sudo-s commited on
Commit
fb60bf4
1 Parent(s): a0b5533

🍻 cheers

Browse files
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.827810650887574,
4
+ "eval_loss": 0.6401488780975342,
5
+ "eval_runtime": 19.2501,
6
+ "eval_samples_per_second": 87.792,
7
+ "eval_steps_per_second": 11.013,
8
+ "total_flos": 2.1353892586861363e+18,
9
+ "train_loss": 1.3106378566387087,
10
+ "train_runtime": 1021.3737,
11
+ "train_samples_per_second": 26.944,
12
+ "train_steps_per_second": 1.684
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.827810650887574,
4
+ "eval_loss": 0.6401488780975342,
5
+ "eval_runtime": 19.2501,
6
+ "eval_samples_per_second": 87.792,
7
+ "eval_steps_per_second": 11.013
8
+ }
runs/Jul22_10-59-53_ef9e17783266/events.out.tfevents.1658488982.ef9e17783266.78.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5450c72234627e1b937b9b39debeaa802a491210eb2919a0545f69dfeb7c2e1b
3
+ size 363
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "total_flos": 2.1353892586861363e+18,
4
+ "train_loss": 1.3106378566387087,
5
+ "train_runtime": 1021.3737,
6
+ "train_samples_per_second": 26.944,
7
+ "train_steps_per_second": 1.684
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1210 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.6401488780975342,
3
+ "best_model_checkpoint": "./exper1_mesum5/checkpoint-1700",
4
+ "epoch": 4.0,
5
+ "global_step": 1720,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 0.0001988372093023256,
13
+ "loss": 4.9469,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 0.00019767441860465116,
19
+ "loss": 4.8606,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 0.00019651162790697676,
25
+ "loss": 4.6917,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.09,
30
+ "learning_rate": 0.00019534883720930232,
31
+ "loss": 4.5534,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 0.0001941860465116279,
37
+ "loss": 4.4467,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 0.0001930232558139535,
43
+ "loss": 4.2795,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.16,
48
+ "learning_rate": 0.0001918604651162791,
49
+ "loss": 4.2333,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 0.00019069767441860466,
55
+ "loss": 4.099,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 0.00018953488372093025,
61
+ "loss": 3.9539,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.23,
66
+ "learning_rate": 0.00018837209302325584,
67
+ "loss": 3.9352,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.23,
72
+ "eval_accuracy": 0.19585798816568048,
73
+ "eval_loss": 3.8550209999084473,
74
+ "eval_runtime": 19.1544,
75
+ "eval_samples_per_second": 88.23,
76
+ "eval_steps_per_second": 11.068,
77
+ "step": 100
78
+ },
79
+ {
80
+ "epoch": 0.26,
81
+ "learning_rate": 0.0001872093023255814,
82
+ "loss": 3.8303,
83
+ "step": 110
84
+ },
85
+ {
86
+ "epoch": 0.28,
87
+ "learning_rate": 0.000186046511627907,
88
+ "loss": 3.7242,
89
+ "step": 120
90
+ },
91
+ {
92
+ "epoch": 0.3,
93
+ "learning_rate": 0.00018488372093023256,
94
+ "loss": 3.5167,
95
+ "step": 130
96
+ },
97
+ {
98
+ "epoch": 0.33,
99
+ "learning_rate": 0.00018372093023255815,
100
+ "loss": 3.57,
101
+ "step": 140
102
+ },
103
+ {
104
+ "epoch": 0.35,
105
+ "learning_rate": 0.0001825581395348837,
106
+ "loss": 3.5231,
107
+ "step": 150
108
+ },
109
+ {
110
+ "epoch": 0.37,
111
+ "learning_rate": 0.0001813953488372093,
112
+ "loss": 3.4206,
113
+ "step": 160
114
+ },
115
+ {
116
+ "epoch": 0.4,
117
+ "learning_rate": 0.0001802325581395349,
118
+ "loss": 3.4456,
119
+ "step": 170
120
+ },
121
+ {
122
+ "epoch": 0.42,
123
+ "learning_rate": 0.00017906976744186048,
124
+ "loss": 3.3114,
125
+ "step": 180
126
+ },
127
+ {
128
+ "epoch": 0.44,
129
+ "learning_rate": 0.00017790697674418605,
130
+ "loss": 3.3379,
131
+ "step": 190
132
+ },
133
+ {
134
+ "epoch": 0.47,
135
+ "learning_rate": 0.00017674418604651164,
136
+ "loss": 3.1536,
137
+ "step": 200
138
+ },
139
+ {
140
+ "epoch": 0.47,
141
+ "eval_accuracy": 0.28875739644970416,
142
+ "eval_loss": 3.175450086593628,
143
+ "eval_runtime": 21.4845,
144
+ "eval_samples_per_second": 78.662,
145
+ "eval_steps_per_second": 9.868,
146
+ "step": 200
147
+ },
148
+ {
149
+ "epoch": 0.49,
150
+ "learning_rate": 0.00017558139534883723,
151
+ "loss": 3.0307,
152
+ "step": 210
153
+ },
154
+ {
155
+ "epoch": 0.51,
156
+ "learning_rate": 0.0001744186046511628,
157
+ "loss": 3.0889,
158
+ "step": 220
159
+ },
160
+ {
161
+ "epoch": 0.53,
162
+ "learning_rate": 0.00017325581395348838,
163
+ "loss": 3.0574,
164
+ "step": 230
165
+ },
166
+ {
167
+ "epoch": 0.56,
168
+ "learning_rate": 0.00017209302325581395,
169
+ "loss": 3.1058,
170
+ "step": 240
171
+ },
172
+ {
173
+ "epoch": 0.58,
174
+ "learning_rate": 0.00017093023255813954,
175
+ "loss": 2.9981,
176
+ "step": 250
177
+ },
178
+ {
179
+ "epoch": 0.6,
180
+ "learning_rate": 0.0001697674418604651,
181
+ "loss": 2.8662,
182
+ "step": 260
183
+ },
184
+ {
185
+ "epoch": 0.63,
186
+ "learning_rate": 0.00016860465116279072,
187
+ "loss": 2.7418,
188
+ "step": 270
189
+ },
190
+ {
191
+ "epoch": 0.65,
192
+ "learning_rate": 0.00016744186046511629,
193
+ "loss": 2.9494,
194
+ "step": 280
195
+ },
196
+ {
197
+ "epoch": 0.67,
198
+ "learning_rate": 0.00016639534883720932,
199
+ "loss": 2.7704,
200
+ "step": 290
201
+ },
202
+ {
203
+ "epoch": 0.7,
204
+ "learning_rate": 0.00016523255813953488,
205
+ "loss": 2.6937,
206
+ "step": 300
207
+ },
208
+ {
209
+ "epoch": 0.7,
210
+ "eval_accuracy": 0.4272189349112426,
211
+ "eval_loss": 2.633167266845703,
212
+ "eval_runtime": 21.2397,
213
+ "eval_samples_per_second": 79.568,
214
+ "eval_steps_per_second": 9.981,
215
+ "step": 300
216
+ },
217
+ {
218
+ "epoch": 0.72,
219
+ "learning_rate": 0.00016406976744186047,
220
+ "loss": 2.6292,
221
+ "step": 310
222
+ },
223
+ {
224
+ "epoch": 0.74,
225
+ "learning_rate": 0.00016290697674418604,
226
+ "loss": 2.641,
227
+ "step": 320
228
+ },
229
+ {
230
+ "epoch": 0.77,
231
+ "learning_rate": 0.00016174418604651163,
232
+ "loss": 2.6534,
233
+ "step": 330
234
+ },
235
+ {
236
+ "epoch": 0.79,
237
+ "learning_rate": 0.00016058139534883722,
238
+ "loss": 2.5655,
239
+ "step": 340
240
+ },
241
+ {
242
+ "epoch": 0.81,
243
+ "learning_rate": 0.0001594186046511628,
244
+ "loss": 2.4129,
245
+ "step": 350
246
+ },
247
+ {
248
+ "epoch": 0.84,
249
+ "learning_rate": 0.00015825581395348837,
250
+ "loss": 2.3514,
251
+ "step": 360
252
+ },
253
+ {
254
+ "epoch": 0.86,
255
+ "learning_rate": 0.00015709302325581396,
256
+ "loss": 2.4527,
257
+ "step": 370
258
+ },
259
+ {
260
+ "epoch": 0.88,
261
+ "learning_rate": 0.00015593023255813955,
262
+ "loss": 2.4411,
263
+ "step": 380
264
+ },
265
+ {
266
+ "epoch": 0.91,
267
+ "learning_rate": 0.00015476744186046512,
268
+ "loss": 2.2639,
269
+ "step": 390
270
+ },
271
+ {
272
+ "epoch": 0.93,
273
+ "learning_rate": 0.0001536046511627907,
274
+ "loss": 2.3748,
275
+ "step": 400
276
+ },
277
+ {
278
+ "epoch": 0.93,
279
+ "eval_accuracy": 0.4970414201183432,
280
+ "eval_loss": 2.28334641456604,
281
+ "eval_runtime": 21.0937,
282
+ "eval_samples_per_second": 80.119,
283
+ "eval_steps_per_second": 10.05,
284
+ "step": 400
285
+ },
286
+ {
287
+ "epoch": 0.95,
288
+ "learning_rate": 0.00015244186046511627,
289
+ "loss": 2.3598,
290
+ "step": 410
291
+ },
292
+ {
293
+ "epoch": 0.98,
294
+ "learning_rate": 0.00015127906976744186,
295
+ "loss": 2.3091,
296
+ "step": 420
297
+ },
298
+ {
299
+ "epoch": 1.0,
300
+ "learning_rate": 0.00015011627906976743,
301
+ "loss": 2.1695,
302
+ "step": 430
303
+ },
304
+ {
305
+ "epoch": 1.02,
306
+ "learning_rate": 0.00014895348837209302,
307
+ "loss": 1.951,
308
+ "step": 440
309
+ },
310
+ {
311
+ "epoch": 1.05,
312
+ "learning_rate": 0.0001477906976744186,
313
+ "loss": 1.8001,
314
+ "step": 450
315
+ },
316
+ {
317
+ "epoch": 1.07,
318
+ "learning_rate": 0.0001466279069767442,
319
+ "loss": 1.7164,
320
+ "step": 460
321
+ },
322
+ {
323
+ "epoch": 1.09,
324
+ "learning_rate": 0.0001454651162790698,
325
+ "loss": 1.8812,
326
+ "step": 470
327
+ },
328
+ {
329
+ "epoch": 1.12,
330
+ "learning_rate": 0.00014430232558139536,
331
+ "loss": 1.7576,
332
+ "step": 480
333
+ },
334
+ {
335
+ "epoch": 1.14,
336
+ "learning_rate": 0.00014313953488372095,
337
+ "loss": 1.7459,
338
+ "step": 490
339
+ },
340
+ {
341
+ "epoch": 1.16,
342
+ "learning_rate": 0.0001419767441860465,
343
+ "loss": 1.5575,
344
+ "step": 500
345
+ },
346
+ {
347
+ "epoch": 1.16,
348
+ "eval_accuracy": 0.5887573964497042,
349
+ "eval_loss": 1.8712348937988281,
350
+ "eval_runtime": 21.4456,
351
+ "eval_samples_per_second": 78.804,
352
+ "eval_steps_per_second": 9.885,
353
+ "step": 500
354
+ },
355
+ {
356
+ "epoch": 1.19,
357
+ "learning_rate": 0.0001408139534883721,
358
+ "loss": 1.6987,
359
+ "step": 510
360
+ },
361
+ {
362
+ "epoch": 1.21,
363
+ "learning_rate": 0.00013965116279069767,
364
+ "loss": 1.6239,
365
+ "step": 520
366
+ },
367
+ {
368
+ "epoch": 1.23,
369
+ "learning_rate": 0.00013848837209302326,
370
+ "loss": 1.656,
371
+ "step": 530
372
+ },
373
+ {
374
+ "epoch": 1.26,
375
+ "learning_rate": 0.00013732558139534885,
376
+ "loss": 1.43,
377
+ "step": 540
378
+ },
379
+ {
380
+ "epoch": 1.28,
381
+ "learning_rate": 0.00013616279069767444,
382
+ "loss": 1.5609,
383
+ "step": 550
384
+ },
385
+ {
386
+ "epoch": 1.3,
387
+ "learning_rate": 0.00013500000000000003,
388
+ "loss": 1.5339,
389
+ "step": 560
390
+ },
391
+ {
392
+ "epoch": 1.33,
393
+ "learning_rate": 0.0001338372093023256,
394
+ "loss": 1.4529,
395
+ "step": 570
396
+ },
397
+ {
398
+ "epoch": 1.35,
399
+ "learning_rate": 0.00013267441860465118,
400
+ "loss": 1.453,
401
+ "step": 580
402
+ },
403
+ {
404
+ "epoch": 1.37,
405
+ "learning_rate": 0.00013151162790697675,
406
+ "loss": 1.4002,
407
+ "step": 590
408
+ },
409
+ {
410
+ "epoch": 1.4,
411
+ "learning_rate": 0.00013034883720930234,
412
+ "loss": 1.4063,
413
+ "step": 600
414
+ },
415
+ {
416
+ "epoch": 1.4,
417
+ "eval_accuracy": 0.6313609467455621,
418
+ "eval_loss": 1.6047611236572266,
419
+ "eval_runtime": 21.4699,
420
+ "eval_samples_per_second": 78.715,
421
+ "eval_steps_per_second": 9.874,
422
+ "step": 600
423
+ },
424
+ {
425
+ "epoch": 1.42,
426
+ "learning_rate": 0.0001291860465116279,
427
+ "loss": 1.4063,
428
+ "step": 610
429
+ },
430
+ {
431
+ "epoch": 1.44,
432
+ "learning_rate": 0.0001280232558139535,
433
+ "loss": 1.3656,
434
+ "step": 620
435
+ },
436
+ {
437
+ "epoch": 1.47,
438
+ "learning_rate": 0.00012686046511627906,
439
+ "loss": 1.2191,
440
+ "step": 630
441
+ },
442
+ {
443
+ "epoch": 1.49,
444
+ "learning_rate": 0.00012569767441860465,
445
+ "loss": 1.3991,
446
+ "step": 640
447
+ },
448
+ {
449
+ "epoch": 1.51,
450
+ "learning_rate": 0.00012453488372093024,
451
+ "loss": 1.2237,
452
+ "step": 650
453
+ },
454
+ {
455
+ "epoch": 1.53,
456
+ "learning_rate": 0.00012337209302325583,
457
+ "loss": 1.1939,
458
+ "step": 660
459
+ },
460
+ {
461
+ "epoch": 1.56,
462
+ "learning_rate": 0.00012220930232558142,
463
+ "loss": 1.2683,
464
+ "step": 670
465
+ },
466
+ {
467
+ "epoch": 1.58,
468
+ "learning_rate": 0.00012104651162790698,
469
+ "loss": 1.2969,
470
+ "step": 680
471
+ },
472
+ {
473
+ "epoch": 1.6,
474
+ "learning_rate": 0.00011988372093023256,
475
+ "loss": 1.1475,
476
+ "step": 690
477
+ },
478
+ {
479
+ "epoch": 1.63,
480
+ "learning_rate": 0.00011872093023255815,
481
+ "loss": 1.1841,
482
+ "step": 700
483
+ },
484
+ {
485
+ "epoch": 1.63,
486
+ "eval_accuracy": 0.6621301775147929,
487
+ "eval_loss": 1.4109498262405396,
488
+ "eval_runtime": 19.5984,
489
+ "eval_samples_per_second": 86.231,
490
+ "eval_steps_per_second": 10.817,
491
+ "step": 700
492
+ },
493
+ {
494
+ "epoch": 1.65,
495
+ "learning_rate": 0.00011755813953488373,
496
+ "loss": 1.2236,
497
+ "step": 710
498
+ },
499
+ {
500
+ "epoch": 1.67,
501
+ "learning_rate": 0.00011639534883720931,
502
+ "loss": 1.3392,
503
+ "step": 720
504
+ },
505
+ {
506
+ "epoch": 1.7,
507
+ "learning_rate": 0.00011523255813953489,
508
+ "loss": 1.2305,
509
+ "step": 730
510
+ },
511
+ {
512
+ "epoch": 1.72,
513
+ "learning_rate": 0.00011406976744186046,
514
+ "loss": 1.1766,
515
+ "step": 740
516
+ },
517
+ {
518
+ "epoch": 1.74,
519
+ "learning_rate": 0.00011290697674418604,
520
+ "loss": 1.164,
521
+ "step": 750
522
+ },
523
+ {
524
+ "epoch": 1.77,
525
+ "learning_rate": 0.00011174418604651162,
526
+ "loss": 1.2152,
527
+ "step": 760
528
+ },
529
+ {
530
+ "epoch": 1.79,
531
+ "learning_rate": 0.00011058139534883722,
532
+ "loss": 1.0498,
533
+ "step": 770
534
+ },
535
+ {
536
+ "epoch": 1.81,
537
+ "learning_rate": 0.0001094186046511628,
538
+ "loss": 1.1407,
539
+ "step": 780
540
+ },
541
+ {
542
+ "epoch": 1.84,
543
+ "learning_rate": 0.00010825581395348838,
544
+ "loss": 1.1325,
545
+ "step": 790
546
+ },
547
+ {
548
+ "epoch": 1.86,
549
+ "learning_rate": 0.00010709302325581397,
550
+ "loss": 1.0857,
551
+ "step": 800
552
+ },
553
+ {
554
+ "epoch": 1.86,
555
+ "eval_accuracy": 0.7112426035502959,
556
+ "eval_loss": 1.1831614971160889,
557
+ "eval_runtime": 19.5421,
558
+ "eval_samples_per_second": 86.48,
559
+ "eval_steps_per_second": 10.848,
560
+ "step": 800
561
+ },
562
+ {
563
+ "epoch": 1.88,
564
+ "learning_rate": 0.00010593023255813955,
565
+ "loss": 0.9263,
566
+ "step": 810
567
+ },
568
+ {
569
+ "epoch": 1.91,
570
+ "learning_rate": 0.00010476744186046512,
571
+ "loss": 0.9742,
572
+ "step": 820
573
+ },
574
+ {
575
+ "epoch": 1.93,
576
+ "learning_rate": 0.0001036046511627907,
577
+ "loss": 1.0176,
578
+ "step": 830
579
+ },
580
+ {
581
+ "epoch": 1.95,
582
+ "learning_rate": 0.00010244186046511628,
583
+ "loss": 1.0105,
584
+ "step": 840
585
+ },
586
+ {
587
+ "epoch": 1.98,
588
+ "learning_rate": 0.00010127906976744185,
589
+ "loss": 1.0131,
590
+ "step": 850
591
+ },
592
+ {
593
+ "epoch": 2.0,
594
+ "learning_rate": 0.00010011627906976743,
595
+ "loss": 0.9204,
596
+ "step": 860
597
+ },
598
+ {
599
+ "epoch": 2.02,
600
+ "learning_rate": 9.895348837209302e-05,
601
+ "loss": 0.6221,
602
+ "step": 870
603
+ },
604
+ {
605
+ "epoch": 2.05,
606
+ "learning_rate": 9.779069767441861e-05,
607
+ "loss": 0.6229,
608
+ "step": 880
609
+ },
610
+ {
611
+ "epoch": 2.07,
612
+ "learning_rate": 9.662790697674419e-05,
613
+ "loss": 0.7227,
614
+ "step": 890
615
+ },
616
+ {
617
+ "epoch": 2.09,
618
+ "learning_rate": 9.546511627906978e-05,
619
+ "loss": 0.582,
620
+ "step": 900
621
+ },
622
+ {
623
+ "epoch": 2.09,
624
+ "eval_accuracy": 0.7479289940828402,
625
+ "eval_loss": 1.0371233224868774,
626
+ "eval_runtime": 19.7303,
627
+ "eval_samples_per_second": 85.655,
628
+ "eval_steps_per_second": 10.745,
629
+ "step": 900
630
+ },
631
+ {
632
+ "epoch": 2.12,
633
+ "learning_rate": 9.430232558139536e-05,
634
+ "loss": 0.6454,
635
+ "step": 910
636
+ },
637
+ {
638
+ "epoch": 2.14,
639
+ "learning_rate": 9.313953488372094e-05,
640
+ "loss": 0.6227,
641
+ "step": 920
642
+ },
643
+ {
644
+ "epoch": 2.16,
645
+ "learning_rate": 9.197674418604651e-05,
646
+ "loss": 0.4867,
647
+ "step": 930
648
+ },
649
+ {
650
+ "epoch": 2.19,
651
+ "learning_rate": 9.081395348837209e-05,
652
+ "loss": 0.4923,
653
+ "step": 940
654
+ },
655
+ {
656
+ "epoch": 2.21,
657
+ "learning_rate": 8.965116279069767e-05,
658
+ "loss": 0.5217,
659
+ "step": 950
660
+ },
661
+ {
662
+ "epoch": 2.23,
663
+ "learning_rate": 8.848837209302326e-05,
664
+ "loss": 0.4372,
665
+ "step": 960
666
+ },
667
+ {
668
+ "epoch": 2.26,
669
+ "learning_rate": 8.732558139534884e-05,
670
+ "loss": 0.567,
671
+ "step": 970
672
+ },
673
+ {
674
+ "epoch": 2.28,
675
+ "learning_rate": 8.616279069767443e-05,
676
+ "loss": 0.5837,
677
+ "step": 980
678
+ },
679
+ {
680
+ "epoch": 2.3,
681
+ "learning_rate": 8.5e-05,
682
+ "loss": 0.5339,
683
+ "step": 990
684
+ },
685
+ {
686
+ "epoch": 2.33,
687
+ "learning_rate": 8.383720930232558e-05,
688
+ "loss": 0.5971,
689
+ "step": 1000
690
+ },
691
+ {
692
+ "epoch": 2.33,
693
+ "eval_accuracy": 0.7461538461538462,
694
+ "eval_loss": 0.9839160442352295,
695
+ "eval_runtime": 19.6324,
696
+ "eval_samples_per_second": 86.082,
697
+ "eval_steps_per_second": 10.798,
698
+ "step": 1000
699
+ },
700
+ {
701
+ "epoch": 2.35,
702
+ "learning_rate": 8.267441860465117e-05,
703
+ "loss": 0.5588,
704
+ "step": 1010
705
+ },
706
+ {
707
+ "epoch": 2.37,
708
+ "learning_rate": 8.151162790697675e-05,
709
+ "loss": 0.5793,
710
+ "step": 1020
711
+ },
712
+ {
713
+ "epoch": 2.4,
714
+ "learning_rate": 8.034883720930233e-05,
715
+ "loss": 0.475,
716
+ "step": 1030
717
+ },
718
+ {
719
+ "epoch": 2.42,
720
+ "learning_rate": 7.91860465116279e-05,
721
+ "loss": 0.5391,
722
+ "step": 1040
723
+ },
724
+ {
725
+ "epoch": 2.44,
726
+ "learning_rate": 7.802325581395348e-05,
727
+ "loss": 0.487,
728
+ "step": 1050
729
+ },
730
+ {
731
+ "epoch": 2.47,
732
+ "learning_rate": 7.686046511627908e-05,
733
+ "loss": 0.4474,
734
+ "step": 1060
735
+ },
736
+ {
737
+ "epoch": 2.49,
738
+ "learning_rate": 7.569767441860465e-05,
739
+ "loss": 0.4753,
740
+ "step": 1070
741
+ },
742
+ {
743
+ "epoch": 2.51,
744
+ "learning_rate": 7.453488372093024e-05,
745
+ "loss": 0.3932,
746
+ "step": 1080
747
+ },
748
+ {
749
+ "epoch": 2.53,
750
+ "learning_rate": 7.337209302325582e-05,
751
+ "loss": 0.4907,
752
+ "step": 1090
753
+ },
754
+ {
755
+ "epoch": 2.56,
756
+ "learning_rate": 7.22093023255814e-05,
757
+ "loss": 0.4617,
758
+ "step": 1100
759
+ },
760
+ {
761
+ "epoch": 2.56,
762
+ "eval_accuracy": 0.7656804733727811,
763
+ "eval_loss": 0.9233230948448181,
764
+ "eval_runtime": 19.3219,
765
+ "eval_samples_per_second": 87.465,
766
+ "eval_steps_per_second": 10.972,
767
+ "step": 1100
768
+ },
769
+ {
770
+ "epoch": 2.58,
771
+ "learning_rate": 7.104651162790698e-05,
772
+ "loss": 0.4143,
773
+ "step": 1110
774
+ },
775
+ {
776
+ "epoch": 2.6,
777
+ "learning_rate": 6.988372093023257e-05,
778
+ "loss": 0.5355,
779
+ "step": 1120
780
+ },
781
+ {
782
+ "epoch": 2.63,
783
+ "learning_rate": 6.872093023255814e-05,
784
+ "loss": 0.4717,
785
+ "step": 1130
786
+ },
787
+ {
788
+ "epoch": 2.65,
789
+ "learning_rate": 6.755813953488372e-05,
790
+ "loss": 0.3827,
791
+ "step": 1140
792
+ },
793
+ {
794
+ "epoch": 2.67,
795
+ "learning_rate": 6.63953488372093e-05,
796
+ "loss": 0.443,
797
+ "step": 1150
798
+ },
799
+ {
800
+ "epoch": 2.7,
801
+ "learning_rate": 6.523255813953488e-05,
802
+ "loss": 0.4257,
803
+ "step": 1160
804
+ },
805
+ {
806
+ "epoch": 2.72,
807
+ "learning_rate": 6.406976744186047e-05,
808
+ "loss": 0.3875,
809
+ "step": 1170
810
+ },
811
+ {
812
+ "epoch": 2.74,
813
+ "learning_rate": 6.290697674418606e-05,
814
+ "loss": 0.3956,
815
+ "step": 1180
816
+ },
817
+ {
818
+ "epoch": 2.77,
819
+ "learning_rate": 6.174418604651164e-05,
820
+ "loss": 0.3629,
821
+ "step": 1190
822
+ },
823
+ {
824
+ "epoch": 2.79,
825
+ "learning_rate": 6.058139534883721e-05,
826
+ "loss": 0.4621,
827
+ "step": 1200
828
+ },
829
+ {
830
+ "epoch": 2.79,
831
+ "eval_accuracy": 0.7828402366863906,
832
+ "eval_loss": 0.8416844606399536,
833
+ "eval_runtime": 19.2964,
834
+ "eval_samples_per_second": 87.581,
835
+ "eval_steps_per_second": 10.986,
836
+ "step": 1200
837
+ },
838
+ {
839
+ "epoch": 2.81,
840
+ "learning_rate": 5.941860465116279e-05,
841
+ "loss": 0.4833,
842
+ "step": 1210
843
+ },
844
+ {
845
+ "epoch": 2.84,
846
+ "learning_rate": 5.825581395348837e-05,
847
+ "loss": 0.3749,
848
+ "step": 1220
849
+ },
850
+ {
851
+ "epoch": 2.86,
852
+ "learning_rate": 5.709302325581396e-05,
853
+ "loss": 0.4963,
854
+ "step": 1230
855
+ },
856
+ {
857
+ "epoch": 2.88,
858
+ "learning_rate": 5.5930232558139536e-05,
859
+ "loss": 0.2743,
860
+ "step": 1240
861
+ },
862
+ {
863
+ "epoch": 2.91,
864
+ "learning_rate": 5.476744186046512e-05,
865
+ "loss": 0.3608,
866
+ "step": 1250
867
+ },
868
+ {
869
+ "epoch": 2.93,
870
+ "learning_rate": 5.36046511627907e-05,
871
+ "loss": 0.3941,
872
+ "step": 1260
873
+ },
874
+ {
875
+ "epoch": 2.95,
876
+ "learning_rate": 5.2441860465116275e-05,
877
+ "loss": 0.3464,
878
+ "step": 1270
879
+ },
880
+ {
881
+ "epoch": 2.98,
882
+ "learning_rate": 5.1279069767441866e-05,
883
+ "loss": 0.3461,
884
+ "step": 1280
885
+ },
886
+ {
887
+ "epoch": 3.0,
888
+ "learning_rate": 5.011627906976745e-05,
889
+ "loss": 0.368,
890
+ "step": 1290
891
+ },
892
+ {
893
+ "epoch": 3.02,
894
+ "learning_rate": 4.895348837209303e-05,
895
+ "loss": 0.2128,
896
+ "step": 1300
897
+ },
898
+ {
899
+ "epoch": 3.02,
900
+ "eval_accuracy": 0.7970414201183432,
901
+ "eval_loss": 0.76435786485672,
902
+ "eval_runtime": 19.4425,
903
+ "eval_samples_per_second": 86.923,
904
+ "eval_steps_per_second": 10.904,
905
+ "step": 1300
906
+ },
907
+ {
908
+ "epoch": 3.05,
909
+ "learning_rate": 4.7790697674418605e-05,
910
+ "loss": 0.2484,
911
+ "step": 1310
912
+ },
913
+ {
914
+ "epoch": 3.07,
915
+ "learning_rate": 4.662790697674419e-05,
916
+ "loss": 0.2213,
917
+ "step": 1320
918
+ },
919
+ {
920
+ "epoch": 3.09,
921
+ "learning_rate": 4.5465116279069766e-05,
922
+ "loss": 0.2305,
923
+ "step": 1330
924
+ },
925
+ {
926
+ "epoch": 3.12,
927
+ "learning_rate": 4.430232558139535e-05,
928
+ "loss": 0.208,
929
+ "step": 1340
930
+ },
931
+ {
932
+ "epoch": 3.14,
933
+ "learning_rate": 4.3139534883720935e-05,
934
+ "loss": 0.2191,
935
+ "step": 1350
936
+ },
937
+ {
938
+ "epoch": 3.16,
939
+ "learning_rate": 4.197674418604651e-05,
940
+ "loss": 0.2032,
941
+ "step": 1360
942
+ },
943
+ {
944
+ "epoch": 3.19,
945
+ "learning_rate": 4.0813953488372096e-05,
946
+ "loss": 0.216,
947
+ "step": 1370
948
+ },
949
+ {
950
+ "epoch": 3.21,
951
+ "learning_rate": 3.9651162790697674e-05,
952
+ "loss": 0.1977,
953
+ "step": 1380
954
+ },
955
+ {
956
+ "epoch": 3.23,
957
+ "learning_rate": 3.848837209302326e-05,
958
+ "loss": 0.1767,
959
+ "step": 1390
960
+ },
961
+ {
962
+ "epoch": 3.26,
963
+ "learning_rate": 3.732558139534884e-05,
964
+ "loss": 0.1883,
965
+ "step": 1400
966
+ },
967
+ {
968
+ "epoch": 3.26,
969
+ "eval_accuracy": 0.8183431952662722,
970
+ "eval_loss": 0.7001345157623291,
971
+ "eval_runtime": 19.4125,
972
+ "eval_samples_per_second": 87.057,
973
+ "eval_steps_per_second": 10.921,
974
+ "step": 1400
975
+ },
976
+ {
977
+ "epoch": 3.28,
978
+ "learning_rate": 3.616279069767442e-05,
979
+ "loss": 0.1768,
980
+ "step": 1410
981
+ },
982
+ {
983
+ "epoch": 3.3,
984
+ "learning_rate": 3.5e-05,
985
+ "loss": 0.1436,
986
+ "step": 1420
987
+ },
988
+ {
989
+ "epoch": 3.33,
990
+ "learning_rate": 3.383720930232558e-05,
991
+ "loss": 0.1781,
992
+ "step": 1430
993
+ },
994
+ {
995
+ "epoch": 3.35,
996
+ "learning_rate": 3.2674418604651165e-05,
997
+ "loss": 0.1816,
998
+ "step": 1440
999
+ },
1000
+ {
1001
+ "epoch": 3.37,
1002
+ "learning_rate": 3.151162790697675e-05,
1003
+ "loss": 0.1516,
1004
+ "step": 1450
1005
+ },
1006
+ {
1007
+ "epoch": 3.4,
1008
+ "learning_rate": 3.0348837209302327e-05,
1009
+ "loss": 0.182,
1010
+ "step": 1460
1011
+ },
1012
+ {
1013
+ "epoch": 3.42,
1014
+ "learning_rate": 2.9186046511627908e-05,
1015
+ "loss": 0.1435,
1016
+ "step": 1470
1017
+ },
1018
+ {
1019
+ "epoch": 3.44,
1020
+ "learning_rate": 2.8023255813953492e-05,
1021
+ "loss": 0.1575,
1022
+ "step": 1480
1023
+ },
1024
+ {
1025
+ "epoch": 3.47,
1026
+ "learning_rate": 2.6860465116279073e-05,
1027
+ "loss": 0.1361,
1028
+ "step": 1490
1029
+ },
1030
+ {
1031
+ "epoch": 3.49,
1032
+ "learning_rate": 2.569767441860465e-05,
1033
+ "loss": 0.1501,
1034
+ "step": 1500
1035
+ },
1036
+ {
1037
+ "epoch": 3.49,
1038
+ "eval_accuracy": 0.8201183431952662,
1039
+ "eval_loss": 0.6825571060180664,
1040
+ "eval_runtime": 19.3331,
1041
+ "eval_samples_per_second": 87.415,
1042
+ "eval_steps_per_second": 10.966,
1043
+ "step": 1500
1044
+ },
1045
+ {
1046
+ "epoch": 3.51,
1047
+ "learning_rate": 2.453488372093023e-05,
1048
+ "loss": 0.1719,
1049
+ "step": 1510
1050
+ },
1051
+ {
1052
+ "epoch": 3.53,
1053
+ "learning_rate": 2.3372093023255815e-05,
1054
+ "loss": 0.2121,
1055
+ "step": 1520
1056
+ },
1057
+ {
1058
+ "epoch": 3.56,
1059
+ "learning_rate": 2.2209302325581396e-05,
1060
+ "loss": 0.1895,
1061
+ "step": 1530
1062
+ },
1063
+ {
1064
+ "epoch": 3.58,
1065
+ "learning_rate": 2.104651162790698e-05,
1066
+ "loss": 0.1466,
1067
+ "step": 1540
1068
+ },
1069
+ {
1070
+ "epoch": 3.6,
1071
+ "learning_rate": 1.9883720930232557e-05,
1072
+ "loss": 0.2179,
1073
+ "step": 1550
1074
+ },
1075
+ {
1076
+ "epoch": 3.63,
1077
+ "learning_rate": 1.8720930232558138e-05,
1078
+ "loss": 0.1219,
1079
+ "step": 1560
1080
+ },
1081
+ {
1082
+ "epoch": 3.65,
1083
+ "learning_rate": 1.7558139534883722e-05,
1084
+ "loss": 0.1275,
1085
+ "step": 1570
1086
+ },
1087
+ {
1088
+ "epoch": 3.67,
1089
+ "learning_rate": 1.6395348837209303e-05,
1090
+ "loss": 0.177,
1091
+ "step": 1580
1092
+ },
1093
+ {
1094
+ "epoch": 3.7,
1095
+ "learning_rate": 1.5232558139534884e-05,
1096
+ "loss": 0.1242,
1097
+ "step": 1590
1098
+ },
1099
+ {
1100
+ "epoch": 3.72,
1101
+ "learning_rate": 1.4069767441860465e-05,
1102
+ "loss": 0.1626,
1103
+ "step": 1600
1104
+ },
1105
+ {
1106
+ "epoch": 3.72,
1107
+ "eval_accuracy": 0.8254437869822485,
1108
+ "eval_loss": 0.6568493843078613,
1109
+ "eval_runtime": 19.3691,
1110
+ "eval_samples_per_second": 87.253,
1111
+ "eval_steps_per_second": 10.945,
1112
+ "step": 1600
1113
+ },
1114
+ {
1115
+ "epoch": 3.74,
1116
+ "learning_rate": 1.2906976744186047e-05,
1117
+ "loss": 0.1078,
1118
+ "step": 1610
1119
+ },
1120
+ {
1121
+ "epoch": 3.77,
1122
+ "learning_rate": 1.1744186046511628e-05,
1123
+ "loss": 0.1073,
1124
+ "step": 1620
1125
+ },
1126
+ {
1127
+ "epoch": 3.79,
1128
+ "learning_rate": 1.058139534883721e-05,
1129
+ "loss": 0.106,
1130
+ "step": 1630
1131
+ },
1132
+ {
1133
+ "epoch": 3.81,
1134
+ "learning_rate": 9.418604651162791e-06,
1135
+ "loss": 0.1707,
1136
+ "step": 1640
1137
+ },
1138
+ {
1139
+ "epoch": 3.84,
1140
+ "learning_rate": 8.255813953488372e-06,
1141
+ "loss": 0.1292,
1142
+ "step": 1650
1143
+ },
1144
+ {
1145
+ "epoch": 3.86,
1146
+ "learning_rate": 7.093023255813954e-06,
1147
+ "loss": 0.1311,
1148
+ "step": 1660
1149
+ },
1150
+ {
1151
+ "epoch": 3.88,
1152
+ "learning_rate": 5.930232558139535e-06,
1153
+ "loss": 0.1477,
1154
+ "step": 1670
1155
+ },
1156
+ {
1157
+ "epoch": 3.91,
1158
+ "learning_rate": 4.767441860465117e-06,
1159
+ "loss": 0.1223,
1160
+ "step": 1680
1161
+ },
1162
+ {
1163
+ "epoch": 3.93,
1164
+ "learning_rate": 3.604651162790698e-06,
1165
+ "loss": 0.1227,
1166
+ "step": 1690
1167
+ },
1168
+ {
1169
+ "epoch": 3.95,
1170
+ "learning_rate": 2.4418604651162793e-06,
1171
+ "loss": 0.1053,
1172
+ "step": 1700
1173
+ },
1174
+ {
1175
+ "epoch": 3.95,
1176
+ "eval_accuracy": 0.827810650887574,
1177
+ "eval_loss": 0.6401488780975342,
1178
+ "eval_runtime": 19.3377,
1179
+ "eval_samples_per_second": 87.394,
1180
+ "eval_steps_per_second": 10.963,
1181
+ "step": 1700
1182
+ },
1183
+ {
1184
+ "epoch": 3.98,
1185
+ "learning_rate": 1.2790697674418605e-06,
1186
+ "loss": 0.1335,
1187
+ "step": 1710
1188
+ },
1189
+ {
1190
+ "epoch": 4.0,
1191
+ "learning_rate": 1.1627906976744187e-07,
1192
+ "loss": 0.0918,
1193
+ "step": 1720
1194
+ },
1195
+ {
1196
+ "epoch": 4.0,
1197
+ "step": 1720,
1198
+ "total_flos": 2.1353892586861363e+18,
1199
+ "train_loss": 1.3106378566387087,
1200
+ "train_runtime": 1021.3737,
1201
+ "train_samples_per_second": 26.944,
1202
+ "train_steps_per_second": 1.684
1203
+ }
1204
+ ],
1205
+ "max_steps": 1720,
1206
+ "num_train_epochs": 4,
1207
+ "total_flos": 2.1353892586861363e+18,
1208
+ "trial_name": null,
1209
+ "trial_params": null
1210
+ }