emrecan commited on
Commit
29c4bf7
1 Parent(s): 3974fd6

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": 0.5041872973027437,
4
+ "train_runtime": 53589.3429,
5
+ "train_samples_per_second": 52.738,
6
+ "train_steps_per_second": 1.648
7
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.5041872973027437,
4
+ "train_runtime": 53589.3429,
5
+ "train_samples_per_second": 52.738,
6
+ "train_steps_per_second": 1.648
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1345 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.7648497198166072,
3
+ "best_model_checkpoint": "/home/user/emrecan/models/bert-base-multilingual-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": 0.8623,
14
+ "step": 1000
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "eval_accuracy": 0.5917473255221599,
19
+ "eval_loss": 0.9075943827629089,
20
+ "eval_runtime": 49.3629,
21
+ "eval_samples_per_second": 198.833,
22
+ "eval_steps_per_second": 6.219,
23
+ "step": 1000
24
+ },
25
+ {
26
+ "epoch": 0.07,
27
+ "learning_rate": 1.9547101449275363e-05,
28
+ "loss": 0.7528,
29
+ "step": 2000
30
+ },
31
+ {
32
+ "epoch": 0.07,
33
+ "eval_accuracy": 0.6119205298013245,
34
+ "eval_loss": 0.8586783409118652,
35
+ "eval_runtime": 49.2395,
36
+ "eval_samples_per_second": 199.332,
37
+ "eval_steps_per_second": 6.235,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.1,
42
+ "learning_rate": 1.9320652173913047e-05,
43
+ "loss": 0.7074,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.1,
48
+ "eval_accuracy": 0.664696892511462,
49
+ "eval_loss": 0.7866753935813904,
50
+ "eval_runtime": 49.3359,
51
+ "eval_samples_per_second": 198.942,
52
+ "eval_steps_per_second": 6.223,
53
+ "step": 3000
54
+ },
55
+ {
56
+ "epoch": 0.14,
57
+ "learning_rate": 1.9094202898550727e-05,
58
+ "loss": 0.6949,
59
+ "step": 4000
60
+ },
61
+ {
62
+ "epoch": 0.14,
63
+ "eval_accuracy": 0.6772287315333673,
64
+ "eval_loss": 0.7473878264427185,
65
+ "eval_runtime": 49.2961,
66
+ "eval_samples_per_second": 199.103,
67
+ "eval_steps_per_second": 6.228,
68
+ "step": 4000
69
+ },
70
+ {
71
+ "epoch": 0.17,
72
+ "learning_rate": 1.8867753623188408e-05,
73
+ "loss": 0.6681,
74
+ "step": 5000
75
+ },
76
+ {
77
+ "epoch": 0.17,
78
+ "eval_accuracy": 0.6814060112073357,
79
+ "eval_loss": 0.7661426067352295,
80
+ "eval_runtime": 49.2761,
81
+ "eval_samples_per_second": 199.184,
82
+ "eval_steps_per_second": 6.23,
83
+ "step": 5000
84
+ },
85
+ {
86
+ "epoch": 0.2,
87
+ "learning_rate": 1.864130434782609e-05,
88
+ "loss": 0.6597,
89
+ "step": 6000
90
+ },
91
+ {
92
+ "epoch": 0.2,
93
+ "eval_accuracy": 0.6943453897096281,
94
+ "eval_loss": 0.7264170050621033,
95
+ "eval_runtime": 49.2718,
96
+ "eval_samples_per_second": 199.201,
97
+ "eval_steps_per_second": 6.231,
98
+ "step": 6000
99
+ },
100
+ {
101
+ "epoch": 0.24,
102
+ "learning_rate": 1.841485507246377e-05,
103
+ "loss": 0.6495,
104
+ "step": 7000
105
+ },
106
+ {
107
+ "epoch": 0.24,
108
+ "eval_accuracy": 0.6781456953642384,
109
+ "eval_loss": 0.7841250896453857,
110
+ "eval_runtime": 49.2905,
111
+ "eval_samples_per_second": 199.126,
112
+ "eval_steps_per_second": 6.228,
113
+ "step": 7000
114
+ },
115
+ {
116
+ "epoch": 0.27,
117
+ "learning_rate": 1.818840579710145e-05,
118
+ "loss": 0.6323,
119
+ "step": 8000
120
+ },
121
+ {
122
+ "epoch": 0.27,
123
+ "eval_accuracy": 0.6951604686704025,
124
+ "eval_loss": 0.7256377339363098,
125
+ "eval_runtime": 49.2057,
126
+ "eval_samples_per_second": 199.469,
127
+ "eval_steps_per_second": 6.239,
128
+ "step": 8000
129
+ },
130
+ {
131
+ "epoch": 0.31,
132
+ "learning_rate": 1.7961956521739134e-05,
133
+ "loss": 0.6308,
134
+ "step": 9000
135
+ },
136
+ {
137
+ "epoch": 0.31,
138
+ "eval_accuracy": 0.6957717778909832,
139
+ "eval_loss": 0.7319093346595764,
140
+ "eval_runtime": 49.2073,
141
+ "eval_samples_per_second": 199.462,
142
+ "eval_steps_per_second": 6.239,
143
+ "step": 9000
144
+ },
145
+ {
146
+ "epoch": 0.34,
147
+ "learning_rate": 1.7735507246376815e-05,
148
+ "loss": 0.6254,
149
+ "step": 10000
150
+ },
151
+ {
152
+ "epoch": 0.34,
153
+ "eval_accuracy": 0.7003565970453388,
154
+ "eval_loss": 0.7053707838058472,
155
+ "eval_runtime": 49.2797,
156
+ "eval_samples_per_second": 199.169,
157
+ "eval_steps_per_second": 6.23,
158
+ "step": 10000
159
+ },
160
+ {
161
+ "epoch": 0.37,
162
+ "learning_rate": 1.7509057971014495e-05,
163
+ "loss": 0.6233,
164
+ "step": 11000
165
+ },
166
+ {
167
+ "epoch": 0.37,
168
+ "eval_accuracy": 0.708507386653082,
169
+ "eval_loss": 0.7068630456924438,
170
+ "eval_runtime": 55.8024,
171
+ "eval_samples_per_second": 175.889,
172
+ "eval_steps_per_second": 5.502,
173
+ "step": 11000
174
+ },
175
+ {
176
+ "epoch": 0.41,
177
+ "learning_rate": 1.7282608695652176e-05,
178
+ "loss": 0.6165,
179
+ "step": 12000
180
+ },
181
+ {
182
+ "epoch": 0.41,
183
+ "eval_accuracy": 0.7180845644421804,
184
+ "eval_loss": 0.687969982624054,
185
+ "eval_runtime": 50.9115,
186
+ "eval_samples_per_second": 192.785,
187
+ "eval_steps_per_second": 6.03,
188
+ "step": 12000
189
+ },
190
+ {
191
+ "epoch": 0.44,
192
+ "learning_rate": 1.7056159420289856e-05,
193
+ "loss": 0.6033,
194
+ "step": 13000
195
+ },
196
+ {
197
+ "epoch": 0.44,
198
+ "eval_accuracy": 0.719714722363729,
199
+ "eval_loss": 0.6844114065170288,
200
+ "eval_runtime": 49.3739,
201
+ "eval_samples_per_second": 198.789,
202
+ "eval_steps_per_second": 6.218,
203
+ "step": 13000
204
+ },
205
+ {
206
+ "epoch": 0.48,
207
+ "learning_rate": 1.6829710144927537e-05,
208
+ "loss": 0.6014,
209
+ "step": 14000
210
+ },
211
+ {
212
+ "epoch": 0.48,
213
+ "eval_accuracy": 0.7128884360672441,
214
+ "eval_loss": 0.6752753853797913,
215
+ "eval_runtime": 49.3697,
216
+ "eval_samples_per_second": 198.806,
217
+ "eval_steps_per_second": 6.218,
218
+ "step": 14000
219
+ },
220
+ {
221
+ "epoch": 0.51,
222
+ "learning_rate": 1.6603260869565218e-05,
223
+ "loss": 0.5947,
224
+ "step": 15000
225
+ },
226
+ {
227
+ "epoch": 0.51,
228
+ "eval_accuracy": 0.7039225674987264,
229
+ "eval_loss": 0.7000291347503662,
230
+ "eval_runtime": 49.3889,
231
+ "eval_samples_per_second": 198.729,
232
+ "eval_steps_per_second": 6.216,
233
+ "step": 15000
234
+ },
235
+ {
236
+ "epoch": 0.54,
237
+ "learning_rate": 1.6376811594202898e-05,
238
+ "loss": 0.5965,
239
+ "step": 16000
240
+ },
241
+ {
242
+ "epoch": 0.54,
243
+ "eval_accuracy": 0.7263372389200203,
244
+ "eval_loss": 0.670754075050354,
245
+ "eval_runtime": 49.311,
246
+ "eval_samples_per_second": 199.043,
247
+ "eval_steps_per_second": 6.226,
248
+ "step": 16000
249
+ },
250
+ {
251
+ "epoch": 0.58,
252
+ "learning_rate": 1.615036231884058e-05,
253
+ "loss": 0.5979,
254
+ "step": 17000
255
+ },
256
+ {
257
+ "epoch": 0.58,
258
+ "eval_accuracy": 0.7284768211920529,
259
+ "eval_loss": 0.656209409236908,
260
+ "eval_runtime": 49.4043,
261
+ "eval_samples_per_second": 198.667,
262
+ "eval_steps_per_second": 6.214,
263
+ "step": 17000
264
+ },
265
+ {
266
+ "epoch": 0.61,
267
+ "learning_rate": 1.5923913043478263e-05,
268
+ "loss": 0.5787,
269
+ "step": 18000
270
+ },
271
+ {
272
+ "epoch": 0.61,
273
+ "eval_accuracy": 0.7296994396332145,
274
+ "eval_loss": 0.6554355621337891,
275
+ "eval_runtime": 50.2795,
276
+ "eval_samples_per_second": 195.209,
277
+ "eval_steps_per_second": 6.106,
278
+ "step": 18000
279
+ },
280
+ {
281
+ "epoch": 0.65,
282
+ "learning_rate": 1.5697463768115943e-05,
283
+ "loss": 0.58,
284
+ "step": 19000
285
+ },
286
+ {
287
+ "epoch": 0.65,
288
+ "eval_accuracy": 0.7315333672949567,
289
+ "eval_loss": 0.6544056534767151,
290
+ "eval_runtime": 49.3813,
291
+ "eval_samples_per_second": 198.76,
292
+ "eval_steps_per_second": 6.217,
293
+ "step": 19000
294
+ },
295
+ {
296
+ "epoch": 0.68,
297
+ "learning_rate": 1.5471014492753624e-05,
298
+ "loss": 0.574,
299
+ "step": 20000
300
+ },
301
+ {
302
+ "epoch": 0.68,
303
+ "eval_accuracy": 0.7338767193071829,
304
+ "eval_loss": 0.6549123525619507,
305
+ "eval_runtime": 49.3189,
306
+ "eval_samples_per_second": 199.011,
307
+ "eval_steps_per_second": 6.225,
308
+ "step": 20000
309
+ },
310
+ {
311
+ "epoch": 0.71,
312
+ "learning_rate": 1.5244565217391305e-05,
313
+ "loss": 0.5751,
314
+ "step": 21000
315
+ },
316
+ {
317
+ "epoch": 0.71,
318
+ "eval_accuracy": 0.7288843606724401,
319
+ "eval_loss": 0.6545295715332031,
320
+ "eval_runtime": 52.3727,
321
+ "eval_samples_per_second": 187.407,
322
+ "eval_steps_per_second": 5.862,
323
+ "step": 21000
324
+ },
325
+ {
326
+ "epoch": 0.75,
327
+ "learning_rate": 1.5018115942028985e-05,
328
+ "loss": 0.5659,
329
+ "step": 22000
330
+ },
331
+ {
332
+ "epoch": 0.75,
333
+ "eval_accuracy": 0.7371370351502802,
334
+ "eval_loss": 0.6466848850250244,
335
+ "eval_runtime": 51.7409,
336
+ "eval_samples_per_second": 189.695,
337
+ "eval_steps_per_second": 5.933,
338
+ "step": 22000
339
+ },
340
+ {
341
+ "epoch": 0.78,
342
+ "learning_rate": 1.479166666666667e-05,
343
+ "loss": 0.5732,
344
+ "step": 23000
345
+ },
346
+ {
347
+ "epoch": 0.78,
348
+ "eval_accuracy": 0.736220071319409,
349
+ "eval_loss": 0.6447662711143494,
350
+ "eval_runtime": 51.1755,
351
+ "eval_samples_per_second": 191.791,
352
+ "eval_steps_per_second": 5.999,
353
+ "step": 23000
354
+ },
355
+ {
356
+ "epoch": 0.82,
357
+ "learning_rate": 1.456521739130435e-05,
358
+ "loss": 0.5637,
359
+ "step": 24000
360
+ },
361
+ {
362
+ "epoch": 0.82,
363
+ "eval_accuracy": 0.7355068772287315,
364
+ "eval_loss": 0.6520141959190369,
365
+ "eval_runtime": 51.0797,
366
+ "eval_samples_per_second": 192.151,
367
+ "eval_steps_per_second": 6.01,
368
+ "step": 24000
369
+ },
370
+ {
371
+ "epoch": 0.85,
372
+ "learning_rate": 1.433876811594203e-05,
373
+ "loss": 0.5648,
374
+ "step": 25000
375
+ },
376
+ {
377
+ "epoch": 0.85,
378
+ "eval_accuracy": 0.7344880285277636,
379
+ "eval_loss": 0.6411919593811035,
380
+ "eval_runtime": 51.1242,
381
+ "eval_samples_per_second": 191.983,
382
+ "eval_steps_per_second": 6.005,
383
+ "step": 25000
384
+ },
385
+ {
386
+ "epoch": 0.88,
387
+ "learning_rate": 1.4112318840579711e-05,
388
+ "loss": 0.5622,
389
+ "step": 26000
390
+ },
391
+ {
392
+ "epoch": 0.88,
393
+ "eval_accuracy": 0.735812531839022,
394
+ "eval_loss": 0.6350020170211792,
395
+ "eval_runtime": 51.0256,
396
+ "eval_samples_per_second": 192.355,
397
+ "eval_steps_per_second": 6.017,
398
+ "step": 26000
399
+ },
400
+ {
401
+ "epoch": 0.92,
402
+ "learning_rate": 1.3885869565217392e-05,
403
+ "loss": 0.5579,
404
+ "step": 27000
405
+ },
406
+ {
407
+ "epoch": 0.92,
408
+ "eval_accuracy": 0.7392766174223128,
409
+ "eval_loss": 0.6346594095230103,
410
+ "eval_runtime": 51.1443,
411
+ "eval_samples_per_second": 191.908,
412
+ "eval_steps_per_second": 6.003,
413
+ "step": 27000
414
+ },
415
+ {
416
+ "epoch": 0.95,
417
+ "learning_rate": 1.3659420289855074e-05,
418
+ "loss": 0.5518,
419
+ "step": 28000
420
+ },
421
+ {
422
+ "epoch": 0.95,
423
+ "eval_accuracy": 0.739174732552216,
424
+ "eval_loss": 0.6417071223258972,
425
+ "eval_runtime": 52.294,
426
+ "eval_samples_per_second": 187.689,
427
+ "eval_steps_per_second": 5.871,
428
+ "step": 28000
429
+ },
430
+ {
431
+ "epoch": 0.99,
432
+ "learning_rate": 1.3432971014492755e-05,
433
+ "loss": 0.5547,
434
+ "step": 29000
435
+ },
436
+ {
437
+ "epoch": 0.99,
438
+ "eval_accuracy": 0.7436576668364748,
439
+ "eval_loss": 0.6321312785148621,
440
+ "eval_runtime": 51.1442,
441
+ "eval_samples_per_second": 191.908,
442
+ "eval_steps_per_second": 6.003,
443
+ "step": 29000
444
+ },
445
+ {
446
+ "epoch": 1.02,
447
+ "learning_rate": 1.3206521739130435e-05,
448
+ "loss": 0.524,
449
+ "step": 30000
450
+ },
451
+ {
452
+ "epoch": 1.02,
453
+ "eval_accuracy": 0.7412124299541518,
454
+ "eval_loss": 0.6430493593215942,
455
+ "eval_runtime": 51.2108,
456
+ "eval_samples_per_second": 191.659,
457
+ "eval_steps_per_second": 5.995,
458
+ "step": 30000
459
+ },
460
+ {
461
+ "epoch": 1.05,
462
+ "learning_rate": 1.2980072463768116e-05,
463
+ "loss": 0.4982,
464
+ "step": 31000
465
+ },
466
+ {
467
+ "epoch": 1.05,
468
+ "eval_accuracy": 0.7457972491085074,
469
+ "eval_loss": 0.6252649426460266,
470
+ "eval_runtime": 57.7511,
471
+ "eval_samples_per_second": 169.954,
472
+ "eval_steps_per_second": 5.316,
473
+ "step": 31000
474
+ },
475
+ {
476
+ "epoch": 1.09,
477
+ "learning_rate": 1.2753623188405797e-05,
478
+ "loss": 0.5002,
479
+ "step": 32000
480
+ },
481
+ {
482
+ "epoch": 1.09,
483
+ "eval_accuracy": 0.7418237391747325,
484
+ "eval_loss": 0.6316350698471069,
485
+ "eval_runtime": 52.7966,
486
+ "eval_samples_per_second": 185.902,
487
+ "eval_steps_per_second": 5.815,
488
+ "step": 32000
489
+ },
490
+ {
491
+ "epoch": 1.12,
492
+ "learning_rate": 1.252717391304348e-05,
493
+ "loss": 0.4993,
494
+ "step": 33000
495
+ },
496
+ {
497
+ "epoch": 1.12,
498
+ "eval_accuracy": 0.7486500254712175,
499
+ "eval_loss": 0.6196975111961365,
500
+ "eval_runtime": 51.1876,
501
+ "eval_samples_per_second": 191.746,
502
+ "eval_steps_per_second": 5.998,
503
+ "step": 33000
504
+ },
505
+ {
506
+ "epoch": 1.15,
507
+ "learning_rate": 1.2300724637681161e-05,
508
+ "loss": 0.4963,
509
+ "step": 34000
510
+ },
511
+ {
512
+ "epoch": 1.15,
513
+ "eval_accuracy": 0.7462047885888945,
514
+ "eval_loss": 0.6307246685028076,
515
+ "eval_runtime": 51.0522,
516
+ "eval_samples_per_second": 192.254,
517
+ "eval_steps_per_second": 6.013,
518
+ "step": 34000
519
+ },
520
+ {
521
+ "epoch": 1.19,
522
+ "learning_rate": 1.2074275362318842e-05,
523
+ "loss": 0.504,
524
+ "step": 35000
525
+ },
526
+ {
527
+ "epoch": 1.19,
528
+ "eval_accuracy": 0.7480387162506368,
529
+ "eval_loss": 0.627221405506134,
530
+ "eval_runtime": 51.0856,
531
+ "eval_samples_per_second": 192.129,
532
+ "eval_steps_per_second": 6.01,
533
+ "step": 35000
534
+ },
535
+ {
536
+ "epoch": 1.22,
537
+ "learning_rate": 1.1847826086956522e-05,
538
+ "loss": 0.4922,
539
+ "step": 36000
540
+ },
541
+ {
542
+ "epoch": 1.22,
543
+ "eval_accuracy": 0.7432501273560876,
544
+ "eval_loss": 0.6410390138626099,
545
+ "eval_runtime": 51.1349,
546
+ "eval_samples_per_second": 191.943,
547
+ "eval_steps_per_second": 6.004,
548
+ "step": 36000
549
+ },
550
+ {
551
+ "epoch": 1.26,
552
+ "learning_rate": 1.1621376811594205e-05,
553
+ "loss": 0.5016,
554
+ "step": 37000
555
+ },
556
+ {
557
+ "epoch": 1.26,
558
+ "eval_accuracy": 0.7461029037187977,
559
+ "eval_loss": 0.6295490860939026,
560
+ "eval_runtime": 51.0732,
561
+ "eval_samples_per_second": 192.175,
562
+ "eval_steps_per_second": 6.011,
563
+ "step": 37000
564
+ },
565
+ {
566
+ "epoch": 1.29,
567
+ "learning_rate": 1.1394927536231885e-05,
568
+ "loss": 0.4957,
569
+ "step": 38000
570
+ },
571
+ {
572
+ "epoch": 1.29,
573
+ "eval_accuracy": 0.7505858380030566,
574
+ "eval_loss": 0.6182788014411926,
575
+ "eval_runtime": 51.1969,
576
+ "eval_samples_per_second": 191.711,
577
+ "eval_steps_per_second": 5.996,
578
+ "step": 38000
579
+ },
580
+ {
581
+ "epoch": 1.32,
582
+ "learning_rate": 1.1168478260869566e-05,
583
+ "loss": 0.4883,
584
+ "step": 39000
585
+ },
586
+ {
587
+ "epoch": 1.32,
588
+ "eval_accuracy": 0.7501782985226694,
589
+ "eval_loss": 0.6260754466056824,
590
+ "eval_runtime": 51.1903,
591
+ "eval_samples_per_second": 191.735,
592
+ "eval_steps_per_second": 5.997,
593
+ "step": 39000
594
+ },
595
+ {
596
+ "epoch": 1.36,
597
+ "learning_rate": 1.0942028985507247e-05,
598
+ "loss": 0.4985,
599
+ "step": 40000
600
+ },
601
+ {
602
+ "epoch": 1.36,
603
+ "eval_accuracy": 0.7495669893020886,
604
+ "eval_loss": 0.6315430402755737,
605
+ "eval_runtime": 51.1506,
606
+ "eval_samples_per_second": 191.884,
607
+ "eval_steps_per_second": 6.002,
608
+ "step": 40000
609
+ },
610
+ {
611
+ "epoch": 1.39,
612
+ "learning_rate": 1.0715579710144927e-05,
613
+ "loss": 0.4885,
614
+ "step": 41000
615
+ },
616
+ {
617
+ "epoch": 1.39,
618
+ "eval_accuracy": 0.7529291900152827,
619
+ "eval_loss": 0.6188690066337585,
620
+ "eval_runtime": 52.5846,
621
+ "eval_samples_per_second": 186.651,
622
+ "eval_steps_per_second": 5.838,
623
+ "step": 41000
624
+ },
625
+ {
626
+ "epoch": 1.43,
627
+ "learning_rate": 1.0489130434782611e-05,
628
+ "loss": 0.4909,
629
+ "step": 42000
630
+ },
631
+ {
632
+ "epoch": 1.43,
633
+ "eval_accuracy": 0.7473255221599593,
634
+ "eval_loss": 0.6188654899597168,
635
+ "eval_runtime": 51.6594,
636
+ "eval_samples_per_second": 189.994,
637
+ "eval_steps_per_second": 5.943,
638
+ "step": 42000
639
+ },
640
+ {
641
+ "epoch": 1.46,
642
+ "learning_rate": 1.0262681159420292e-05,
643
+ "loss": 0.4894,
644
+ "step": 43000
645
+ },
646
+ {
647
+ "epoch": 1.46,
648
+ "eval_accuracy": 0.7432501273560876,
649
+ "eval_loss": 0.631429135799408,
650
+ "eval_runtime": 51.1492,
651
+ "eval_samples_per_second": 191.889,
652
+ "eval_steps_per_second": 6.002,
653
+ "step": 43000
654
+ },
655
+ {
656
+ "epoch": 1.49,
657
+ "learning_rate": 1.0036231884057972e-05,
658
+ "loss": 0.4912,
659
+ "step": 44000
660
+ },
661
+ {
662
+ "epoch": 1.49,
663
+ "eval_accuracy": 0.7445746306673459,
664
+ "eval_loss": 0.6183902025222778,
665
+ "eval_runtime": 51.0914,
666
+ "eval_samples_per_second": 192.107,
667
+ "eval_steps_per_second": 6.009,
668
+ "step": 44000
669
+ },
670
+ {
671
+ "epoch": 1.53,
672
+ "learning_rate": 9.809782608695653e-06,
673
+ "loss": 0.4851,
674
+ "step": 45000
675
+ },
676
+ {
677
+ "epoch": 1.53,
678
+ "eval_accuracy": 0.7461029037187977,
679
+ "eval_loss": 0.6257576942443848,
680
+ "eval_runtime": 50.9837,
681
+ "eval_samples_per_second": 192.513,
682
+ "eval_steps_per_second": 6.022,
683
+ "step": 45000
684
+ },
685
+ {
686
+ "epoch": 1.56,
687
+ "learning_rate": 9.583333333333335e-06,
688
+ "loss": 0.4879,
689
+ "step": 46000
690
+ },
691
+ {
692
+ "epoch": 1.56,
693
+ "eval_accuracy": 0.7480387162506368,
694
+ "eval_loss": 0.6286013126373291,
695
+ "eval_runtime": 51.1172,
696
+ "eval_samples_per_second": 192.01,
697
+ "eval_steps_per_second": 6.006,
698
+ "step": 46000
699
+ },
700
+ {
701
+ "epoch": 1.6,
702
+ "learning_rate": 9.356884057971016e-06,
703
+ "loss": 0.4907,
704
+ "step": 47000
705
+ },
706
+ {
707
+ "epoch": 1.6,
708
+ "eval_accuracy": 0.7511971472236373,
709
+ "eval_loss": 0.6196326613426208,
710
+ "eval_runtime": 51.072,
711
+ "eval_samples_per_second": 192.18,
712
+ "eval_steps_per_second": 6.011,
713
+ "step": 47000
714
+ },
715
+ {
716
+ "epoch": 1.63,
717
+ "learning_rate": 9.130434782608697e-06,
718
+ "loss": 0.4884,
719
+ "step": 48000
720
+ },
721
+ {
722
+ "epoch": 1.63,
723
+ "eval_accuracy": 0.7526235354049924,
724
+ "eval_loss": 0.6156549453735352,
725
+ "eval_runtime": 51.1757,
726
+ "eval_samples_per_second": 191.79,
727
+ "eval_steps_per_second": 5.999,
728
+ "step": 48000
729
+ },
730
+ {
731
+ "epoch": 1.66,
732
+ "learning_rate": 8.903985507246377e-06,
733
+ "loss": 0.4755,
734
+ "step": 49000
735
+ },
736
+ {
737
+ "epoch": 1.66,
738
+ "eval_accuracy": 0.7591441670911869,
739
+ "eval_loss": 0.6055976152420044,
740
+ "eval_runtime": 50.9202,
741
+ "eval_samples_per_second": 192.752,
742
+ "eval_steps_per_second": 6.029,
743
+ "step": 49000
744
+ },
745
+ {
746
+ "epoch": 1.7,
747
+ "learning_rate": 8.677536231884058e-06,
748
+ "loss": 0.4811,
749
+ "step": 50000
750
+ },
751
+ {
752
+ "epoch": 1.7,
753
+ "eval_accuracy": 0.7582272032603159,
754
+ "eval_loss": 0.5976621508598328,
755
+ "eval_runtime": 51.1974,
756
+ "eval_samples_per_second": 191.709,
757
+ "eval_steps_per_second": 5.996,
758
+ "step": 50000
759
+ },
760
+ {
761
+ "epoch": 1.73,
762
+ "learning_rate": 8.45108695652174e-06,
763
+ "loss": 0.4787,
764
+ "step": 51000
765
+ },
766
+ {
767
+ "epoch": 1.73,
768
+ "eval_accuracy": 0.7620988283239939,
769
+ "eval_loss": 0.5914710164070129,
770
+ "eval_runtime": 52.4733,
771
+ "eval_samples_per_second": 187.047,
772
+ "eval_steps_per_second": 5.851,
773
+ "step": 51000
774
+ },
775
+ {
776
+ "epoch": 1.77,
777
+ "learning_rate": 8.22463768115942e-06,
778
+ "loss": 0.4779,
779
+ "step": 52000
780
+ },
781
+ {
782
+ "epoch": 1.77,
783
+ "eval_accuracy": 0.7583290881304127,
784
+ "eval_loss": 0.6014041304588318,
785
+ "eval_runtime": 51.6886,
786
+ "eval_samples_per_second": 189.887,
787
+ "eval_steps_per_second": 5.939,
788
+ "step": 52000
789
+ },
790
+ {
791
+ "epoch": 1.8,
792
+ "learning_rate": 7.998188405797103e-06,
793
+ "loss": 0.4767,
794
+ "step": 53000
795
+ },
796
+ {
797
+ "epoch": 1.8,
798
+ "eval_accuracy": 0.7623025980641874,
799
+ "eval_loss": 0.6041266918182373,
800
+ "eval_runtime": 51.0745,
801
+ "eval_samples_per_second": 192.17,
802
+ "eval_steps_per_second": 6.011,
803
+ "step": 53000
804
+ },
805
+ {
806
+ "epoch": 1.83,
807
+ "learning_rate": 7.771739130434784e-06,
808
+ "loss": 0.4737,
809
+ "step": 54000
810
+ },
811
+ {
812
+ "epoch": 1.83,
813
+ "eval_accuracy": 0.7562913907284768,
814
+ "eval_loss": 0.6093412637710571,
815
+ "eval_runtime": 51.0594,
816
+ "eval_samples_per_second": 192.227,
817
+ "eval_steps_per_second": 6.013,
818
+ "step": 54000
819
+ },
820
+ {
821
+ "epoch": 1.87,
822
+ "learning_rate": 7.545289855072464e-06,
823
+ "loss": 0.4836,
824
+ "step": 55000
825
+ },
826
+ {
827
+ "epoch": 1.87,
828
+ "eval_accuracy": 0.7568008150789608,
829
+ "eval_loss": 0.6001136302947998,
830
+ "eval_runtime": 51.1468,
831
+ "eval_samples_per_second": 191.899,
832
+ "eval_steps_per_second": 6.002,
833
+ "step": 55000
834
+ },
835
+ {
836
+ "epoch": 1.9,
837
+ "learning_rate": 7.318840579710146e-06,
838
+ "loss": 0.4765,
839
+ "step": 56000
840
+ },
841
+ {
842
+ "epoch": 1.9,
843
+ "eval_accuracy": 0.7600611309220581,
844
+ "eval_loss": 0.6109462380409241,
845
+ "eval_runtime": 51.0882,
846
+ "eval_samples_per_second": 192.119,
847
+ "eval_steps_per_second": 6.009,
848
+ "step": 56000
849
+ },
850
+ {
851
+ "epoch": 1.94,
852
+ "learning_rate": 7.092391304347826e-06,
853
+ "loss": 0.4776,
854
+ "step": 57000
855
+ },
856
+ {
857
+ "epoch": 1.94,
858
+ "eval_accuracy": 0.7598573611818645,
859
+ "eval_loss": 0.6045997142791748,
860
+ "eval_runtime": 51.0831,
861
+ "eval_samples_per_second": 192.138,
862
+ "eval_steps_per_second": 6.01,
863
+ "step": 57000
864
+ },
865
+ {
866
+ "epoch": 1.97,
867
+ "learning_rate": 6.865942028985509e-06,
868
+ "loss": 0.4769,
869
+ "step": 58000
870
+ },
871
+ {
872
+ "epoch": 1.97,
873
+ "eval_accuracy": 0.7568008150789608,
874
+ "eval_loss": 0.5969610214233398,
875
+ "eval_runtime": 51.092,
876
+ "eval_samples_per_second": 192.105,
877
+ "eval_steps_per_second": 6.009,
878
+ "step": 58000
879
+ },
880
+ {
881
+ "epoch": 2.0,
882
+ "learning_rate": 6.639492753623189e-06,
883
+ "loss": 0.4654,
884
+ "step": 59000
885
+ },
886
+ {
887
+ "epoch": 2.0,
888
+ "eval_accuracy": 0.7613856342333164,
889
+ "eval_loss": 0.6146702170372009,
890
+ "eval_runtime": 51.0794,
891
+ "eval_samples_per_second": 192.152,
892
+ "eval_steps_per_second": 6.01,
893
+ "step": 59000
894
+ },
895
+ {
896
+ "epoch": 2.04,
897
+ "learning_rate": 6.41304347826087e-06,
898
+ "loss": 0.4144,
899
+ "step": 60000
900
+ },
901
+ {
902
+ "epoch": 2.04,
903
+ "eval_accuracy": 0.7565970453387671,
904
+ "eval_loss": 0.6438983678817749,
905
+ "eval_runtime": 51.1073,
906
+ "eval_samples_per_second": 192.047,
907
+ "eval_steps_per_second": 6.007,
908
+ "step": 60000
909
+ },
910
+ {
911
+ "epoch": 2.07,
912
+ "learning_rate": 6.186594202898551e-06,
913
+ "loss": 0.4101,
914
+ "step": 61000
915
+ },
916
+ {
917
+ "epoch": 2.07,
918
+ "eval_accuracy": 0.7527254202750892,
919
+ "eval_loss": 0.637277364730835,
920
+ "eval_runtime": 52.7247,
921
+ "eval_samples_per_second": 186.156,
922
+ "eval_steps_per_second": 5.823,
923
+ "step": 61000
924
+ },
925
+ {
926
+ "epoch": 2.11,
927
+ "learning_rate": 5.960144927536232e-06,
928
+ "loss": 0.4192,
929
+ "step": 62000
930
+ },
931
+ {
932
+ "epoch": 2.11,
933
+ "eval_accuracy": 0.7575140091696383,
934
+ "eval_loss": 0.6135603189468384,
935
+ "eval_runtime": 52.6526,
936
+ "eval_samples_per_second": 186.411,
937
+ "eval_steps_per_second": 5.831,
938
+ "step": 62000
939
+ },
940
+ {
941
+ "epoch": 2.14,
942
+ "learning_rate": 5.733695652173914e-06,
943
+ "loss": 0.4128,
944
+ "step": 63000
945
+ },
946
+ {
947
+ "epoch": 2.14,
948
+ "eval_accuracy": 0.7559857361181864,
949
+ "eval_loss": 0.6282700896263123,
950
+ "eval_runtime": 51.1146,
951
+ "eval_samples_per_second": 192.02,
952
+ "eval_steps_per_second": 6.006,
953
+ "step": 63000
954
+ },
955
+ {
956
+ "epoch": 2.17,
957
+ "learning_rate": 5.507246376811595e-06,
958
+ "loss": 0.4204,
959
+ "step": 64000
960
+ },
961
+ {
962
+ "epoch": 2.17,
963
+ "eval_accuracy": 0.7625063678043811,
964
+ "eval_loss": 0.618690013885498,
965
+ "eval_runtime": 51.1974,
966
+ "eval_samples_per_second": 191.709,
967
+ "eval_steps_per_second": 5.996,
968
+ "step": 64000
969
+ },
970
+ {
971
+ "epoch": 2.21,
972
+ "learning_rate": 5.2807971014492755e-06,
973
+ "loss": 0.4114,
974
+ "step": 65000
975
+ },
976
+ {
977
+ "epoch": 2.21,
978
+ "eval_accuracy": 0.7620988283239939,
979
+ "eval_loss": 0.6127009987831116,
980
+ "eval_runtime": 51.1304,
981
+ "eval_samples_per_second": 191.96,
982
+ "eval_steps_per_second": 6.004,
983
+ "step": 65000
984
+ },
985
+ {
986
+ "epoch": 2.24,
987
+ "learning_rate": 5.054347826086957e-06,
988
+ "loss": 0.4097,
989
+ "step": 66000
990
+ },
991
+ {
992
+ "epoch": 2.24,
993
+ "eval_accuracy": 0.7626082526744778,
994
+ "eval_loss": 0.618818998336792,
995
+ "eval_runtime": 51.1752,
996
+ "eval_samples_per_second": 191.792,
997
+ "eval_steps_per_second": 5.999,
998
+ "step": 66000
999
+ },
1000
+ {
1001
+ "epoch": 2.28,
1002
+ "learning_rate": 4.8278985507246375e-06,
1003
+ "loss": 0.4129,
1004
+ "step": 67000
1005
+ },
1006
+ {
1007
+ "epoch": 2.28,
1008
+ "eval_accuracy": 0.7639327559857361,
1009
+ "eval_loss": 0.6156466603279114,
1010
+ "eval_runtime": 51.1634,
1011
+ "eval_samples_per_second": 191.836,
1012
+ "eval_steps_per_second": 6.0,
1013
+ "step": 67000
1014
+ },
1015
+ {
1016
+ "epoch": 2.31,
1017
+ "learning_rate": 4.601449275362319e-06,
1018
+ "loss": 0.4085,
1019
+ "step": 68000
1020
+ },
1021
+ {
1022
+ "epoch": 2.31,
1023
+ "eval_accuracy": 0.7615894039735099,
1024
+ "eval_loss": 0.6232104301452637,
1025
+ "eval_runtime": 51.1826,
1026
+ "eval_samples_per_second": 191.764,
1027
+ "eval_steps_per_second": 5.998,
1028
+ "step": 68000
1029
+ },
1030
+ {
1031
+ "epoch": 2.34,
1032
+ "learning_rate": 4.3750000000000005e-06,
1033
+ "loss": 0.4074,
1034
+ "step": 69000
1035
+ },
1036
+ {
1037
+ "epoch": 2.34,
1038
+ "eval_accuracy": 0.7604686704024453,
1039
+ "eval_loss": 0.6239916682243347,
1040
+ "eval_runtime": 51.0688,
1041
+ "eval_samples_per_second": 192.192,
1042
+ "eval_steps_per_second": 6.011,
1043
+ "step": 69000
1044
+ },
1045
+ {
1046
+ "epoch": 2.38,
1047
+ "learning_rate": 4.148550724637682e-06,
1048
+ "loss": 0.409,
1049
+ "step": 70000
1050
+ },
1051
+ {
1052
+ "epoch": 2.38,
1053
+ "eval_accuracy": 0.7591441670911869,
1054
+ "eval_loss": 0.6152721643447876,
1055
+ "eval_runtime": 51.1662,
1056
+ "eval_samples_per_second": 191.826,
1057
+ "eval_steps_per_second": 6.0,
1058
+ "step": 70000
1059
+ },
1060
+ {
1061
+ "epoch": 2.41,
1062
+ "learning_rate": 3.9221014492753625e-06,
1063
+ "loss": 0.4046,
1064
+ "step": 71000
1065
+ },
1066
+ {
1067
+ "epoch": 2.41,
1068
+ "eval_accuracy": 0.7587366276107997,
1069
+ "eval_loss": 0.6375284194946289,
1070
+ "eval_runtime": 52.7067,
1071
+ "eval_samples_per_second": 186.219,
1072
+ "eval_steps_per_second": 5.825,
1073
+ "step": 71000
1074
+ },
1075
+ {
1076
+ "epoch": 2.45,
1077
+ "learning_rate": 3.6956521739130436e-06,
1078
+ "loss": 0.4117,
1079
+ "step": 72000
1080
+ },
1081
+ {
1082
+ "epoch": 2.45,
1083
+ "eval_accuracy": 0.7629139072847683,
1084
+ "eval_loss": 0.6144647598266602,
1085
+ "eval_runtime": 52.5798,
1086
+ "eval_samples_per_second": 186.669,
1087
+ "eval_steps_per_second": 5.839,
1088
+ "step": 72000
1089
+ },
1090
+ {
1091
+ "epoch": 2.48,
1092
+ "learning_rate": 3.4692028985507246e-06,
1093
+ "loss": 0.4002,
1094
+ "step": 73000
1095
+ },
1096
+ {
1097
+ "epoch": 2.48,
1098
+ "eval_accuracy": 0.7609780947529292,
1099
+ "eval_loss": 0.6278636455535889,
1100
+ "eval_runtime": 51.1134,
1101
+ "eval_samples_per_second": 192.024,
1102
+ "eval_steps_per_second": 6.006,
1103
+ "step": 73000
1104
+ },
1105
+ {
1106
+ "epoch": 2.51,
1107
+ "learning_rate": 3.242753623188406e-06,
1108
+ "loss": 0.4042,
1109
+ "step": 74000
1110
+ },
1111
+ {
1112
+ "epoch": 2.51,
1113
+ "eval_accuracy": 0.7646459500764137,
1114
+ "eval_loss": 0.6176004409790039,
1115
+ "eval_runtime": 51.1309,
1116
+ "eval_samples_per_second": 191.958,
1117
+ "eval_steps_per_second": 6.004,
1118
+ "step": 74000
1119
+ },
1120
+ {
1121
+ "epoch": 2.55,
1122
+ "learning_rate": 3.016304347826087e-06,
1123
+ "loss": 0.4055,
1124
+ "step": 75000
1125
+ },
1126
+ {
1127
+ "epoch": 2.55,
1128
+ "eval_accuracy": 0.7643402954661233,
1129
+ "eval_loss": 0.627702534198761,
1130
+ "eval_runtime": 51.1017,
1131
+ "eval_samples_per_second": 192.068,
1132
+ "eval_steps_per_second": 6.008,
1133
+ "step": 75000
1134
+ },
1135
+ {
1136
+ "epoch": 2.58,
1137
+ "learning_rate": 2.7898550724637686e-06,
1138
+ "loss": 0.4021,
1139
+ "step": 76000
1140
+ },
1141
+ {
1142
+ "epoch": 2.58,
1143
+ "eval_accuracy": 0.7642384105960265,
1144
+ "eval_loss": 0.619607150554657,
1145
+ "eval_runtime": 51.1091,
1146
+ "eval_samples_per_second": 192.04,
1147
+ "eval_steps_per_second": 6.007,
1148
+ "step": 76000
1149
+ },
1150
+ {
1151
+ "epoch": 2.62,
1152
+ "learning_rate": 2.563405797101449e-06,
1153
+ "loss": 0.4081,
1154
+ "step": 77000
1155
+ },
1156
+ {
1157
+ "epoch": 2.62,
1158
+ "eval_accuracy": 0.7658685685175751,
1159
+ "eval_loss": 0.6127172708511353,
1160
+ "eval_runtime": 51.0673,
1161
+ "eval_samples_per_second": 192.197,
1162
+ "eval_steps_per_second": 6.012,
1163
+ "step": 77000
1164
+ },
1165
+ {
1166
+ "epoch": 2.65,
1167
+ "learning_rate": 2.3369565217391307e-06,
1168
+ "loss": 0.408,
1169
+ "step": 78000
1170
+ },
1171
+ {
1172
+ "epoch": 2.65,
1173
+ "eval_accuracy": 0.7638308711156393,
1174
+ "eval_loss": 0.6236501336097717,
1175
+ "eval_runtime": 51.0905,
1176
+ "eval_samples_per_second": 192.11,
1177
+ "eval_steps_per_second": 6.009,
1178
+ "step": 78000
1179
+ },
1180
+ {
1181
+ "epoch": 2.68,
1182
+ "learning_rate": 2.1105072463768117e-06,
1183
+ "loss": 0.3997,
1184
+ "step": 79000
1185
+ },
1186
+ {
1187
+ "epoch": 2.68,
1188
+ "eval_accuracy": 0.7636271013754458,
1189
+ "eval_loss": 0.6190339922904968,
1190
+ "eval_runtime": 51.097,
1191
+ "eval_samples_per_second": 192.086,
1192
+ "eval_steps_per_second": 6.008,
1193
+ "step": 79000
1194
+ },
1195
+ {
1196
+ "epoch": 2.72,
1197
+ "learning_rate": 1.884057971014493e-06,
1198
+ "loss": 0.4093,
1199
+ "step": 80000
1200
+ },
1201
+ {
1202
+ "epoch": 2.72,
1203
+ "eval_accuracy": 0.7648497198166072,
1204
+ "eval_loss": 0.615200936794281,
1205
+ "eval_runtime": 51.0806,
1206
+ "eval_samples_per_second": 192.147,
1207
+ "eval_steps_per_second": 6.01,
1208
+ "step": 80000
1209
+ },
1210
+ {
1211
+ "epoch": 2.75,
1212
+ "learning_rate": 1.657608695652174e-06,
1213
+ "loss": 0.4095,
1214
+ "step": 81000
1215
+ },
1216
+ {
1217
+ "epoch": 2.75,
1218
+ "eval_accuracy": 0.7627101375445746,
1219
+ "eval_loss": 0.6154515743255615,
1220
+ "eval_runtime": 52.4793,
1221
+ "eval_samples_per_second": 187.026,
1222
+ "eval_steps_per_second": 5.85,
1223
+ "step": 81000
1224
+ },
1225
+ {
1226
+ "epoch": 2.79,
1227
+ "learning_rate": 1.4311594202898552e-06,
1228
+ "loss": 0.4088,
1229
+ "step": 82000
1230
+ },
1231
+ {
1232
+ "epoch": 2.79,
1233
+ "eval_accuracy": 0.7641365257259297,
1234
+ "eval_loss": 0.6130374073982239,
1235
+ "eval_runtime": 52.6234,
1236
+ "eval_samples_per_second": 186.514,
1237
+ "eval_steps_per_second": 5.834,
1238
+ "step": 82000
1239
+ },
1240
+ {
1241
+ "epoch": 2.82,
1242
+ "learning_rate": 1.2047101449275363e-06,
1243
+ "loss": 0.4063,
1244
+ "step": 83000
1245
+ },
1246
+ {
1247
+ "epoch": 2.82,
1248
+ "eval_accuracy": 0.7646459500764137,
1249
+ "eval_loss": 0.6072085499763489,
1250
+ "eval_runtime": 51.0506,
1251
+ "eval_samples_per_second": 192.26,
1252
+ "eval_steps_per_second": 6.014,
1253
+ "step": 83000
1254
+ },
1255
+ {
1256
+ "epoch": 2.85,
1257
+ "learning_rate": 9.782608695652175e-07,
1258
+ "loss": 0.3978,
1259
+ "step": 84000
1260
+ },
1261
+ {
1262
+ "epoch": 2.85,
1263
+ "eval_accuracy": 0.7661742231278655,
1264
+ "eval_loss": 0.6128284931182861,
1265
+ "eval_runtime": 51.2162,
1266
+ "eval_samples_per_second": 191.639,
1267
+ "eval_steps_per_second": 5.994,
1268
+ "step": 84000
1269
+ },
1270
+ {
1271
+ "epoch": 2.89,
1272
+ "learning_rate": 7.518115942028987e-07,
1273
+ "loss": 0.4034,
1274
+ "step": 85000
1275
+ },
1276
+ {
1277
+ "epoch": 2.89,
1278
+ "eval_accuracy": 0.7627101375445746,
1279
+ "eval_loss": 0.6156770586967468,
1280
+ "eval_runtime": 51.1115,
1281
+ "eval_samples_per_second": 192.031,
1282
+ "eval_steps_per_second": 6.006,
1283
+ "step": 85000
1284
+ },
1285
+ {
1286
+ "epoch": 2.92,
1287
+ "learning_rate": 5.253623188405797e-07,
1288
+ "loss": 0.4044,
1289
+ "step": 86000
1290
+ },
1291
+ {
1292
+ "epoch": 2.92,
1293
+ "eval_accuracy": 0.7660723382577688,
1294
+ "eval_loss": 0.6127380728721619,
1295
+ "eval_runtime": 51.0045,
1296
+ "eval_samples_per_second": 192.434,
1297
+ "eval_steps_per_second": 6.019,
1298
+ "step": 86000
1299
+ },
1300
+ {
1301
+ "epoch": 2.96,
1302
+ "learning_rate": 2.989130434782609e-07,
1303
+ "loss": 0.403,
1304
+ "step": 87000
1305
+ },
1306
+ {
1307
+ "epoch": 2.96,
1308
+ "eval_accuracy": 0.7663779928680591,
1309
+ "eval_loss": 0.612598717212677,
1310
+ "eval_runtime": 51.1269,
1311
+ "eval_samples_per_second": 191.973,
1312
+ "eval_steps_per_second": 6.005,
1313
+ "step": 87000
1314
+ },
1315
+ {
1316
+ "epoch": 2.99,
1317
+ "learning_rate": 7.246376811594204e-08,
1318
+ "loss": 0.4033,
1319
+ "step": 88000
1320
+ },
1321
+ {
1322
+ "epoch": 2.99,
1323
+ "eval_accuracy": 0.7661742231278655,
1324
+ "eval_loss": 0.6143723726272583,
1325
+ "eval_runtime": 50.9677,
1326
+ "eval_samples_per_second": 192.573,
1327
+ "eval_steps_per_second": 6.023,
1328
+ "step": 88000
1329
+ },
1330
+ {
1331
+ "epoch": 3.0,
1332
+ "step": 88320,
1333
+ "total_flos": 1.4721628163172653e+17,
1334
+ "train_loss": 0.5041872973027437,
1335
+ "train_runtime": 53589.3429,
1336
+ "train_samples_per_second": 52.738,
1337
+ "train_steps_per_second": 1.648
1338
+ }
1339
+ ],
1340
+ "max_steps": 88320,
1341
+ "num_train_epochs": 3,
1342
+ "total_flos": 1.4721628163172653e+17,
1343
+ "trial_name": null,
1344
+ "trial_params": null
1345
+ }