ "step": 9750
623
+ },
624
+ {
625
+ "epoch": 2.37,
626
+ "eval_accuracy": 0.674895337086762,
627
+ "eval_f1_score": 0.664902640512849,
628
+ "eval_loss": 1.050571322441101,
629
+ "eval_runtime": 105.0155,
630
+ "eval_samples_per_second": 197.885,
631
+ "eval_steps_per_second": 2.066,
632
+ "step": 9750
633
+ },
634
+ {
635
+ "epoch": 2.43,
636
+ "learning_rate": 2.0223169508525576e-05,
637
+ "loss": 0.8905,
638
+ "step": 10000
639
+ },
640
+ {
641
+ "epoch": 2.43,
642
+ "eval_accuracy": 0.6742697656513161,
643
+ "eval_f1_score": 0.6644634718259311,
644
+ "eval_loss": 1.0541752576828003,
645
+ "eval_runtime": 104.7466,
646
+ "eval_samples_per_second": 198.393,
647
+ "eval_steps_per_second": 2.072,
648
+ "step": 10000
649
+ },
650
+ {
651
+ "epoch": 2.49,
652
+ "learning_rate": 1.9439568706118356e-05,
653
+ "loss": 0.8981,
654
+ "step": 10250
655
+ },
656
+ {
657
+ "epoch": 2.49,
658
+ "eval_accuracy": 0.6703238535200423,
659
+ "eval_f1_score": 0.6590618499027966,
660
+ "eval_loss": 1.0549697875976562,
661
+ "eval_runtime": 105.0016,
662
+ "eval_samples_per_second": 197.911,
663
+ "eval_steps_per_second": 2.067,
664
+ "step": 10250
665
+ },
666
+ {
667
+ "epoch": 2.55,
668
+ "learning_rate": 1.8655967903711136e-05,
669
+ "loss": 0.9038,
670
+ "step": 10500
671
+ },
672
+ {
673
+ "epoch": 2.55,
674
+ "eval_accuracy": 0.671286271113036,
675
+ "eval_f1_score": 0.6622994563067591,
676
+ "eval_loss": 1.047540307044983,
677
+ "eval_runtime": 105.0268,
678
+ "eval_samples_per_second": 197.864,
679
+ "eval_steps_per_second": 2.066,
680
+ "step": 10500
681
+ },
682
+ {
683
+ "epoch": 2.61,
684
+ "learning_rate": 1.7872367101303912e-05,
685
+ "loss": 0.8884,
686
+ "step": 10750
687
+ },
688
+ {
689
+ "epoch": 2.61,
690
+ "eval_accuracy": 0.6739329194937683,
691
+ "eval_f1_score": 0.6640063633487168,
692
+ "eval_loss": 1.0482571125030518,
693
+ "eval_runtime": 104.9667,
694
+ "eval_samples_per_second": 197.977,
695
+ "eval_steps_per_second": 2.067,
696
+ "step": 10750
697
+ },
698
+ {
699
+ "epoch": 2.67,
700
+ "learning_rate": 1.7088766298896692e-05,
701
+ "loss": 0.8993,
702
+ "step": 11000
703
+ },
704
+ {
705
+ "epoch": 2.67,
706
+ "eval_accuracy": 0.6732592271786728,
707
+ "eval_f1_score": 0.6644247957468762,
708
+ "eval_loss": 1.0455670356750488,
709
+ "eval_runtime": 104.9075,
710
+ "eval_samples_per_second": 198.089,
711
+ "eval_steps_per_second": 2.068,
712
+ "step": 11000
713
+ },
714
+ {
715
+ "epoch": 2.74,
716
+ "learning_rate": 1.630516549648947e-05,
717
+ "loss": 0.8881,
718
+ "step": 11250
719
+ },
720
+ {
721
+ "epoch": 2.74,
722
+ "eval_accuracy": 0.6758577546797555,
723
+ "eval_f1_score": 0.6663120997306459,
724
+ "eval_loss": 1.0385903120040894,
725
+ "eval_runtime": 104.9411,
726
+ "eval_samples_per_second": 198.025,
727
+ "eval_steps_per_second": 2.068,
728
+ "step": 11250
729
+ },
730
+ {
731
+ "epoch": 2.8,
732
+ "learning_rate": 1.552156469408225e-05,
733
+ "loss": 0.8774,
734
+ "step": 11500
735
+ },
736
+ {
737
+ "epoch": 2.8,
738
+ "eval_accuracy": 0.677590106347144,
739
+ "eval_f1_score": 0.6679638313459818,
740
+ "eval_loss": 1.0412800312042236,
741
+ "eval_runtime": 104.9104,
742
+ "eval_samples_per_second": 198.083,
743
+ "eval_steps_per_second": 2.068,
744
+ "step": 11500
745
+ },
746
+ {
747
+ "epoch": 2.86,
748
+ "learning_rate": 1.4737963891675025e-05,
749
+ "loss": 0.8725,
750
+ "step": 11750
751
+ },
752
+ {
753
+ "epoch": 2.86,
754
+ "eval_accuracy": 0.6772051393099466,
755
+ "eval_f1_score": 0.6684538458032429,
756
+ "eval_loss": 1.041609287261963,
757
+ "eval_runtime": 104.9692,
758
+ "eval_samples_per_second": 197.972,
759
+ "eval_steps_per_second": 2.067,
760
+ "step": 11750
761
+ },
762
+ {
763
+ "epoch": 2.92,
764
+ "learning_rate": 1.3954363089267805e-05,
765
+ "loss": 0.8814,
766
+ "step": 12000
767
+ },
768
+ {
769
+ "epoch": 2.92,
770
+ "eval_accuracy": 0.6780231942639912,
771
+ "eval_f1_score": 0.668403494357387,
772
+ "eval_loss": 1.0360217094421387,
773
+ "eval_runtime": 104.9925,
774
+ "eval_samples_per_second": 197.929,
775
+ "eval_steps_per_second": 2.067,
776
+ "step": 12000
777
+ },
778
+ {
779
+ "epoch": 2.98,
780
+ "learning_rate": 1.3170762286860583e-05,
781
+ "loss": 0.8707,
782
+ "step": 12250
783
+ },
784
+ {
785
+ "epoch": 2.98,
786
+ "eval_accuracy": 0.6776382272267937,
787
+ "eval_f1_score": 0.6678552195861074,
788
+ "eval_loss": 1.036238193511963,
789
+ "eval_runtime": 104.9928,
790
+ "eval_samples_per_second": 197.928,
791
+ "eval_steps_per_second": 2.067,
792
+ "step": 12250
793
+ },
794
+ {
795
+ "epoch": 3.04,
796
+ "learning_rate": 1.2387161484453361e-05,
797
+ "loss": 0.803,
798
+ "step": 12500
799
+ },
800
+ {
801
+ "epoch": 3.04,
802
+ "eval_accuracy": 0.678504403060488,
803
+ "eval_f1_score": 0.6693071368875216,
804
+ "eval_loss": 1.048478364944458,
805
+ "eval_runtime": 104.9691,
806
+ "eval_samples_per_second": 197.973,
807
+ "eval_steps_per_second": 2.067,
808
+ "step": 12500
809
+ },
810
+ {
811
+ "epoch": 3.1,
812
+ "learning_rate": 1.160356068204614e-05,
813
+ "loss": 0.7647,
814
+ "step": 12750
815
+ },
816
+ {
817
+ "epoch": 3.1,
818
+ "eval_accuracy": 0.6773976228285453,
819
+ "eval_f1_score": 0.667836848673267,
820
+ "eval_loss": 1.0551421642303467,
821
+ "eval_runtime": 104.9474,
822
+ "eval_samples_per_second": 198.013,
823
+ "eval_steps_per_second": 2.068,
824
+ "step": 12750
825
+ },
826
+ {
827
+ "epoch": 3.16,
828
+ "learning_rate": 1.0819959879638917e-05,
829
+ "loss": 0.7595,
830
+ "step": 13000
831
+ },
832
+ {
833
+ "epoch": 3.16,
834
+ "eval_accuracy": 0.6766758096338001,
835
+ "eval_f1_score": 0.6681691840022692,
836
+ "eval_loss": 1.0496253967285156,
837
+ "eval_runtime": 104.8372,
838
+ "eval_samples_per_second": 198.222,
839
+ "eval_steps_per_second": 2.07,
840
+ "step": 13000
841
+ },
842
+ {
843
+ "epoch": 3.22,
844
+ "learning_rate": 1.0036359077231696e-05,
845
+ "loss": 0.7628,
846
+ "step": 13250
847
+ },
848
+ {
849
+ "epoch": 3.22,
850
+ "eval_accuracy": 0.6778788316250421,
851
+ "eval_f1_score": 0.6698784049603582,
852
+ "eval_loss": 1.0489078760147095,
853
+ "eval_runtime": 104.8411,
854
+ "eval_samples_per_second": 198.214,
855
+ "eval_steps_per_second": 2.07,
856
+ "step": 13250
857
+ },
858
+ {
859
+ "epoch": 3.28,
860
+ "learning_rate": 9.252758274824474e-06,
861
+ "loss": 0.7711,
862
+ "step": 13500
863
+ },
864
+ {
865
+ "epoch": 3.28,
866
+ "eval_accuracy": 0.6778307107453925,
867
+ "eval_f1_score": 0.670537104393122,
868
+ "eval_loss": 1.0493261814117432,
869
+ "eval_runtime": 104.9371,
870
+ "eval_samples_per_second": 198.033,
871
+ "eval_steps_per_second": 2.068,
872
+ "step": 13500
873
+ },
874
+ {
875
+ "epoch": 3.34,
876
+ "learning_rate": 8.469157472417252e-06,
877
+ "loss": 0.7659,
878
+ "step": 13750
879
+ },
880
+ {
881
+ "epoch": 3.34,
882
+ "eval_accuracy": 0.677445743708195,
883
+ "eval_f1_score": 0.6690005925034824,
884
+ "eval_loss": 1.0476195812225342,
885
+ "eval_runtime": 104.8587,
886
+ "eval_samples_per_second": 198.181,
887
+ "eval_steps_per_second": 2.069,
888
+ "step": 13750
889
+ },
890
+ {
891
+ "epoch": 3.4,
892
+ "learning_rate": 7.68555667001003e-06,
893
+ "loss": 0.7466,
894
+ "step": 14000
895
+ },
896
+ {
897
+ "epoch": 3.4,
898
+ "eval_accuracy": 0.6767720513930995,
899
+ "eval_f1_score": 0.6682514083974397,
900
+ "eval_loss": 1.0503697395324707,
901
+ "eval_runtime": 104.8988,
902
+ "eval_samples_per_second": 198.105,
903
+ "eval_steps_per_second": 2.069,
904
+ "step": 14000
905
+ },
906
+ {
907
+ "epoch": 3.46,
908
+ "learning_rate": 6.901955867602809e-06,
909
+ "loss": 0.7438,
910
+ "step": 14250
911
+ },
912
+ {
913
+ "epoch": 3.46,
914
+ "eval_accuracy": 0.6778788316250421,
915
+ "eval_f1_score": 0.6704394720403158,
916
+ "eval_loss": 1.0488700866699219,
917
+ "eval_runtime": 104.8654,
918
+ "eval_samples_per_second": 198.168,
919
+ "eval_steps_per_second": 2.069,
920
+ "step": 14250
921
+ },
922
+ {
923
+ "epoch": 3.53,
924
+ "learning_rate": 6.118355065195587e-06,
925
+ "loss": 0.7492,
926
+ "step": 14500
927
+ },
928
+ {
929
+ "epoch": 3.53,
930
+ "eval_accuracy": 0.6762908425966027,
931
+ "eval_f1_score": 0.668875699854832,
932
+ "eval_loss": 1.0487704277038574,
933
+ "eval_runtime": 105.1614,
934
+ "eval_samples_per_second": 197.611,
935
+ "eval_steps_per_second": 2.063,
936
+ "step": 14500
937
+ },
938
+ {
939
+ "epoch": 3.59,
940
+ "learning_rate": 5.334754262788365e-06,
941
+ "loss": 0.7534,
942
+ "step": 14750
943
+ },
944
+ {
945
+ "epoch": 3.59,
946
+ "eval_accuracy": 0.6767239305134498,
947
+ "eval_f1_score": 0.6693352825093323,
948
+ "eval_loss": 1.0485919713974,
949
+ "eval_runtime": 105.4042,
950
+ "eval_samples_per_second": 197.155,
951
+ "eval_steps_per_second": 2.059,
952
+ "step": 14750
953
+ },
954
+ {
955
+ "epoch": 3.65,
956
+ "learning_rate": 4.551153460381144e-06,
957
+ "loss": 0.7391,
958
+ "step": 15000
959
+ },
960
+ {
961
+ "epoch": 3.65,
962
+ "eval_accuracy": 0.6791299744959338,
963
+ "eval_f1_score": 0.6716788527298224,
964
+ "eval_loss": 1.0460212230682373,
965
+ "eval_runtime": 105.024,
966
+ "eval_samples_per_second": 197.869,
967
+ "eval_steps_per_second": 2.066,
968
+ "step": 15000
969
+ },
970
+ {
971
+ "epoch": 3.71,
972
+ "learning_rate": 3.7675526579739217e-06,
973
+ "loss": 0.7543,
974
+ "step": 15250
975
+ },
976
+ {
977
+ "epoch": 3.71,
978
+ "eval_accuracy": 0.6804773591261248,
979
+ "eval_f1_score": 0.6726007691455318,
980
+ "eval_loss": 1.045469880104065,
981
+ "eval_runtime": 104.6353,
982
+ "eval_samples_per_second": 198.604,
983
+ "eval_steps_per_second": 2.074,
984
+ "step": 15250
985
+ },
986
+ {
987
+ "epoch": 3.77,
988
+ "learning_rate": 2.9839518555667003e-06,
989
+ "loss": 0.7382,
990
+ "step": 15500
991
+ },
992
+ {
993
+ "epoch": 3.77,
994
+ "eval_accuracy": 0.6802848756075262,
995
+ "eval_f1_score": 0.6728396464392151,
996
+ "eval_loss": 1.0436148643493652,
997
+ "eval_runtime": 105.0033,
998
+ "eval_samples_per_second": 197.908,
999
+ "eval_steps_per_second": 2.067,
1000
+ "step": 15500
1001
+ }
1002
+ ],
1003
+ "max_steps": 16452,
1004
+ "num_train_epochs": 4,
1005
+ "total_flos": 1.9763710480077926e+17,
1006
+ "trial_name": null,
1007
+ "trial_params": null
1008
+ }