Ricardo-H commited on
Commit
4a806b7
1 Parent(s): 4fefeef

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +13 -0
  2. eval_results.json +8 -0
  3. train_results.json +8 -0
  4. trainer_state.json +1176 -0
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_accuracy": 0.9758,
4
+ "eval_loss": 0.07598920166492462,
5
+ "eval_runtime": 20.502,
6
+ "eval_samples_per_second": 243.878,
7
+ "eval_steps_per_second": 3.853,
8
+ "total_flos": 1.11876953739264e+19,
9
+ "train_loss": 0.3102077527479692,
10
+ "train_runtime": 2497.672,
11
+ "train_samples_per_second": 180.168,
12
+ "train_steps_per_second": 0.705
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_accuracy": 0.9758,
4
+ "eval_loss": 0.07598920166492462,
5
+ "eval_runtime": 20.502,
6
+ "eval_samples_per_second": 243.878,
7
+ "eval_steps_per_second": 3.853
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "total_flos": 1.11876953739264e+19,
4
+ "train_loss": 0.3102077527479692,
5
+ "train_runtime": 2497.672,
6
+ "train_samples_per_second": 180.168,
7
+ "train_steps_per_second": 0.705
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1176 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9758,
3
+ "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-cifar10/checkpoint-1760",
4
+ "epoch": 10.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1760,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.06,
13
+ "learning_rate": 2.840909090909091e-06,
14
+ "loss": 0.402,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.11,
19
+ "learning_rate": 5.681818181818182e-06,
20
+ "loss": 0.3805,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.17,
25
+ "learning_rate": 8.522727272727273e-06,
26
+ "loss": 0.3874,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.23,
31
+ "learning_rate": 1.1363636363636365e-05,
32
+ "loss": 0.3618,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.28,
37
+ "learning_rate": 1.4204545454545456e-05,
38
+ "loss": 0.3786,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.34,
43
+ "learning_rate": 1.7045454545454546e-05,
44
+ "loss": 0.3798,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.4,
49
+ "learning_rate": 1.9886363636363638e-05,
50
+ "loss": 0.3811,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.45,
55
+ "learning_rate": 2.272727272727273e-05,
56
+ "loss": 0.3483,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.51,
61
+ "learning_rate": 2.5568181818181817e-05,
62
+ "loss": 0.3578,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.57,
67
+ "learning_rate": 2.8409090909090912e-05,
68
+ "loss": 0.3346,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.62,
73
+ "learning_rate": 3.125e-05,
74
+ "loss": 0.3361,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.68,
79
+ "learning_rate": 3.409090909090909e-05,
80
+ "loss": 0.3395,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.74,
85
+ "learning_rate": 3.6931818181818184e-05,
86
+ "loss": 0.3291,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.8,
91
+ "learning_rate": 3.9772727272727275e-05,
92
+ "loss": 0.3015,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.85,
97
+ "learning_rate": 4.261363636363637e-05,
98
+ "loss": 0.3264,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.91,
103
+ "learning_rate": 4.545454545454546e-05,
104
+ "loss": 0.3435,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.97,
109
+ "learning_rate": 4.829545454545455e-05,
110
+ "loss": 0.314,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 1.0,
115
+ "eval_accuracy": 0.9612,
116
+ "eval_loss": 0.12112396210432053,
117
+ "eval_runtime": 20.4608,
118
+ "eval_samples_per_second": 244.369,
119
+ "eval_steps_per_second": 3.861,
120
+ "step": 176
121
+ },
122
+ {
123
+ "epoch": 1.02,
124
+ "learning_rate": 4.9873737373737375e-05,
125
+ "loss": 0.2916,
126
+ "step": 180
127
+ },
128
+ {
129
+ "epoch": 1.08,
130
+ "learning_rate": 4.955808080808081e-05,
131
+ "loss": 0.3273,
132
+ "step": 190
133
+ },
134
+ {
135
+ "epoch": 1.14,
136
+ "learning_rate": 4.9242424242424245e-05,
137
+ "loss": 0.3315,
138
+ "step": 200
139
+ },
140
+ {
141
+ "epoch": 1.19,
142
+ "learning_rate": 4.892676767676767e-05,
143
+ "loss": 0.3236,
144
+ "step": 210
145
+ },
146
+ {
147
+ "epoch": 1.25,
148
+ "learning_rate": 4.8611111111111115e-05,
149
+ "loss": 0.2965,
150
+ "step": 220
151
+ },
152
+ {
153
+ "epoch": 1.31,
154
+ "learning_rate": 4.829545454545455e-05,
155
+ "loss": 0.3125,
156
+ "step": 230
157
+ },
158
+ {
159
+ "epoch": 1.36,
160
+ "learning_rate": 4.797979797979798e-05,
161
+ "loss": 0.308,
162
+ "step": 240
163
+ },
164
+ {
165
+ "epoch": 1.42,
166
+ "learning_rate": 4.7664141414141413e-05,
167
+ "loss": 0.2999,
168
+ "step": 250
169
+ },
170
+ {
171
+ "epoch": 1.48,
172
+ "learning_rate": 4.7348484848484855e-05,
173
+ "loss": 0.3437,
174
+ "step": 260
175
+ },
176
+ {
177
+ "epoch": 1.53,
178
+ "learning_rate": 4.7032828282828283e-05,
179
+ "loss": 0.3626,
180
+ "step": 270
181
+ },
182
+ {
183
+ "epoch": 1.59,
184
+ "learning_rate": 4.671717171717172e-05,
185
+ "loss": 0.3189,
186
+ "step": 280
187
+ },
188
+ {
189
+ "epoch": 1.65,
190
+ "learning_rate": 4.6401515151515154e-05,
191
+ "loss": 0.3152,
192
+ "step": 290
193
+ },
194
+ {
195
+ "epoch": 1.7,
196
+ "learning_rate": 4.608585858585859e-05,
197
+ "loss": 0.2978,
198
+ "step": 300
199
+ },
200
+ {
201
+ "epoch": 1.76,
202
+ "learning_rate": 4.5770202020202024e-05,
203
+ "loss": 0.3366,
204
+ "step": 310
205
+ },
206
+ {
207
+ "epoch": 1.82,
208
+ "learning_rate": 4.545454545454546e-05,
209
+ "loss": 0.3209,
210
+ "step": 320
211
+ },
212
+ {
213
+ "epoch": 1.88,
214
+ "learning_rate": 4.5138888888888894e-05,
215
+ "loss": 0.364,
216
+ "step": 330
217
+ },
218
+ {
219
+ "epoch": 1.93,
220
+ "learning_rate": 4.482323232323233e-05,
221
+ "loss": 0.3617,
222
+ "step": 340
223
+ },
224
+ {
225
+ "epoch": 1.99,
226
+ "learning_rate": 4.450757575757576e-05,
227
+ "loss": 0.2992,
228
+ "step": 350
229
+ },
230
+ {
231
+ "epoch": 2.0,
232
+ "eval_accuracy": 0.9622,
233
+ "eval_loss": 0.11857912689447403,
234
+ "eval_runtime": 20.7515,
235
+ "eval_samples_per_second": 240.947,
236
+ "eval_steps_per_second": 3.807,
237
+ "step": 352
238
+ },
239
+ {
240
+ "epoch": 2.05,
241
+ "learning_rate": 4.41919191919192e-05,
242
+ "loss": 0.3473,
243
+ "step": 360
244
+ },
245
+ {
246
+ "epoch": 2.1,
247
+ "learning_rate": 4.387626262626263e-05,
248
+ "loss": 0.3637,
249
+ "step": 370
250
+ },
251
+ {
252
+ "epoch": 2.16,
253
+ "learning_rate": 4.356060606060606e-05,
254
+ "loss": 0.3319,
255
+ "step": 380
256
+ },
257
+ {
258
+ "epoch": 2.22,
259
+ "learning_rate": 4.32449494949495e-05,
260
+ "loss": 0.3378,
261
+ "step": 390
262
+ },
263
+ {
264
+ "epoch": 2.27,
265
+ "learning_rate": 4.292929292929293e-05,
266
+ "loss": 0.3677,
267
+ "step": 400
268
+ },
269
+ {
270
+ "epoch": 2.33,
271
+ "learning_rate": 4.261363636363637e-05,
272
+ "loss": 0.3273,
273
+ "step": 410
274
+ },
275
+ {
276
+ "epoch": 2.39,
277
+ "learning_rate": 4.2297979797979795e-05,
278
+ "loss": 0.3504,
279
+ "step": 420
280
+ },
281
+ {
282
+ "epoch": 2.44,
283
+ "learning_rate": 4.198232323232324e-05,
284
+ "loss": 0.3253,
285
+ "step": 430
286
+ },
287
+ {
288
+ "epoch": 2.5,
289
+ "learning_rate": 4.166666666666667e-05,
290
+ "loss": 0.3388,
291
+ "step": 440
292
+ },
293
+ {
294
+ "epoch": 2.56,
295
+ "learning_rate": 4.13510101010101e-05,
296
+ "loss": 0.3406,
297
+ "step": 450
298
+ },
299
+ {
300
+ "epoch": 2.61,
301
+ "learning_rate": 4.1035353535353535e-05,
302
+ "loss": 0.3566,
303
+ "step": 460
304
+ },
305
+ {
306
+ "epoch": 2.67,
307
+ "learning_rate": 4.071969696969698e-05,
308
+ "loss": 0.3601,
309
+ "step": 470
310
+ },
311
+ {
312
+ "epoch": 2.73,
313
+ "learning_rate": 4.0404040404040405e-05,
314
+ "loss": 0.3421,
315
+ "step": 480
316
+ },
317
+ {
318
+ "epoch": 2.78,
319
+ "learning_rate": 4.008838383838384e-05,
320
+ "loss": 0.3579,
321
+ "step": 490
322
+ },
323
+ {
324
+ "epoch": 2.84,
325
+ "learning_rate": 3.9772727272727275e-05,
326
+ "loss": 0.379,
327
+ "step": 500
328
+ },
329
+ {
330
+ "epoch": 2.9,
331
+ "learning_rate": 3.945707070707071e-05,
332
+ "loss": 0.3544,
333
+ "step": 510
334
+ },
335
+ {
336
+ "epoch": 2.95,
337
+ "learning_rate": 3.9141414141414145e-05,
338
+ "loss": 0.3544,
339
+ "step": 520
340
+ },
341
+ {
342
+ "epoch": 3.0,
343
+ "eval_accuracy": 0.968,
344
+ "eval_loss": 0.09892657399177551,
345
+ "eval_runtime": 20.3604,
346
+ "eval_samples_per_second": 245.575,
347
+ "eval_steps_per_second": 3.88,
348
+ "step": 528
349
+ },
350
+ {
351
+ "epoch": 3.01,
352
+ "learning_rate": 3.8825757575757574e-05,
353
+ "loss": 0.3691,
354
+ "step": 530
355
+ },
356
+ {
357
+ "epoch": 3.07,
358
+ "learning_rate": 3.8510101010101015e-05,
359
+ "loss": 0.3584,
360
+ "step": 540
361
+ },
362
+ {
363
+ "epoch": 3.12,
364
+ "learning_rate": 3.8194444444444444e-05,
365
+ "loss": 0.3797,
366
+ "step": 550
367
+ },
368
+ {
369
+ "epoch": 3.18,
370
+ "learning_rate": 3.787878787878788e-05,
371
+ "loss": 0.3441,
372
+ "step": 560
373
+ },
374
+ {
375
+ "epoch": 3.24,
376
+ "learning_rate": 3.7563131313131314e-05,
377
+ "loss": 0.3692,
378
+ "step": 570
379
+ },
380
+ {
381
+ "epoch": 3.3,
382
+ "learning_rate": 3.724747474747475e-05,
383
+ "loss": 0.3625,
384
+ "step": 580
385
+ },
386
+ {
387
+ "epoch": 3.35,
388
+ "learning_rate": 3.6931818181818184e-05,
389
+ "loss": 0.3478,
390
+ "step": 590
391
+ },
392
+ {
393
+ "epoch": 3.41,
394
+ "learning_rate": 3.661616161616162e-05,
395
+ "loss": 0.3453,
396
+ "step": 600
397
+ },
398
+ {
399
+ "epoch": 3.47,
400
+ "learning_rate": 3.6300505050505054e-05,
401
+ "loss": 0.3723,
402
+ "step": 610
403
+ },
404
+ {
405
+ "epoch": 3.52,
406
+ "learning_rate": 3.598484848484849e-05,
407
+ "loss": 0.3609,
408
+ "step": 620
409
+ },
410
+ {
411
+ "epoch": 3.58,
412
+ "learning_rate": 3.566919191919192e-05,
413
+ "loss": 0.3288,
414
+ "step": 630
415
+ },
416
+ {
417
+ "epoch": 3.64,
418
+ "learning_rate": 3.535353535353535e-05,
419
+ "loss": 0.3409,
420
+ "step": 640
421
+ },
422
+ {
423
+ "epoch": 3.69,
424
+ "learning_rate": 3.5037878787878794e-05,
425
+ "loss": 0.3486,
426
+ "step": 650
427
+ },
428
+ {
429
+ "epoch": 3.75,
430
+ "learning_rate": 3.472222222222222e-05,
431
+ "loss": 0.3645,
432
+ "step": 660
433
+ },
434
+ {
435
+ "epoch": 3.81,
436
+ "learning_rate": 3.440656565656566e-05,
437
+ "loss": 0.3198,
438
+ "step": 670
439
+ },
440
+ {
441
+ "epoch": 3.86,
442
+ "learning_rate": 3.409090909090909e-05,
443
+ "loss": 0.3134,
444
+ "step": 680
445
+ },
446
+ {
447
+ "epoch": 3.92,
448
+ "learning_rate": 3.377525252525253e-05,
449
+ "loss": 0.3544,
450
+ "step": 690
451
+ },
452
+ {
453
+ "epoch": 3.98,
454
+ "learning_rate": 3.345959595959596e-05,
455
+ "loss": 0.3068,
456
+ "step": 700
457
+ },
458
+ {
459
+ "epoch": 4.0,
460
+ "eval_accuracy": 0.9724,
461
+ "eval_loss": 0.08715949207544327,
462
+ "eval_runtime": 20.9754,
463
+ "eval_samples_per_second": 238.375,
464
+ "eval_steps_per_second": 3.766,
465
+ "step": 704
466
+ },
467
+ {
468
+ "epoch": 4.03,
469
+ "learning_rate": 3.314393939393939e-05,
470
+ "loss": 0.3443,
471
+ "step": 710
472
+ },
473
+ {
474
+ "epoch": 4.09,
475
+ "learning_rate": 3.282828282828283e-05,
476
+ "loss": 0.325,
477
+ "step": 720
478
+ },
479
+ {
480
+ "epoch": 4.15,
481
+ "learning_rate": 3.251262626262627e-05,
482
+ "loss": 0.3266,
483
+ "step": 730
484
+ },
485
+ {
486
+ "epoch": 4.2,
487
+ "learning_rate": 3.2196969696969696e-05,
488
+ "loss": 0.3468,
489
+ "step": 740
490
+ },
491
+ {
492
+ "epoch": 4.26,
493
+ "learning_rate": 3.188131313131314e-05,
494
+ "loss": 0.3182,
495
+ "step": 750
496
+ },
497
+ {
498
+ "epoch": 4.32,
499
+ "learning_rate": 3.1565656565656566e-05,
500
+ "loss": 0.2887,
501
+ "step": 760
502
+ },
503
+ {
504
+ "epoch": 4.38,
505
+ "learning_rate": 3.125e-05,
506
+ "loss": 0.3317,
507
+ "step": 770
508
+ },
509
+ {
510
+ "epoch": 4.43,
511
+ "learning_rate": 3.0934343434343436e-05,
512
+ "loss": 0.3089,
513
+ "step": 780
514
+ },
515
+ {
516
+ "epoch": 4.49,
517
+ "learning_rate": 3.061868686868687e-05,
518
+ "loss": 0.3096,
519
+ "step": 790
520
+ },
521
+ {
522
+ "epoch": 4.55,
523
+ "learning_rate": 3.0303030303030306e-05,
524
+ "loss": 0.3073,
525
+ "step": 800
526
+ },
527
+ {
528
+ "epoch": 4.6,
529
+ "learning_rate": 2.9987373737373737e-05,
530
+ "loss": 0.3138,
531
+ "step": 810
532
+ },
533
+ {
534
+ "epoch": 4.66,
535
+ "learning_rate": 2.9671717171717172e-05,
536
+ "loss": 0.3206,
537
+ "step": 820
538
+ },
539
+ {
540
+ "epoch": 4.72,
541
+ "learning_rate": 2.935606060606061e-05,
542
+ "loss": 0.3221,
543
+ "step": 830
544
+ },
545
+ {
546
+ "epoch": 4.77,
547
+ "learning_rate": 2.904040404040404e-05,
548
+ "loss": 0.3274,
549
+ "step": 840
550
+ },
551
+ {
552
+ "epoch": 4.83,
553
+ "learning_rate": 2.8724747474747477e-05,
554
+ "loss": 0.3171,
555
+ "step": 850
556
+ },
557
+ {
558
+ "epoch": 4.89,
559
+ "learning_rate": 2.8409090909090912e-05,
560
+ "loss": 0.3084,
561
+ "step": 860
562
+ },
563
+ {
564
+ "epoch": 4.94,
565
+ "learning_rate": 2.8093434343434344e-05,
566
+ "loss": 0.3076,
567
+ "step": 870
568
+ },
569
+ {
570
+ "epoch": 5.0,
571
+ "learning_rate": 2.777777777777778e-05,
572
+ "loss": 0.3421,
573
+ "step": 880
574
+ },
575
+ {
576
+ "epoch": 5.0,
577
+ "eval_accuracy": 0.972,
578
+ "eval_loss": 0.08578282594680786,
579
+ "eval_runtime": 20.54,
580
+ "eval_samples_per_second": 243.427,
581
+ "eval_steps_per_second": 3.846,
582
+ "step": 880
583
+ },
584
+ {
585
+ "epoch": 5.06,
586
+ "learning_rate": 2.746212121212121e-05,
587
+ "loss": 0.3177,
588
+ "step": 890
589
+ },
590
+ {
591
+ "epoch": 5.11,
592
+ "learning_rate": 2.714646464646465e-05,
593
+ "loss": 0.3161,
594
+ "step": 900
595
+ },
596
+ {
597
+ "epoch": 5.17,
598
+ "learning_rate": 2.6830808080808084e-05,
599
+ "loss": 0.3014,
600
+ "step": 910
601
+ },
602
+ {
603
+ "epoch": 5.23,
604
+ "learning_rate": 2.6515151515151516e-05,
605
+ "loss": 0.3328,
606
+ "step": 920
607
+ },
608
+ {
609
+ "epoch": 5.28,
610
+ "learning_rate": 2.619949494949495e-05,
611
+ "loss": 0.3236,
612
+ "step": 930
613
+ },
614
+ {
615
+ "epoch": 5.34,
616
+ "learning_rate": 2.5883838383838382e-05,
617
+ "loss": 0.3287,
618
+ "step": 940
619
+ },
620
+ {
621
+ "epoch": 5.4,
622
+ "learning_rate": 2.5568181818181817e-05,
623
+ "loss": 0.3083,
624
+ "step": 950
625
+ },
626
+ {
627
+ "epoch": 5.45,
628
+ "learning_rate": 2.5252525252525256e-05,
629
+ "loss": 0.3076,
630
+ "step": 960
631
+ },
632
+ {
633
+ "epoch": 5.51,
634
+ "learning_rate": 2.4936868686868688e-05,
635
+ "loss": 0.31,
636
+ "step": 970
637
+ },
638
+ {
639
+ "epoch": 5.57,
640
+ "learning_rate": 2.4621212121212123e-05,
641
+ "loss": 0.2931,
642
+ "step": 980
643
+ },
644
+ {
645
+ "epoch": 5.62,
646
+ "learning_rate": 2.4305555555555558e-05,
647
+ "loss": 0.3151,
648
+ "step": 990
649
+ },
650
+ {
651
+ "epoch": 5.68,
652
+ "learning_rate": 2.398989898989899e-05,
653
+ "loss": 0.3292,
654
+ "step": 1000
655
+ },
656
+ {
657
+ "epoch": 5.74,
658
+ "learning_rate": 2.3674242424242428e-05,
659
+ "loss": 0.3149,
660
+ "step": 1010
661
+ },
662
+ {
663
+ "epoch": 5.8,
664
+ "learning_rate": 2.335858585858586e-05,
665
+ "loss": 0.3153,
666
+ "step": 1020
667
+ },
668
+ {
669
+ "epoch": 5.85,
670
+ "learning_rate": 2.3042929292929294e-05,
671
+ "loss": 0.3039,
672
+ "step": 1030
673
+ },
674
+ {
675
+ "epoch": 5.91,
676
+ "learning_rate": 2.272727272727273e-05,
677
+ "loss": 0.289,
678
+ "step": 1040
679
+ },
680
+ {
681
+ "epoch": 5.97,
682
+ "learning_rate": 2.2411616161616164e-05,
683
+ "loss": 0.2915,
684
+ "step": 1050
685
+ },
686
+ {
687
+ "epoch": 6.0,
688
+ "eval_accuracy": 0.9724,
689
+ "eval_loss": 0.08235723525285721,
690
+ "eval_runtime": 20.3521,
691
+ "eval_samples_per_second": 245.674,
692
+ "eval_steps_per_second": 3.882,
693
+ "step": 1056
694
+ },
695
+ {
696
+ "epoch": 6.02,
697
+ "learning_rate": 2.20959595959596e-05,
698
+ "loss": 0.3045,
699
+ "step": 1060
700
+ },
701
+ {
702
+ "epoch": 6.08,
703
+ "learning_rate": 2.178030303030303e-05,
704
+ "loss": 0.2974,
705
+ "step": 1070
706
+ },
707
+ {
708
+ "epoch": 6.14,
709
+ "learning_rate": 2.1464646464646466e-05,
710
+ "loss": 0.2713,
711
+ "step": 1080
712
+ },
713
+ {
714
+ "epoch": 6.19,
715
+ "learning_rate": 2.1148989898989898e-05,
716
+ "loss": 0.3115,
717
+ "step": 1090
718
+ },
719
+ {
720
+ "epoch": 6.25,
721
+ "learning_rate": 2.0833333333333336e-05,
722
+ "loss": 0.2762,
723
+ "step": 1100
724
+ },
725
+ {
726
+ "epoch": 6.31,
727
+ "learning_rate": 2.0517676767676768e-05,
728
+ "loss": 0.2833,
729
+ "step": 1110
730
+ },
731
+ {
732
+ "epoch": 6.36,
733
+ "learning_rate": 2.0202020202020203e-05,
734
+ "loss": 0.2867,
735
+ "step": 1120
736
+ },
737
+ {
738
+ "epoch": 6.42,
739
+ "learning_rate": 1.9886363636363638e-05,
740
+ "loss": 0.3164,
741
+ "step": 1130
742
+ },
743
+ {
744
+ "epoch": 6.48,
745
+ "learning_rate": 1.9570707070707073e-05,
746
+ "loss": 0.2845,
747
+ "step": 1140
748
+ },
749
+ {
750
+ "epoch": 6.53,
751
+ "learning_rate": 1.9255050505050508e-05,
752
+ "loss": 0.2943,
753
+ "step": 1150
754
+ },
755
+ {
756
+ "epoch": 6.59,
757
+ "learning_rate": 1.893939393939394e-05,
758
+ "loss": 0.2883,
759
+ "step": 1160
760
+ },
761
+ {
762
+ "epoch": 6.65,
763
+ "learning_rate": 1.8623737373737374e-05,
764
+ "loss": 0.3085,
765
+ "step": 1170
766
+ },
767
+ {
768
+ "epoch": 6.7,
769
+ "learning_rate": 1.830808080808081e-05,
770
+ "loss": 0.2938,
771
+ "step": 1180
772
+ },
773
+ {
774
+ "epoch": 6.76,
775
+ "learning_rate": 1.7992424242424244e-05,
776
+ "loss": 0.2886,
777
+ "step": 1190
778
+ },
779
+ {
780
+ "epoch": 6.82,
781
+ "learning_rate": 1.7676767676767676e-05,
782
+ "loss": 0.2913,
783
+ "step": 1200
784
+ },
785
+ {
786
+ "epoch": 6.88,
787
+ "learning_rate": 1.736111111111111e-05,
788
+ "loss": 0.2845,
789
+ "step": 1210
790
+ },
791
+ {
792
+ "epoch": 6.93,
793
+ "learning_rate": 1.7045454545454546e-05,
794
+ "loss": 0.2913,
795
+ "step": 1220
796
+ },
797
+ {
798
+ "epoch": 6.99,
799
+ "learning_rate": 1.672979797979798e-05,
800
+ "loss": 0.3051,
801
+ "step": 1230
802
+ },
803
+ {
804
+ "epoch": 7.0,
805
+ "eval_accuracy": 0.974,
806
+ "eval_loss": 0.08220645785331726,
807
+ "eval_runtime": 20.6232,
808
+ "eval_samples_per_second": 242.446,
809
+ "eval_steps_per_second": 3.831,
810
+ "step": 1232
811
+ },
812
+ {
813
+ "epoch": 7.05,
814
+ "learning_rate": 1.6414141414141416e-05,
815
+ "loss": 0.2587,
816
+ "step": 1240
817
+ },
818
+ {
819
+ "epoch": 7.1,
820
+ "learning_rate": 1.6098484848484848e-05,
821
+ "loss": 0.2601,
822
+ "step": 1250
823
+ },
824
+ {
825
+ "epoch": 7.16,
826
+ "learning_rate": 1.5782828282828283e-05,
827
+ "loss": 0.274,
828
+ "step": 1260
829
+ },
830
+ {
831
+ "epoch": 7.22,
832
+ "learning_rate": 1.5467171717171718e-05,
833
+ "loss": 0.2968,
834
+ "step": 1270
835
+ },
836
+ {
837
+ "epoch": 7.27,
838
+ "learning_rate": 1.5151515151515153e-05,
839
+ "loss": 0.2883,
840
+ "step": 1280
841
+ },
842
+ {
843
+ "epoch": 7.33,
844
+ "learning_rate": 1.4835858585858586e-05,
845
+ "loss": 0.2637,
846
+ "step": 1290
847
+ },
848
+ {
849
+ "epoch": 7.39,
850
+ "learning_rate": 1.452020202020202e-05,
851
+ "loss": 0.2529,
852
+ "step": 1300
853
+ },
854
+ {
855
+ "epoch": 7.44,
856
+ "learning_rate": 1.4204545454545456e-05,
857
+ "loss": 0.2808,
858
+ "step": 1310
859
+ },
860
+ {
861
+ "epoch": 7.5,
862
+ "learning_rate": 1.388888888888889e-05,
863
+ "loss": 0.269,
864
+ "step": 1320
865
+ },
866
+ {
867
+ "epoch": 7.56,
868
+ "learning_rate": 1.3573232323232325e-05,
869
+ "loss": 0.2782,
870
+ "step": 1330
871
+ },
872
+ {
873
+ "epoch": 7.61,
874
+ "learning_rate": 1.3257575757575758e-05,
875
+ "loss": 0.2754,
876
+ "step": 1340
877
+ },
878
+ {
879
+ "epoch": 7.67,
880
+ "learning_rate": 1.2941919191919191e-05,
881
+ "loss": 0.2678,
882
+ "step": 1350
883
+ },
884
+ {
885
+ "epoch": 7.73,
886
+ "learning_rate": 1.2626262626262628e-05,
887
+ "loss": 0.2855,
888
+ "step": 1360
889
+ },
890
+ {
891
+ "epoch": 7.78,
892
+ "learning_rate": 1.2310606060606061e-05,
893
+ "loss": 0.2705,
894
+ "step": 1370
895
+ },
896
+ {
897
+ "epoch": 7.84,
898
+ "learning_rate": 1.1994949494949495e-05,
899
+ "loss": 0.2789,
900
+ "step": 1380
901
+ },
902
+ {
903
+ "epoch": 7.9,
904
+ "learning_rate": 1.167929292929293e-05,
905
+ "loss": 0.2668,
906
+ "step": 1390
907
+ },
908
+ {
909
+ "epoch": 7.95,
910
+ "learning_rate": 1.1363636363636365e-05,
911
+ "loss": 0.2849,
912
+ "step": 1400
913
+ },
914
+ {
915
+ "epoch": 8.0,
916
+ "eval_accuracy": 0.975,
917
+ "eval_loss": 0.0769709050655365,
918
+ "eval_runtime": 20.7637,
919
+ "eval_samples_per_second": 240.804,
920
+ "eval_steps_per_second": 3.805,
921
+ "step": 1408
922
+ },
923
+ {
924
+ "epoch": 8.01,
925
+ "learning_rate": 1.10479797979798e-05,
926
+ "loss": 0.2466,
927
+ "step": 1410
928
+ },
929
+ {
930
+ "epoch": 8.07,
931
+ "learning_rate": 1.0732323232323233e-05,
932
+ "loss": 0.2666,
933
+ "step": 1420
934
+ },
935
+ {
936
+ "epoch": 8.12,
937
+ "learning_rate": 1.0416666666666668e-05,
938
+ "loss": 0.2764,
939
+ "step": 1430
940
+ },
941
+ {
942
+ "epoch": 8.18,
943
+ "learning_rate": 1.0101010101010101e-05,
944
+ "loss": 0.2841,
945
+ "step": 1440
946
+ },
947
+ {
948
+ "epoch": 8.24,
949
+ "learning_rate": 9.785353535353536e-06,
950
+ "loss": 0.2809,
951
+ "step": 1450
952
+ },
953
+ {
954
+ "epoch": 8.3,
955
+ "learning_rate": 9.46969696969697e-06,
956
+ "loss": 0.2708,
957
+ "step": 1460
958
+ },
959
+ {
960
+ "epoch": 8.35,
961
+ "learning_rate": 9.154040404040405e-06,
962
+ "loss": 0.2986,
963
+ "step": 1470
964
+ },
965
+ {
966
+ "epoch": 8.41,
967
+ "learning_rate": 8.838383838383838e-06,
968
+ "loss": 0.3001,
969
+ "step": 1480
970
+ },
971
+ {
972
+ "epoch": 8.47,
973
+ "learning_rate": 8.522727272727273e-06,
974
+ "loss": 0.2697,
975
+ "step": 1490
976
+ },
977
+ {
978
+ "epoch": 8.52,
979
+ "learning_rate": 8.207070707070708e-06,
980
+ "loss": 0.2524,
981
+ "step": 1500
982
+ },
983
+ {
984
+ "epoch": 8.58,
985
+ "learning_rate": 7.891414141414141e-06,
986
+ "loss": 0.2972,
987
+ "step": 1510
988
+ },
989
+ {
990
+ "epoch": 8.64,
991
+ "learning_rate": 7.5757575757575764e-06,
992
+ "loss": 0.256,
993
+ "step": 1520
994
+ },
995
+ {
996
+ "epoch": 8.69,
997
+ "learning_rate": 7.26010101010101e-06,
998
+ "loss": 0.2703,
999
+ "step": 1530
1000
+ },
1001
+ {
1002
+ "epoch": 8.75,
1003
+ "learning_rate": 6.944444444444445e-06,
1004
+ "loss": 0.2786,
1005
+ "step": 1540
1006
+ },
1007
+ {
1008
+ "epoch": 8.81,
1009
+ "learning_rate": 6.628787878787879e-06,
1010
+ "loss": 0.2536,
1011
+ "step": 1550
1012
+ },
1013
+ {
1014
+ "epoch": 8.86,
1015
+ "learning_rate": 6.313131313131314e-06,
1016
+ "loss": 0.2516,
1017
+ "step": 1560
1018
+ },
1019
+ {
1020
+ "epoch": 8.92,
1021
+ "learning_rate": 5.997474747474747e-06,
1022
+ "loss": 0.2649,
1023
+ "step": 1570
1024
+ },
1025
+ {
1026
+ "epoch": 8.98,
1027
+ "learning_rate": 5.681818181818182e-06,
1028
+ "loss": 0.2661,
1029
+ "step": 1580
1030
+ },
1031
+ {
1032
+ "epoch": 9.0,
1033
+ "eval_accuracy": 0.9756,
1034
+ "eval_loss": 0.07725635915994644,
1035
+ "eval_runtime": 20.3739,
1036
+ "eval_samples_per_second": 245.412,
1037
+ "eval_steps_per_second": 3.878,
1038
+ "step": 1584
1039
+ },
1040
+ {
1041
+ "epoch": 9.03,
1042
+ "learning_rate": 5.3661616161616165e-06,
1043
+ "loss": 0.2733,
1044
+ "step": 1590
1045
+ },
1046
+ {
1047
+ "epoch": 9.09,
1048
+ "learning_rate": 5.050505050505051e-06,
1049
+ "loss": 0.2386,
1050
+ "step": 1600
1051
+ },
1052
+ {
1053
+ "epoch": 9.15,
1054
+ "learning_rate": 4.734848484848485e-06,
1055
+ "loss": 0.2804,
1056
+ "step": 1610
1057
+ },
1058
+ {
1059
+ "epoch": 9.2,
1060
+ "learning_rate": 4.419191919191919e-06,
1061
+ "loss": 0.2626,
1062
+ "step": 1620
1063
+ },
1064
+ {
1065
+ "epoch": 9.26,
1066
+ "learning_rate": 4.103535353535354e-06,
1067
+ "loss": 0.2528,
1068
+ "step": 1630
1069
+ },
1070
+ {
1071
+ "epoch": 9.32,
1072
+ "learning_rate": 3.7878787878787882e-06,
1073
+ "loss": 0.2668,
1074
+ "step": 1640
1075
+ },
1076
+ {
1077
+ "epoch": 9.38,
1078
+ "learning_rate": 3.4722222222222224e-06,
1079
+ "loss": 0.2393,
1080
+ "step": 1650
1081
+ },
1082
+ {
1083
+ "epoch": 9.43,
1084
+ "learning_rate": 3.156565656565657e-06,
1085
+ "loss": 0.2549,
1086
+ "step": 1660
1087
+ },
1088
+ {
1089
+ "epoch": 9.49,
1090
+ "learning_rate": 2.840909090909091e-06,
1091
+ "loss": 0.2457,
1092
+ "step": 1670
1093
+ },
1094
+ {
1095
+ "epoch": 9.55,
1096
+ "learning_rate": 2.5252525252525253e-06,
1097
+ "loss": 0.2752,
1098
+ "step": 1680
1099
+ },
1100
+ {
1101
+ "epoch": 9.6,
1102
+ "learning_rate": 2.2095959595959595e-06,
1103
+ "loss": 0.2419,
1104
+ "step": 1690
1105
+ },
1106
+ {
1107
+ "epoch": 9.66,
1108
+ "learning_rate": 1.8939393939393941e-06,
1109
+ "loss": 0.2609,
1110
+ "step": 1700
1111
+ },
1112
+ {
1113
+ "epoch": 9.72,
1114
+ "learning_rate": 1.5782828282828285e-06,
1115
+ "loss": 0.2548,
1116
+ "step": 1710
1117
+ },
1118
+ {
1119
+ "epoch": 9.77,
1120
+ "learning_rate": 1.2626262626262627e-06,
1121
+ "loss": 0.284,
1122
+ "step": 1720
1123
+ },
1124
+ {
1125
+ "epoch": 9.83,
1126
+ "learning_rate": 9.469696969696971e-07,
1127
+ "loss": 0.2615,
1128
+ "step": 1730
1129
+ },
1130
+ {
1131
+ "epoch": 9.89,
1132
+ "learning_rate": 6.313131313131313e-07,
1133
+ "loss": 0.2843,
1134
+ "step": 1740
1135
+ },
1136
+ {
1137
+ "epoch": 9.94,
1138
+ "learning_rate": 3.1565656565656567e-07,
1139
+ "loss": 0.2549,
1140
+ "step": 1750
1141
+ },
1142
+ {
1143
+ "epoch": 10.0,
1144
+ "learning_rate": 0.0,
1145
+ "loss": 0.2504,
1146
+ "step": 1760
1147
+ },
1148
+ {
1149
+ "epoch": 10.0,
1150
+ "eval_accuracy": 0.9758,
1151
+ "eval_loss": 0.07598920166492462,
1152
+ "eval_runtime": 20.5812,
1153
+ "eval_samples_per_second": 242.94,
1154
+ "eval_steps_per_second": 3.838,
1155
+ "step": 1760
1156
+ },
1157
+ {
1158
+ "epoch": 10.0,
1159
+ "step": 1760,
1160
+ "total_flos": 1.11876953739264e+19,
1161
+ "train_loss": 0.3102077527479692,
1162
+ "train_runtime": 2497.672,
1163
+ "train_samples_per_second": 180.168,
1164
+ "train_steps_per_second": 0.705
1165
+ }
1166
+ ],
1167
+ "logging_steps": 10,
1168
+ "max_steps": 1760,
1169
+ "num_input_tokens_seen": 0,
1170
+ "num_train_epochs": 10,
1171
+ "save_steps": 500,
1172
+ "total_flos": 1.11876953739264e+19,
1173
+ "train_batch_size": 64,
1174
+ "trial_name": null,
1175
+ "trial_params": null
1176
+ }