emrecan commited on
Commit
1f0ff20
1 Parent(s): 91de5c1

End of training

Browse files
Files changed (3) hide show
  1. all_results.json +7 -0
  2. train_results.json +7 -0
  3. trainer_state.json +1345 -0
all_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.3844692539477694,
4
+ "train_runtime": 45994.9721,
5
+ "train_samples_per_second": 61.446,
6
+ "train_steps_per_second": 1.92
7
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.3844692539477694,
4
+ "train_runtime": 45994.9721,
5
+ "train_samples_per_second": 61.446,
6
+ "train_steps_per_second": 1.92
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1345 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8122261844116149,
3
+ "best_model_checkpoint": "/home/user/emrecan/models/dbmdz_convbert-base-turkish-mc4-cased_allnli_tr/checkpoint-80000",
4
+ "epoch": 3.0,
5
+ "global_step": 88320,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.03,
12
+ "learning_rate": 1.9773550724637682e-05,
13
+ "loss": 0.7338,
14
+ "step": 1000
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "eval_accuracy": 0.723586347427407,
19
+ "eval_loss": 0.6721901297569275,
20
+ "eval_runtime": 44.7079,
21
+ "eval_samples_per_second": 219.536,
22
+ "eval_steps_per_second": 6.867,
23
+ "step": 1000
24
+ },
25
+ {
26
+ "epoch": 0.07,
27
+ "learning_rate": 1.9547101449275363e-05,
28
+ "loss": 0.603,
29
+ "step": 2000
30
+ },
31
+ {
32
+ "epoch": 0.07,
33
+ "eval_accuracy": 0.7398879266428935,
34
+ "eval_loss": 0.6465004682540894,
35
+ "eval_runtime": 44.7809,
36
+ "eval_samples_per_second": 219.178,
37
+ "eval_steps_per_second": 6.856,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.1,
42
+ "learning_rate": 1.9320652173913047e-05,
43
+ "loss": 0.5605,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.1,
48
+ "eval_accuracy": 0.7727967396841569,
49
+ "eval_loss": 0.5801343321800232,
50
+ "eval_runtime": 44.4418,
51
+ "eval_samples_per_second": 220.851,
52
+ "eval_steps_per_second": 6.908,
53
+ "step": 3000
54
+ },
55
+ {
56
+ "epoch": 0.14,
57
+ "learning_rate": 1.9094202898550727e-05,
58
+ "loss": 0.55,
59
+ "step": 4000
60
+ },
61
+ {
62
+ "epoch": 0.14,
63
+ "eval_accuracy": 0.7626082526744778,
64
+ "eval_loss": 0.5994369983673096,
65
+ "eval_runtime": 44.603,
66
+ "eval_samples_per_second": 220.052,
67
+ "eval_steps_per_second": 6.883,
68
+ "step": 4000
69
+ },
70
+ {
71
+ "epoch": 0.17,
72
+ "learning_rate": 1.8867753623188408e-05,
73
+ "loss": 0.529,
74
+ "step": 5000
75
+ },
76
+ {
77
+ "epoch": 0.17,
78
+ "eval_accuracy": 0.7697401935812532,
79
+ "eval_loss": 0.5720144510269165,
80
+ "eval_runtime": 44.7705,
81
+ "eval_samples_per_second": 219.229,
82
+ "eval_steps_per_second": 6.857,
83
+ "step": 5000
84
+ },
85
+ {
86
+ "epoch": 0.2,
87
+ "learning_rate": 1.864130434782609e-05,
88
+ "loss": 0.5196,
89
+ "step": 6000
90
+ },
91
+ {
92
+ "epoch": 0.2,
93
+ "eval_accuracy": 0.7768721344880285,
94
+ "eval_loss": 0.5691649317741394,
95
+ "eval_runtime": 44.7386,
96
+ "eval_samples_per_second": 219.385,
97
+ "eval_steps_per_second": 6.862,
98
+ "step": 6000
99
+ },
100
+ {
101
+ "epoch": 0.24,
102
+ "learning_rate": 1.841485507246377e-05,
103
+ "loss": 0.5117,
104
+ "step": 7000
105
+ },
106
+ {
107
+ "epoch": 0.24,
108
+ "eval_accuracy": 0.7785022924095771,
109
+ "eval_loss": 0.5724737644195557,
110
+ "eval_runtime": 44.7709,
111
+ "eval_samples_per_second": 219.227,
112
+ "eval_steps_per_second": 6.857,
113
+ "step": 7000
114
+ },
115
+ {
116
+ "epoch": 0.27,
117
+ "learning_rate": 1.818840579710145e-05,
118
+ "loss": 0.5044,
119
+ "step": 8000
120
+ },
121
+ {
122
+ "epoch": 0.27,
123
+ "eval_accuracy": 0.7787060621497708,
124
+ "eval_loss": 0.5532070398330688,
125
+ "eval_runtime": 44.6683,
126
+ "eval_samples_per_second": 219.731,
127
+ "eval_steps_per_second": 6.873,
128
+ "step": 8000
129
+ },
130
+ {
131
+ "epoch": 0.31,
132
+ "learning_rate": 1.7961956521739134e-05,
133
+ "loss": 0.5016,
134
+ "step": 9000
135
+ },
136
+ {
137
+ "epoch": 0.31,
138
+ "eval_accuracy": 0.7811512990320937,
139
+ "eval_loss": 0.5546181797981262,
140
+ "eval_runtime": 44.6036,
141
+ "eval_samples_per_second": 220.049,
142
+ "eval_steps_per_second": 6.883,
143
+ "step": 9000
144
+ },
145
+ {
146
+ "epoch": 0.34,
147
+ "learning_rate": 1.7735507246376815e-05,
148
+ "loss": 0.5031,
149
+ "step": 10000
150
+ },
151
+ {
152
+ "epoch": 0.34,
153
+ "eval_accuracy": 0.7869587366276108,
154
+ "eval_loss": 0.5461075901985168,
155
+ "eval_runtime": 44.6087,
156
+ "eval_samples_per_second": 220.024,
157
+ "eval_steps_per_second": 6.882,
158
+ "step": 10000
159
+ },
160
+ {
161
+ "epoch": 0.37,
162
+ "learning_rate": 1.7509057971014495e-05,
163
+ "loss": 0.4949,
164
+ "step": 11000
165
+ },
166
+ {
167
+ "epoch": 0.37,
168
+ "eval_accuracy": 0.7825776872134488,
169
+ "eval_loss": 0.5724930763244629,
170
+ "eval_runtime": 44.6365,
171
+ "eval_samples_per_second": 219.888,
172
+ "eval_steps_per_second": 6.878,
173
+ "step": 11000
174
+ },
175
+ {
176
+ "epoch": 0.41,
177
+ "learning_rate": 1.7282608695652176e-05,
178
+ "loss": 0.4894,
179
+ "step": 12000
180
+ },
181
+ {
182
+ "epoch": 0.41,
183
+ "eval_accuracy": 0.7932755985736119,
184
+ "eval_loss": 0.5418519377708435,
185
+ "eval_runtime": 44.7444,
186
+ "eval_samples_per_second": 219.357,
187
+ "eval_steps_per_second": 6.861,
188
+ "step": 12000
189
+ },
190
+ {
191
+ "epoch": 0.44,
192
+ "learning_rate": 1.7056159420289856e-05,
193
+ "loss": 0.4796,
194
+ "step": 13000
195
+ },
196
+ {
197
+ "epoch": 0.44,
198
+ "eval_accuracy": 0.7914416709118696,
199
+ "eval_loss": 0.5278186798095703,
200
+ "eval_runtime": 44.5917,
201
+ "eval_samples_per_second": 220.108,
202
+ "eval_steps_per_second": 6.885,
203
+ "step": 13000
204
+ },
205
+ {
206
+ "epoch": 0.48,
207
+ "learning_rate": 1.6829710144927537e-05,
208
+ "loss": 0.4795,
209
+ "step": 14000
210
+ },
211
+ {
212
+ "epoch": 0.48,
213
+ "eval_accuracy": 0.7953132959755477,
214
+ "eval_loss": 0.5193336606025696,
215
+ "eval_runtime": 44.673,
216
+ "eval_samples_per_second": 219.708,
217
+ "eval_steps_per_second": 6.872,
218
+ "step": 14000
219
+ },
220
+ {
221
+ "epoch": 0.51,
222
+ "learning_rate": 1.6603260869565218e-05,
223
+ "loss": 0.4713,
224
+ "step": 15000
225
+ },
226
+ {
227
+ "epoch": 0.51,
228
+ "eval_accuracy": 0.7770759042282221,
229
+ "eval_loss": 0.5534396171569824,
230
+ "eval_runtime": 44.688,
231
+ "eval_samples_per_second": 219.634,
232
+ "eval_steps_per_second": 6.87,
233
+ "step": 15000
234
+ },
235
+ {
236
+ "epoch": 0.54,
237
+ "learning_rate": 1.6376811594202898e-05,
238
+ "loss": 0.4738,
239
+ "step": 16000
240
+ },
241
+ {
242
+ "epoch": 0.54,
243
+ "eval_accuracy": 0.803871625063678,
244
+ "eval_loss": 0.5097627639770508,
245
+ "eval_runtime": 44.7421,
246
+ "eval_samples_per_second": 219.368,
247
+ "eval_steps_per_second": 6.862,
248
+ "step": 16000
249
+ },
250
+ {
251
+ "epoch": 0.58,
252
+ "learning_rate": 1.615036231884058e-05,
253
+ "loss": 0.481,
254
+ "step": 17000
255
+ },
256
+ {
257
+ "epoch": 0.58,
258
+ "eval_accuracy": 0.7958227203260316,
259
+ "eval_loss": 0.5244179368019104,
260
+ "eval_runtime": 44.7909,
261
+ "eval_samples_per_second": 219.129,
262
+ "eval_steps_per_second": 6.854,
263
+ "step": 17000
264
+ },
265
+ {
266
+ "epoch": 0.61,
267
+ "learning_rate": 1.5923913043478263e-05,
268
+ "loss": 0.4634,
269
+ "step": 18000
270
+ },
271
+ {
272
+ "epoch": 0.61,
273
+ "eval_accuracy": 0.7972491085073866,
274
+ "eval_loss": 0.5214974880218506,
275
+ "eval_runtime": 44.5586,
276
+ "eval_samples_per_second": 220.272,
277
+ "eval_steps_per_second": 6.89,
278
+ "step": 18000
279
+ },
280
+ {
281
+ "epoch": 0.65,
282
+ "learning_rate": 1.5697463768115943e-05,
283
+ "loss": 0.465,
284
+ "step": 19000
285
+ },
286
+ {
287
+ "epoch": 0.65,
288
+ "eval_accuracy": 0.7984717269485482,
289
+ "eval_loss": 0.5128703117370605,
290
+ "eval_runtime": 44.7294,
291
+ "eval_samples_per_second": 219.431,
292
+ "eval_steps_per_second": 6.863,
293
+ "step": 19000
294
+ },
295
+ {
296
+ "epoch": 0.68,
297
+ "learning_rate": 1.5471014492753624e-05,
298
+ "loss": 0.4624,
299
+ "step": 20000
300
+ },
301
+ {
302
+ "epoch": 0.68,
303
+ "eval_accuracy": 0.8046867040244524,
304
+ "eval_loss": 0.5061507821083069,
305
+ "eval_runtime": 44.7949,
306
+ "eval_samples_per_second": 219.11,
307
+ "eval_steps_per_second": 6.853,
308
+ "step": 20000
309
+ },
310
+ {
311
+ "epoch": 0.71,
312
+ "learning_rate": 1.5244565217391305e-05,
313
+ "loss": 0.4597,
314
+ "step": 21000
315
+ },
316
+ {
317
+ "epoch": 0.71,
318
+ "eval_accuracy": 0.8028527763627101,
319
+ "eval_loss": 0.5114405751228333,
320
+ "eval_runtime": 47.1654,
321
+ "eval_samples_per_second": 208.098,
322
+ "eval_steps_per_second": 6.509,
323
+ "step": 21000
324
+ },
325
+ {
326
+ "epoch": 0.75,
327
+ "learning_rate": 1.5018115942028985e-05,
328
+ "loss": 0.4571,
329
+ "step": 22000
330
+ },
331
+ {
332
+ "epoch": 0.75,
333
+ "eval_accuracy": 0.807335710646969,
334
+ "eval_loss": 0.5069620609283447,
335
+ "eval_runtime": 45.7859,
336
+ "eval_samples_per_second": 214.368,
337
+ "eval_steps_per_second": 6.705,
338
+ "step": 22000
339
+ },
340
+ {
341
+ "epoch": 0.78,
342
+ "learning_rate": 1.479166666666667e-05,
343
+ "loss": 0.4602,
344
+ "step": 23000
345
+ },
346
+ {
347
+ "epoch": 0.78,
348
+ "eval_accuracy": 0.7992868059093224,
349
+ "eval_loss": 0.5114511847496033,
350
+ "eval_runtime": 45.8642,
351
+ "eval_samples_per_second": 214.001,
352
+ "eval_steps_per_second": 6.694,
353
+ "step": 23000
354
+ },
355
+ {
356
+ "epoch": 0.82,
357
+ "learning_rate": 1.456521739130435e-05,
358
+ "loss": 0.4552,
359
+ "step": 24000
360
+ },
361
+ {
362
+ "epoch": 0.82,
363
+ "eval_accuracy": 0.8051961283749364,
364
+ "eval_loss": 0.5085225105285645,
365
+ "eval_runtime": 46.055,
366
+ "eval_samples_per_second": 213.115,
367
+ "eval_steps_per_second": 6.666,
368
+ "step": 24000
369
+ },
370
+ {
371
+ "epoch": 0.85,
372
+ "learning_rate": 1.433876811594203e-05,
373
+ "loss": 0.4538,
374
+ "step": 25000
375
+ },
376
+ {
377
+ "epoch": 0.85,
378
+ "eval_accuracy": 0.7973509933774835,
379
+ "eval_loss": 0.5118097066879272,
380
+ "eval_runtime": 45.9419,
381
+ "eval_samples_per_second": 213.639,
382
+ "eval_steps_per_second": 6.682,
383
+ "step": 25000
384
+ },
385
+ {
386
+ "epoch": 0.88,
387
+ "learning_rate": 1.4112318840579711e-05,
388
+ "loss": 0.4517,
389
+ "step": 26000
390
+ },
391
+ {
392
+ "epoch": 0.88,
393
+ "eval_accuracy": 0.804381049414162,
394
+ "eval_loss": 0.5036212205886841,
395
+ "eval_runtime": 45.8528,
396
+ "eval_samples_per_second": 214.055,
397
+ "eval_steps_per_second": 6.695,
398
+ "step": 26000
399
+ },
400
+ {
401
+ "epoch": 0.92,
402
+ "learning_rate": 1.3885869565217392e-05,
403
+ "loss": 0.4517,
404
+ "step": 27000
405
+ },
406
+ {
407
+ "epoch": 0.92,
408
+ "eval_accuracy": 0.8062149770759043,
409
+ "eval_loss": 0.4929669201374054,
410
+ "eval_runtime": 45.8655,
411
+ "eval_samples_per_second": 213.995,
412
+ "eval_steps_per_second": 6.693,
413
+ "step": 27000
414
+ },
415
+ {
416
+ "epoch": 0.95,
417
+ "learning_rate": 1.3659420289855074e-05,
418
+ "loss": 0.4413,
419
+ "step": 28000
420
+ },
421
+ {
422
+ "epoch": 0.95,
423
+ "eval_accuracy": 0.7964340295466124,
424
+ "eval_loss": 0.5306665897369385,
425
+ "eval_runtime": 45.9438,
426
+ "eval_samples_per_second": 213.631,
427
+ "eval_steps_per_second": 6.682,
428
+ "step": 28000
429
+ },
430
+ {
431
+ "epoch": 0.99,
432
+ "learning_rate": 1.3432971014492755e-05,
433
+ "loss": 0.4483,
434
+ "step": 29000
435
+ },
436
+ {
437
+ "epoch": 0.99,
438
+ "eval_accuracy": 0.7937850229240958,
439
+ "eval_loss": 0.5194875001907349,
440
+ "eval_runtime": 45.8704,
441
+ "eval_samples_per_second": 213.972,
442
+ "eval_steps_per_second": 6.693,
443
+ "step": 29000
444
+ },
445
+ {
446
+ "epoch": 1.02,
447
+ "learning_rate": 1.3206521739130435e-05,
448
+ "loss": 0.4036,
449
+ "step": 30000
450
+ },
451
+ {
452
+ "epoch": 1.02,
453
+ "eval_accuracy": 0.8028527763627101,
454
+ "eval_loss": 0.5238322615623474,
455
+ "eval_runtime": 45.9862,
456
+ "eval_samples_per_second": 213.434,
457
+ "eval_steps_per_second": 6.676,
458
+ "step": 30000
459
+ },
460
+ {
461
+ "epoch": 1.05,
462
+ "learning_rate": 1.2980072463768116e-05,
463
+ "loss": 0.3724,
464
+ "step": 31000
465
+ },
466
+ {
467
+ "epoch": 1.05,
468
+ "eval_accuracy": 0.8081507896077432,
469
+ "eval_loss": 0.5124598741531372,
470
+ "eval_runtime": 47.042,
471
+ "eval_samples_per_second": 208.643,
472
+ "eval_steps_per_second": 6.526,
473
+ "step": 31000
474
+ },
475
+ {
476
+ "epoch": 1.09,
477
+ "learning_rate": 1.2753623188405797e-05,
478
+ "loss": 0.3777,
479
+ "step": 32000
480
+ },
481
+ {
482
+ "epoch": 1.09,
483
+ "eval_accuracy": 0.8075394803871625,
484
+ "eval_loss": 0.509867787361145,
485
+ "eval_runtime": 46.0494,
486
+ "eval_samples_per_second": 213.141,
487
+ "eval_steps_per_second": 6.667,
488
+ "step": 32000
489
+ },
490
+ {
491
+ "epoch": 1.12,
492
+ "learning_rate": 1.252717391304348e-05,
493
+ "loss": 0.3753,
494
+ "step": 33000
495
+ },
496
+ {
497
+ "epoch": 1.12,
498
+ "eval_accuracy": 0.8052980132450331,
499
+ "eval_loss": 0.5171676278114319,
500
+ "eval_runtime": 45.8652,
501
+ "eval_samples_per_second": 213.997,
502
+ "eval_steps_per_second": 6.694,
503
+ "step": 33000
504
+ },
505
+ {
506
+ "epoch": 1.15,
507
+ "learning_rate": 1.2300724637681161e-05,
508
+ "loss": 0.367,
509
+ "step": 34000
510
+ },
511
+ {
512
+ "epoch": 1.15,
513
+ "eval_accuracy": 0.8052980132450331,
514
+ "eval_loss": 0.5187755823135376,
515
+ "eval_runtime": 45.7807,
516
+ "eval_samples_per_second": 214.391,
517
+ "eval_steps_per_second": 6.706,
518
+ "step": 34000
519
+ },
520
+ {
521
+ "epoch": 1.19,
522
+ "learning_rate": 1.2074275362318842e-05,
523
+ "loss": 0.3819,
524
+ "step": 35000
525
+ },
526
+ {
527
+ "epoch": 1.19,
528
+ "eval_accuracy": 0.8045848191543555,
529
+ "eval_loss": 0.5218426585197449,
530
+ "eval_runtime": 46.9682,
531
+ "eval_samples_per_second": 208.971,
532
+ "eval_steps_per_second": 6.536,
533
+ "step": 35000
534
+ },
535
+ {
536
+ "epoch": 1.22,
537
+ "learning_rate": 1.1847826086956522e-05,
538
+ "loss": 0.363,
539
+ "step": 36000
540
+ },
541
+ {
542
+ "epoch": 1.22,
543
+ "eval_accuracy": 0.7992868059093224,
544
+ "eval_loss": 0.5201919078826904,
545
+ "eval_runtime": 46.028,
546
+ "eval_samples_per_second": 213.24,
547
+ "eval_steps_per_second": 6.67,
548
+ "step": 36000
549
+ },
550
+ {
551
+ "epoch": 1.26,
552
+ "learning_rate": 1.1621376811594205e-05,
553
+ "loss": 0.3794,
554
+ "step": 37000
555
+ },
556
+ {
557
+ "epoch": 1.26,
558
+ "eval_accuracy": 0.8047885888945492,
559
+ "eval_loss": 0.5240486860275269,
560
+ "eval_runtime": 46.0171,
561
+ "eval_samples_per_second": 213.29,
562
+ "eval_steps_per_second": 6.671,
563
+ "step": 37000
564
+ },
565
+ {
566
+ "epoch": 1.29,
567
+ "learning_rate": 1.1394927536231885e-05,
568
+ "loss": 0.3749,
569
+ "step": 38000
570
+ },
571
+ {
572
+ "epoch": 1.29,
573
+ "eval_accuracy": 0.8053998981151299,
574
+ "eval_loss": 0.5025837421417236,
575
+ "eval_runtime": 45.9936,
576
+ "eval_samples_per_second": 213.399,
577
+ "eval_steps_per_second": 6.675,
578
+ "step": 38000
579
+ },
580
+ {
581
+ "epoch": 1.32,
582
+ "learning_rate": 1.1168478260869566e-05,
583
+ "loss": 0.367,
584
+ "step": 39000
585
+ },
586
+ {
587
+ "epoch": 1.32,
588
+ "eval_accuracy": 0.8075394803871625,
589
+ "eval_loss": 0.5197691917419434,
590
+ "eval_runtime": 46.0223,
591
+ "eval_samples_per_second": 213.266,
592
+ "eval_steps_per_second": 6.671,
593
+ "step": 39000
594
+ },
595
+ {
596
+ "epoch": 1.36,
597
+ "learning_rate": 1.0942028985507247e-05,
598
+ "loss": 0.3759,
599
+ "step": 40000
600
+ },
601
+ {
602
+ "epoch": 1.36,
603
+ "eval_accuracy": 0.7992868059093224,
604
+ "eval_loss": 0.5298164486885071,
605
+ "eval_runtime": 46.0659,
606
+ "eval_samples_per_second": 213.064,
607
+ "eval_steps_per_second": 6.664,
608
+ "step": 40000
609
+ },
610
+ {
611
+ "epoch": 1.39,
612
+ "learning_rate": 1.0715579710144927e-05,
613
+ "loss": 0.3701,
614
+ "step": 41000
615
+ },
616
+ {
617
+ "epoch": 1.39,
618
+ "eval_accuracy": 0.8090677534386144,
619
+ "eval_loss": 0.5072407722473145,
620
+ "eval_runtime": 47.2083,
621
+ "eval_samples_per_second": 207.908,
622
+ "eval_steps_per_second": 6.503,
623
+ "step": 41000
624
+ },
625
+ {
626
+ "epoch": 1.43,
627
+ "learning_rate": 1.0489130434782611e-05,
628
+ "loss": 0.3742,
629
+ "step": 42000
630
+ },
631
+ {
632
+ "epoch": 1.43,
633
+ "eval_accuracy": 0.8097809475292919,
634
+ "eval_loss": 0.507131814956665,
635
+ "eval_runtime": 46.0803,
636
+ "eval_samples_per_second": 212.998,
637
+ "eval_steps_per_second": 6.662,
638
+ "step": 42000
639
+ },
640
+ {
641
+ "epoch": 1.46,
642
+ "learning_rate": 1.0262681159420292e-05,
643
+ "loss": 0.3706,
644
+ "step": 43000
645
+ },
646
+ {
647
+ "epoch": 1.46,
648
+ "eval_accuracy": 0.8036678553234845,
649
+ "eval_loss": 0.5317161083221436,
650
+ "eval_runtime": 46.1098,
651
+ "eval_samples_per_second": 212.861,
652
+ "eval_steps_per_second": 6.658,
653
+ "step": 43000
654
+ },
655
+ {
656
+ "epoch": 1.49,
657
+ "learning_rate": 1.0036231884057972e-05,
658
+ "loss": 0.3716,
659
+ "step": 44000
660
+ },
661
+ {
662
+ "epoch": 1.49,
663
+ "eval_accuracy": 0.8051961283749364,
664
+ "eval_loss": 0.5034372210502625,
665
+ "eval_runtime": 46.0657,
666
+ "eval_samples_per_second": 213.065,
667
+ "eval_steps_per_second": 6.664,
668
+ "step": 44000
669
+ },
670
+ {
671
+ "epoch": 1.53,
672
+ "learning_rate": 9.809782608695653e-06,
673
+ "loss": 0.3717,
674
+ "step": 45000
675
+ },
676
+ {
677
+ "epoch": 1.53,
678
+ "eval_accuracy": 0.8012226184411615,
679
+ "eval_loss": 0.525775671005249,
680
+ "eval_runtime": 45.917,
681
+ "eval_samples_per_second": 213.755,
682
+ "eval_steps_per_second": 6.686,
683
+ "step": 45000
684
+ },
685
+ {
686
+ "epoch": 1.56,
687
+ "learning_rate": 9.583333333333335e-06,
688
+ "loss": 0.3714,
689
+ "step": 46000
690
+ },
691
+ {
692
+ "epoch": 1.56,
693
+ "eval_accuracy": 0.8049923586347427,
694
+ "eval_loss": 0.5195213556289673,
695
+ "eval_runtime": 46.0299,
696
+ "eval_samples_per_second": 213.231,
697
+ "eval_steps_per_second": 6.67,
698
+ "step": 46000
699
+ },
700
+ {
701
+ "epoch": 1.6,
702
+ "learning_rate": 9.356884057971016e-06,
703
+ "loss": 0.3781,
704
+ "step": 47000
705
+ },
706
+ {
707
+ "epoch": 1.6,
708
+ "eval_accuracy": 0.8103922567498727,
709
+ "eval_loss": 0.5003909468650818,
710
+ "eval_runtime": 46.0896,
711
+ "eval_samples_per_second": 212.955,
712
+ "eval_steps_per_second": 6.661,
713
+ "step": 47000
714
+ },
715
+ {
716
+ "epoch": 1.63,
717
+ "learning_rate": 9.130434782608697e-06,
718
+ "loss": 0.3725,
719
+ "step": 48000
720
+ },
721
+ {
722
+ "epoch": 1.63,
723
+ "eval_accuracy": 0.8113092205807437,
724
+ "eval_loss": 0.5124202966690063,
725
+ "eval_runtime": 46.0326,
726
+ "eval_samples_per_second": 213.218,
727
+ "eval_steps_per_second": 6.669,
728
+ "step": 48000
729
+ },
730
+ {
731
+ "epoch": 1.66,
732
+ "learning_rate": 8.903985507246377e-06,
733
+ "loss": 0.3624,
734
+ "step": 49000
735
+ },
736
+ {
737
+ "epoch": 1.66,
738
+ "eval_accuracy": 0.8093734080489048,
739
+ "eval_loss": 0.5039945840835571,
740
+ "eval_runtime": 46.0265,
741
+ "eval_samples_per_second": 213.247,
742
+ "eval_steps_per_second": 6.67,
743
+ "step": 49000
744
+ },
745
+ {
746
+ "epoch": 1.7,
747
+ "learning_rate": 8.677536231884058e-06,
748
+ "loss": 0.3657,
749
+ "step": 50000
750
+ },
751
+ {
752
+ "epoch": 1.7,
753
+ "eval_accuracy": 0.8111054508405502,
754
+ "eval_loss": 0.49794963002204895,
755
+ "eval_runtime": 46.0671,
756
+ "eval_samples_per_second": 213.059,
757
+ "eval_steps_per_second": 6.664,
758
+ "step": 50000
759
+ },
760
+ {
761
+ "epoch": 1.73,
762
+ "learning_rate": 8.45108695652174e-06,
763
+ "loss": 0.3669,
764
+ "step": 51000
765
+ },
766
+ {
767
+ "epoch": 1.73,
768
+ "eval_accuracy": 0.8099847172694855,
769
+ "eval_loss": 0.4967914819717407,
770
+ "eval_runtime": 47.5013,
771
+ "eval_samples_per_second": 206.626,
772
+ "eval_steps_per_second": 6.463,
773
+ "step": 51000
774
+ },
775
+ {
776
+ "epoch": 1.77,
777
+ "learning_rate": 8.22463768115942e-06,
778
+ "loss": 0.3636,
779
+ "step": 52000
780
+ },
781
+ {
782
+ "epoch": 1.77,
783
+ "eval_accuracy": 0.8079470198675497,
784
+ "eval_loss": 0.5075203776359558,
785
+ "eval_runtime": 45.998,
786
+ "eval_samples_per_second": 213.379,
787
+ "eval_steps_per_second": 6.674,
788
+ "step": 52000
789
+ },
790
+ {
791
+ "epoch": 1.8,
792
+ "learning_rate": 7.998188405797103e-06,
793
+ "loss": 0.36,
794
+ "step": 53000
795
+ },
796
+ {
797
+ "epoch": 1.8,
798
+ "eval_accuracy": 0.8110035659704534,
799
+ "eval_loss": 0.498517245054245,
800
+ "eval_runtime": 46.2214,
801
+ "eval_samples_per_second": 212.347,
802
+ "eval_steps_per_second": 6.642,
803
+ "step": 53000
804
+ },
805
+ {
806
+ "epoch": 1.83,
807
+ "learning_rate": 7.771739130434784e-06,
808
+ "loss": 0.3624,
809
+ "step": 54000
810
+ },
811
+ {
812
+ "epoch": 1.83,
813
+ "eval_accuracy": 0.8070300560366785,
814
+ "eval_loss": 0.5125023722648621,
815
+ "eval_runtime": 46.2947,
816
+ "eval_samples_per_second": 212.011,
817
+ "eval_steps_per_second": 6.631,
818
+ "step": 54000
819
+ },
820
+ {
821
+ "epoch": 1.87,
822
+ "learning_rate": 7.545289855072464e-06,
823
+ "loss": 0.366,
824
+ "step": 55000
825
+ },
826
+ {
827
+ "epoch": 1.87,
828
+ "eval_accuracy": 0.8117167600611309,
829
+ "eval_loss": 0.4918198585510254,
830
+ "eval_runtime": 46.0126,
831
+ "eval_samples_per_second": 213.311,
832
+ "eval_steps_per_second": 6.672,
833
+ "step": 55000
834
+ },
835
+ {
836
+ "epoch": 1.9,
837
+ "learning_rate": 7.318840579710146e-06,
838
+ "loss": 0.3655,
839
+ "step": 56000
840
+ },
841
+ {
842
+ "epoch": 1.9,
843
+ "eval_accuracy": 0.8109016811003565,
844
+ "eval_loss": 0.505085289478302,
845
+ "eval_runtime": 46.0789,
846
+ "eval_samples_per_second": 213.004,
847
+ "eval_steps_per_second": 6.662,
848
+ "step": 56000
849
+ },
850
+ {
851
+ "epoch": 1.94,
852
+ "learning_rate": 7.092391304347826e-06,
853
+ "loss": 0.3609,
854
+ "step": 57000
855
+ },
856
+ {
857
+ "epoch": 1.94,
858
+ "eval_accuracy": 0.8104941416199695,
859
+ "eval_loss": 0.5082952976226807,
860
+ "eval_runtime": 46.1283,
861
+ "eval_samples_per_second": 212.776,
862
+ "eval_steps_per_second": 6.655,
863
+ "step": 57000
864
+ },
865
+ {
866
+ "epoch": 1.97,
867
+ "learning_rate": 6.865942028985509e-06,
868
+ "loss": 0.3672,
869
+ "step": 58000
870
+ },
871
+ {
872
+ "epoch": 1.97,
873
+ "eval_accuracy": 0.8084564442180336,
874
+ "eval_loss": 0.5128825902938843,
875
+ "eval_runtime": 46.1412,
876
+ "eval_samples_per_second": 212.717,
877
+ "eval_steps_per_second": 6.653,
878
+ "step": 58000
879
+ },
880
+ {
881
+ "epoch": 2.0,
882
+ "learning_rate": 6.639492753623189e-06,
883
+ "loss": 0.3545,
884
+ "step": 59000
885
+ },
886
+ {
887
+ "epoch": 2.0,
888
+ "eval_accuracy": 0.8109016811003565,
889
+ "eval_loss": 0.5467332601547241,
890
+ "eval_runtime": 46.0619,
891
+ "eval_samples_per_second": 213.083,
892
+ "eval_steps_per_second": 6.665,
893
+ "step": 59000
894
+ },
895
+ {
896
+ "epoch": 2.04,
897
+ "learning_rate": 6.41304347826087e-06,
898
+ "loss": 0.2938,
899
+ "step": 60000
900
+ },
901
+ {
902
+ "epoch": 2.04,
903
+ "eval_accuracy": 0.804890473764646,
904
+ "eval_loss": 0.5634797811508179,
905
+ "eval_runtime": 46.0511,
906
+ "eval_samples_per_second": 213.133,
907
+ "eval_steps_per_second": 6.667,
908
+ "step": 60000
909
+ },
910
+ {
911
+ "epoch": 2.07,
912
+ "learning_rate": 6.186594202898551e-06,
913
+ "loss": 0.29,
914
+ "step": 61000
915
+ },
916
+ {
917
+ "epoch": 2.07,
918
+ "eval_accuracy": 0.8040753948038716,
919
+ "eval_loss": 0.5781472325325012,
920
+ "eval_runtime": 47.2218,
921
+ "eval_samples_per_second": 207.849,
922
+ "eval_steps_per_second": 6.501,
923
+ "step": 61000
924
+ },
925
+ {
926
+ "epoch": 2.11,
927
+ "learning_rate": 5.960144927536232e-06,
928
+ "loss": 0.2992,
929
+ "step": 62000
930
+ },
931
+ {
932
+ "epoch": 2.11,
933
+ "eval_accuracy": 0.807743250127356,
934
+ "eval_loss": 0.5470119714736938,
935
+ "eval_runtime": 46.0932,
936
+ "eval_samples_per_second": 212.938,
937
+ "eval_steps_per_second": 6.66,
938
+ "step": 62000
939
+ },
940
+ {
941
+ "epoch": 2.14,
942
+ "learning_rate": 5.733695652173914e-06,
943
+ "loss": 0.2957,
944
+ "step": 63000
945
+ },
946
+ {
947
+ "epoch": 2.14,
948
+ "eval_accuracy": 0.807335710646969,
949
+ "eval_loss": 0.5764774084091187,
950
+ "eval_runtime": 45.9878,
951
+ "eval_samples_per_second": 213.426,
952
+ "eval_steps_per_second": 6.676,
953
+ "step": 63000
954
+ },
955
+ {
956
+ "epoch": 2.17,
957
+ "learning_rate": 5.507246376811595e-06,
958
+ "loss": 0.292,
959
+ "step": 64000
960
+ },
961
+ {
962
+ "epoch": 2.17,
963
+ "eval_accuracy": 0.8105960264900662,
964
+ "eval_loss": 0.547173798084259,
965
+ "eval_runtime": 45.972,
966
+ "eval_samples_per_second": 213.5,
967
+ "eval_steps_per_second": 6.678,
968
+ "step": 64000
969
+ },
970
+ {
971
+ "epoch": 2.21,
972
+ "learning_rate": 5.2807971014492755e-06,
973
+ "loss": 0.2893,
974
+ "step": 65000
975
+ },
976
+ {
977
+ "epoch": 2.21,
978
+ "eval_accuracy": 0.8084564442180336,
979
+ "eval_loss": 0.5590358376502991,
980
+ "eval_runtime": 45.9212,
981
+ "eval_samples_per_second": 213.736,
982
+ "eval_steps_per_second": 6.685,
983
+ "step": 65000
984
+ },
985
+ {
986
+ "epoch": 2.24,
987
+ "learning_rate": 5.054347826086957e-06,
988
+ "loss": 0.2883,
989
+ "step": 66000
990
+ },
991
+ {
992
+ "epoch": 2.24,
993
+ "eval_accuracy": 0.8064187468160978,
994
+ "eval_loss": 0.5534654855728149,
995
+ "eval_runtime": 45.9485,
996
+ "eval_samples_per_second": 213.609,
997
+ "eval_steps_per_second": 6.681,
998
+ "step": 66000
999
+ },
1000
+ {
1001
+ "epoch": 2.28,
1002
+ "learning_rate": 4.8278985507246375e-06,
1003
+ "loss": 0.2923,
1004
+ "step": 67000
1005
+ },
1006
+ {
1007
+ "epoch": 2.28,
1008
+ "eval_accuracy": 0.8094752929190016,
1009
+ "eval_loss": 0.5508309602737427,
1010
+ "eval_runtime": 45.9406,
1011
+ "eval_samples_per_second": 213.645,
1012
+ "eval_steps_per_second": 6.683,
1013
+ "step": 67000
1014
+ },
1015
+ {
1016
+ "epoch": 2.31,
1017
+ "learning_rate": 4.601449275362319e-06,
1018
+ "loss": 0.2868,
1019
+ "step": 68000
1020
+ },
1021
+ {
1022
+ "epoch": 2.31,
1023
+ "eval_accuracy": 0.8097809475292919,
1024
+ "eval_loss": 0.5678631067276001,
1025
+ "eval_runtime": 45.9005,
1026
+ "eval_samples_per_second": 213.832,
1027
+ "eval_steps_per_second": 6.688,
1028
+ "step": 68000
1029
+ },
1030
+ {
1031
+ "epoch": 2.34,
1032
+ "learning_rate": 4.3750000000000005e-06,
1033
+ "loss": 0.2892,
1034
+ "step": 69000
1035
+ },
1036
+ {
1037
+ "epoch": 2.34,
1038
+ "eval_accuracy": 0.8057055527254203,
1039
+ "eval_loss": 0.5659688115119934,
1040
+ "eval_runtime": 46.0172,
1041
+ "eval_samples_per_second": 213.29,
1042
+ "eval_steps_per_second": 6.671,
1043
+ "step": 69000
1044
+ },
1045
+ {
1046
+ "epoch": 2.38,
1047
+ "learning_rate": 4.148550724637682e-06,
1048
+ "loss": 0.292,
1049
+ "step": 70000
1050
+ },
1051
+ {
1052
+ "epoch": 2.38,
1053
+ "eval_accuracy": 0.808762098828324,
1054
+ "eval_loss": 0.54941326379776,
1055
+ "eval_runtime": 46.0397,
1056
+ "eval_samples_per_second": 213.186,
1057
+ "eval_steps_per_second": 6.668,
1058
+ "step": 70000
1059
+ },
1060
+ {
1061
+ "epoch": 2.41,
1062
+ "learning_rate": 3.9221014492753625e-06,
1063
+ "loss": 0.286,
1064
+ "step": 71000
1065
+ },
1066
+ {
1067
+ "epoch": 2.41,
1068
+ "eval_accuracy": 0.8084564442180336,
1069
+ "eval_loss": 0.5653256177902222,
1070
+ "eval_runtime": 47.1102,
1071
+ "eval_samples_per_second": 208.341,
1072
+ "eval_steps_per_second": 6.517,
1073
+ "step": 71000
1074
+ },
1075
+ {
1076
+ "epoch": 2.45,
1077
+ "learning_rate": 3.6956521739130436e-06,
1078
+ "loss": 0.2939,
1079
+ "step": 72000
1080
+ },
1081
+ {
1082
+ "epoch": 2.45,
1083
+ "eval_accuracy": 0.8070300560366785,
1084
+ "eval_loss": 0.5673069953918457,
1085
+ "eval_runtime": 46.0492,
1086
+ "eval_samples_per_second": 213.142,
1087
+ "eval_steps_per_second": 6.667,
1088
+ "step": 72000
1089
+ },
1090
+ {
1091
+ "epoch": 2.48,
1092
+ "learning_rate": 3.4692028985507246e-06,
1093
+ "loss": 0.286,
1094
+ "step": 73000
1095
+ },
1096
+ {
1097
+ "epoch": 2.48,
1098
+ "eval_accuracy": 0.8091696383087111,
1099
+ "eval_loss": 0.56004798412323,
1100
+ "eval_runtime": 45.9369,
1101
+ "eval_samples_per_second": 213.663,
1102
+ "eval_steps_per_second": 6.683,
1103
+ "step": 73000
1104
+ },
1105
+ {
1106
+ "epoch": 2.51,
1107
+ "learning_rate": 3.242753623188406e-06,
1108
+ "loss": 0.2844,
1109
+ "step": 74000
1110
+ },
1111
+ {
1112
+ "epoch": 2.51,
1113
+ "eval_accuracy": 0.8094752929190016,
1114
+ "eval_loss": 0.5507674217224121,
1115
+ "eval_runtime": 45.8631,
1116
+ "eval_samples_per_second": 214.006,
1117
+ "eval_steps_per_second": 6.694,
1118
+ "step": 74000
1119
+ },
1120
+ {
1121
+ "epoch": 2.55,
1122
+ "learning_rate": 3.016304347826087e-06,
1123
+ "loss": 0.2913,
1124
+ "step": 75000
1125
+ },
1126
+ {
1127
+ "epoch": 2.55,
1128
+ "eval_accuracy": 0.808762098828324,
1129
+ "eval_loss": 0.5644803047180176,
1130
+ "eval_runtime": 45.9087,
1131
+ "eval_samples_per_second": 213.794,
1132
+ "eval_steps_per_second": 6.687,
1133
+ "step": 75000
1134
+ },
1135
+ {
1136
+ "epoch": 2.58,
1137
+ "learning_rate": 2.7898550724637686e-06,
1138
+ "loss": 0.2859,
1139
+ "step": 76000
1140
+ },
1141
+ {
1142
+ "epoch": 2.58,
1143
+ "eval_accuracy": 0.8094752929190016,
1144
+ "eval_loss": 0.5677375197410583,
1145
+ "eval_runtime": 46.1274,
1146
+ "eval_samples_per_second": 212.78,
1147
+ "eval_steps_per_second": 6.655,
1148
+ "step": 76000
1149
+ },
1150
+ {
1151
+ "epoch": 2.62,
1152
+ "learning_rate": 2.563405797101449e-06,
1153
+ "loss": 0.2892,
1154
+ "step": 77000
1155
+ },
1156
+ {
1157
+ "epoch": 2.62,
1158
+ "eval_accuracy": 0.8113092205807437,
1159
+ "eval_loss": 0.5597835183143616,
1160
+ "eval_runtime": 46.0994,
1161
+ "eval_samples_per_second": 212.909,
1162
+ "eval_steps_per_second": 6.66,
1163
+ "step": 77000
1164
+ },
1165
+ {
1166
+ "epoch": 2.65,
1167
+ "learning_rate": 2.3369565217391307e-06,
1168
+ "loss": 0.2898,
1169
+ "step": 78000
1170
+ },
1171
+ {
1172
+ "epoch": 2.65,
1173
+ "eval_accuracy": 0.8095771777890983,
1174
+ "eval_loss": 0.5617640018463135,
1175
+ "eval_runtime": 45.9468,
1176
+ "eval_samples_per_second": 213.617,
1177
+ "eval_steps_per_second": 6.682,
1178
+ "step": 78000
1179
+ },
1180
+ {
1181
+ "epoch": 2.68,
1182
+ "learning_rate": 2.1105072463768117e-06,
1183
+ "loss": 0.2814,
1184
+ "step": 79000
1185
+ },
1186
+ {
1187
+ "epoch": 2.68,
1188
+ "eval_accuracy": 0.8102903718797758,
1189
+ "eval_loss": 0.5664277672767639,
1190
+ "eval_runtime": 46.0837,
1191
+ "eval_samples_per_second": 212.982,
1192
+ "eval_steps_per_second": 6.662,
1193
+ "step": 79000
1194
+ },
1195
+ {
1196
+ "epoch": 2.72,
1197
+ "learning_rate": 1.884057971014493e-06,
1198
+ "loss": 0.2917,
1199
+ "step": 80000
1200
+ },
1201
+ {
1202
+ "epoch": 2.72,
1203
+ "eval_accuracy": 0.8122261844116149,
1204
+ "eval_loss": 0.5484075546264648,
1205
+ "eval_runtime": 46.0805,
1206
+ "eval_samples_per_second": 212.997,
1207
+ "eval_steps_per_second": 6.662,
1208
+ "step": 80000
1209
+ },
1210
+ {
1211
+ "epoch": 2.75,
1212
+ "learning_rate": 1.657608695652174e-06,
1213
+ "loss": 0.2907,
1214
+ "step": 81000
1215
+ },
1216
+ {
1217
+ "epoch": 2.75,
1218
+ "eval_accuracy": 0.8116148751910341,
1219
+ "eval_loss": 0.5522238612174988,
1220
+ "eval_runtime": 47.1215,
1221
+ "eval_samples_per_second": 208.291,
1222
+ "eval_steps_per_second": 6.515,
1223
+ "step": 81000
1224
+ },
1225
+ {
1226
+ "epoch": 2.79,
1227
+ "learning_rate": 1.4311594202898552e-06,
1228
+ "loss": 0.2896,
1229
+ "step": 82000
1230
+ },
1231
+ {
1232
+ "epoch": 2.79,
1233
+ "eval_accuracy": 0.8092715231788079,
1234
+ "eval_loss": 0.5539634823799133,
1235
+ "eval_runtime": 46.0412,
1236
+ "eval_samples_per_second": 213.179,
1237
+ "eval_steps_per_second": 6.668,
1238
+ "step": 82000
1239
+ },
1240
+ {
1241
+ "epoch": 2.82,
1242
+ "learning_rate": 1.2047101449275363e-06,
1243
+ "loss": 0.2907,
1244
+ "step": 83000
1245
+ },
1246
+ {
1247
+ "epoch": 2.82,
1248
+ "eval_accuracy": 0.8103922567498727,
1249
+ "eval_loss": 0.5468757152557373,
1250
+ "eval_runtime": 46.1268,
1251
+ "eval_samples_per_second": 212.783,
1252
+ "eval_steps_per_second": 6.656,
1253
+ "step": 83000
1254
+ },
1255
+ {
1256
+ "epoch": 2.85,
1257
+ "learning_rate": 9.782608695652175e-07,
1258
+ "loss": 0.2882,
1259
+ "step": 84000
1260
+ },
1261
+ {
1262
+ "epoch": 2.85,
1263
+ "eval_accuracy": 0.8122261844116149,
1264
+ "eval_loss": 0.5470659732818604,
1265
+ "eval_runtime": 46.0408,
1266
+ "eval_samples_per_second": 213.18,
1267
+ "eval_steps_per_second": 6.668,
1268
+ "step": 84000
1269
+ },
1270
+ {
1271
+ "epoch": 2.89,
1272
+ "learning_rate": 7.518115942028987e-07,
1273
+ "loss": 0.2878,
1274
+ "step": 85000
1275
+ },
1276
+ {
1277
+ "epoch": 2.89,
1278
+ "eval_accuracy": 0.8107997962302598,
1279
+ "eval_loss": 0.5532234907150269,
1280
+ "eval_runtime": 46.1642,
1281
+ "eval_samples_per_second": 212.61,
1282
+ "eval_steps_per_second": 6.65,
1283
+ "step": 85000
1284
+ },
1285
+ {
1286
+ "epoch": 2.92,
1287
+ "learning_rate": 5.253623188405797e-07,
1288
+ "loss": 0.2858,
1289
+ "step": 86000
1290
+ },
1291
+ {
1292
+ "epoch": 2.92,
1293
+ "eval_accuracy": 0.8115129903209374,
1294
+ "eval_loss": 0.5510849952697754,
1295
+ "eval_runtime": 46.0337,
1296
+ "eval_samples_per_second": 213.214,
1297
+ "eval_steps_per_second": 6.669,
1298
+ "step": 86000
1299
+ },
1300
+ {
1301
+ "epoch": 2.96,
1302
+ "learning_rate": 2.989130434782609e-07,
1303
+ "loss": 0.288,
1304
+ "step": 87000
1305
+ },
1306
+ {
1307
+ "epoch": 2.96,
1308
+ "eval_accuracy": 0.8111054508405502,
1309
+ "eval_loss": 0.5491093397140503,
1310
+ "eval_runtime": 45.9486,
1311
+ "eval_samples_per_second": 213.608,
1312
+ "eval_steps_per_second": 6.681,
1313
+ "step": 87000
1314
+ },
1315
+ {
1316
+ "epoch": 2.99,
1317
+ "learning_rate": 7.246376811594204e-08,
1318
+ "loss": 0.2834,
1319
+ "step": 88000
1320
+ },
1321
+ {
1322
+ "epoch": 2.99,
1323
+ "eval_accuracy": 0.8111054508405502,
1324
+ "eval_loss": 0.5540585517883301,
1325
+ "eval_runtime": 45.8903,
1326
+ "eval_samples_per_second": 213.879,
1327
+ "eval_steps_per_second": 6.69,
1328
+ "step": 88000
1329
+ },
1330
+ {
1331
+ "epoch": 3.0,
1332
+ "step": 88320,
1333
+ "total_flos": 9.711893491230248e+16,
1334
+ "train_loss": 0.3844692539477694,
1335
+ "train_runtime": 45994.9721,
1336
+ "train_samples_per_second": 61.446,
1337
+ "train_steps_per_second": 1.92
1338
+ }
1339
+ ],
1340
+ "max_steps": 88320,
1341
+ "num_train_epochs": 3,
1342
+ "total_flos": 9.711893491230248e+16,
1343
+ "trial_name": null,
1344
+ "trial_params": null
1345
+ }