Text Generation
Transformers
PyTorch
llama
conversational
text-generation-inference
yapeichang commited on
Commit
684b289
·
verified ·
1 Parent(s): 3972deb

Upload trainer_state.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. trainer_state.json +1438 -0
trainer_state.json ADDED
@@ -0,0 +1,1438 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9984,
6
+ "eval_steps": 500,
7
+ "global_step": 156,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0064,
14
+ "grad_norm": 4.46052360534668,
15
+ "learning_rate": 0.0,
16
+ "loss": 1.1926,
17
+ "mean_token_accuracy": 0.7001730054616928,
18
+ "num_tokens": 14640.0,
19
+ "step": 1
20
+ },
21
+ {
22
+ "epoch": 0.0128,
23
+ "grad_norm": 3.8626222610473633,
24
+ "learning_rate": 6.25e-07,
25
+ "loss": 1.3138,
26
+ "mean_token_accuracy": 0.6548218131065369,
27
+ "num_tokens": 27693.0,
28
+ "step": 2
29
+ },
30
+ {
31
+ "epoch": 0.0192,
32
+ "grad_norm": 3.952249050140381,
33
+ "learning_rate": 1.25e-06,
34
+ "loss": 1.3569,
35
+ "mean_token_accuracy": 0.6562561392784119,
36
+ "num_tokens": 40525.0,
37
+ "step": 3
38
+ },
39
+ {
40
+ "epoch": 0.0256,
41
+ "grad_norm": 4.109567165374756,
42
+ "learning_rate": 1.8750000000000003e-06,
43
+ "loss": 1.3472,
44
+ "mean_token_accuracy": 0.656329445540905,
45
+ "num_tokens": 55822.0,
46
+ "step": 4
47
+ },
48
+ {
49
+ "epoch": 0.032,
50
+ "grad_norm": 4.119779586791992,
51
+ "learning_rate": 2.5e-06,
52
+ "loss": 1.3473,
53
+ "mean_token_accuracy": 0.6558413058519363,
54
+ "num_tokens": 70670.0,
55
+ "step": 5
56
+ },
57
+ {
58
+ "epoch": 0.0384,
59
+ "grad_norm": 3.3819665908813477,
60
+ "learning_rate": 3.125e-06,
61
+ "loss": 1.4108,
62
+ "mean_token_accuracy": 0.6545125097036362,
63
+ "num_tokens": 84335.0,
64
+ "step": 6
65
+ },
66
+ {
67
+ "epoch": 0.0448,
68
+ "grad_norm": 3.7010326385498047,
69
+ "learning_rate": 3.7500000000000005e-06,
70
+ "loss": 1.2498,
71
+ "mean_token_accuracy": 0.6681988388299942,
72
+ "num_tokens": 95857.0,
73
+ "step": 7
74
+ },
75
+ {
76
+ "epoch": 0.0512,
77
+ "grad_norm": 3.496901035308838,
78
+ "learning_rate": 4.3750000000000005e-06,
79
+ "loss": 1.2894,
80
+ "mean_token_accuracy": 0.6635143309831619,
81
+ "num_tokens": 109806.0,
82
+ "step": 8
83
+ },
84
+ {
85
+ "epoch": 0.0576,
86
+ "grad_norm": 3.2937991619110107,
87
+ "learning_rate": 5e-06,
88
+ "loss": 1.2092,
89
+ "mean_token_accuracy": 0.6767952516674995,
90
+ "num_tokens": 125631.0,
91
+ "step": 9
92
+ },
93
+ {
94
+ "epoch": 0.064,
95
+ "grad_norm": 6.788699626922607,
96
+ "learning_rate": 5e-06,
97
+ "loss": 1.4017,
98
+ "mean_token_accuracy": 0.6475312933325768,
99
+ "num_tokens": 140029.0,
100
+ "step": 10
101
+ },
102
+ {
103
+ "epoch": 0.0704,
104
+ "grad_norm": 8.16016674041748,
105
+ "learning_rate": 5e-06,
106
+ "loss": 1.4182,
107
+ "mean_token_accuracy": 0.6417016983032227,
108
+ "num_tokens": 153863.0,
109
+ "step": 11
110
+ },
111
+ {
112
+ "epoch": 0.0768,
113
+ "grad_norm": 7.313711643218994,
114
+ "learning_rate": 5e-06,
115
+ "loss": 1.2493,
116
+ "mean_token_accuracy": 0.6722374185919762,
117
+ "num_tokens": 169366.0,
118
+ "step": 12
119
+ },
120
+ {
121
+ "epoch": 0.0832,
122
+ "grad_norm": 7.297321319580078,
123
+ "learning_rate": 5e-06,
124
+ "loss": 1.2984,
125
+ "mean_token_accuracy": 0.6654682904481888,
126
+ "num_tokens": 182285.0,
127
+ "step": 13
128
+ },
129
+ {
130
+ "epoch": 0.0896,
131
+ "grad_norm": 5.980355262756348,
132
+ "learning_rate": 5e-06,
133
+ "loss": 1.3728,
134
+ "mean_token_accuracy": 0.6474195793271065,
135
+ "num_tokens": 200009.0,
136
+ "step": 14
137
+ },
138
+ {
139
+ "epoch": 0.096,
140
+ "grad_norm": 5.341892242431641,
141
+ "learning_rate": 5e-06,
142
+ "loss": 1.1786,
143
+ "mean_token_accuracy": 0.6898260414600372,
144
+ "num_tokens": 217172.0,
145
+ "step": 15
146
+ },
147
+ {
148
+ "epoch": 0.1024,
149
+ "grad_norm": 5.044528961181641,
150
+ "learning_rate": 5e-06,
151
+ "loss": 1.473,
152
+ "mean_token_accuracy": 0.6259094700217247,
153
+ "num_tokens": 232308.0,
154
+ "step": 16
155
+ },
156
+ {
157
+ "epoch": 0.1088,
158
+ "grad_norm": 4.829463005065918,
159
+ "learning_rate": 5e-06,
160
+ "loss": 1.4725,
161
+ "mean_token_accuracy": 0.6119906902313232,
162
+ "num_tokens": 243772.0,
163
+ "step": 17
164
+ },
165
+ {
166
+ "epoch": 0.1152,
167
+ "grad_norm": 3.969775676727295,
168
+ "learning_rate": 5e-06,
169
+ "loss": 1.2455,
170
+ "mean_token_accuracy": 0.6761033684015274,
171
+ "num_tokens": 260674.0,
172
+ "step": 18
173
+ },
174
+ {
175
+ "epoch": 0.1216,
176
+ "grad_norm": 4.059275150299072,
177
+ "learning_rate": 5e-06,
178
+ "loss": 1.3388,
179
+ "mean_token_accuracy": 0.6495725736021996,
180
+ "num_tokens": 276352.0,
181
+ "step": 19
182
+ },
183
+ {
184
+ "epoch": 0.128,
185
+ "grad_norm": 3.6464030742645264,
186
+ "learning_rate": 5e-06,
187
+ "loss": 1.277,
188
+ "mean_token_accuracy": 0.6567010655999184,
189
+ "num_tokens": 291765.0,
190
+ "step": 20
191
+ },
192
+ {
193
+ "epoch": 0.1344,
194
+ "grad_norm": 4.572549819946289,
195
+ "learning_rate": 5e-06,
196
+ "loss": 1.349,
197
+ "mean_token_accuracy": 0.6503036767244339,
198
+ "num_tokens": 303401.0,
199
+ "step": 21
200
+ },
201
+ {
202
+ "epoch": 0.1408,
203
+ "grad_norm": 3.3478147983551025,
204
+ "learning_rate": 5e-06,
205
+ "loss": 1.3734,
206
+ "mean_token_accuracy": 0.657492570579052,
207
+ "num_tokens": 318861.0,
208
+ "step": 22
209
+ },
210
+ {
211
+ "epoch": 0.1472,
212
+ "grad_norm": 3.8337855339050293,
213
+ "learning_rate": 5e-06,
214
+ "loss": 1.2852,
215
+ "mean_token_accuracy": 0.6548745334148407,
216
+ "num_tokens": 330566.0,
217
+ "step": 23
218
+ },
219
+ {
220
+ "epoch": 0.1536,
221
+ "grad_norm": 4.150561809539795,
222
+ "learning_rate": 5e-06,
223
+ "loss": 1.2668,
224
+ "mean_token_accuracy": 0.6780209168791771,
225
+ "num_tokens": 341999.0,
226
+ "step": 24
227
+ },
228
+ {
229
+ "epoch": 0.16,
230
+ "grad_norm": 5.090829849243164,
231
+ "learning_rate": 5e-06,
232
+ "loss": 1.3751,
233
+ "mean_token_accuracy": 0.6607428789138794,
234
+ "num_tokens": 358011.0,
235
+ "step": 25
236
+ },
237
+ {
238
+ "epoch": 0.1664,
239
+ "grad_norm": 4.033487319946289,
240
+ "learning_rate": 5e-06,
241
+ "loss": 1.2884,
242
+ "mean_token_accuracy": 0.6732476428151131,
243
+ "num_tokens": 371643.0,
244
+ "step": 26
245
+ },
246
+ {
247
+ "epoch": 0.1728,
248
+ "grad_norm": 3.5941057205200195,
249
+ "learning_rate": 5e-06,
250
+ "loss": 1.047,
251
+ "mean_token_accuracy": 0.7033858001232147,
252
+ "num_tokens": 383610.0,
253
+ "step": 27
254
+ },
255
+ {
256
+ "epoch": 0.1792,
257
+ "grad_norm": 3.3021464347839355,
258
+ "learning_rate": 5e-06,
259
+ "loss": 1.233,
260
+ "mean_token_accuracy": 0.6834521070122719,
261
+ "num_tokens": 398590.0,
262
+ "step": 28
263
+ },
264
+ {
265
+ "epoch": 0.1856,
266
+ "grad_norm": 4.259124279022217,
267
+ "learning_rate": 5e-06,
268
+ "loss": 1.2961,
269
+ "mean_token_accuracy": 0.6622931659221649,
270
+ "num_tokens": 409575.0,
271
+ "step": 29
272
+ },
273
+ {
274
+ "epoch": 0.192,
275
+ "grad_norm": 3.220076322555542,
276
+ "learning_rate": 5e-06,
277
+ "loss": 1.0836,
278
+ "mean_token_accuracy": 0.7075582221150398,
279
+ "num_tokens": 423461.0,
280
+ "step": 30
281
+ },
282
+ {
283
+ "epoch": 0.1984,
284
+ "grad_norm": 3.3759732246398926,
285
+ "learning_rate": 5e-06,
286
+ "loss": 1.2671,
287
+ "mean_token_accuracy": 0.6733162626624107,
288
+ "num_tokens": 439184.0,
289
+ "step": 31
290
+ },
291
+ {
292
+ "epoch": 0.2048,
293
+ "grad_norm": 13.060998916625977,
294
+ "learning_rate": 5e-06,
295
+ "loss": 1.263,
296
+ "mean_token_accuracy": 0.6846785917878151,
297
+ "num_tokens": 451887.0,
298
+ "step": 32
299
+ },
300
+ {
301
+ "epoch": 0.2112,
302
+ "grad_norm": 3.4759275913238525,
303
+ "learning_rate": 5e-06,
304
+ "loss": 1.2905,
305
+ "mean_token_accuracy": 0.6659640669822693,
306
+ "num_tokens": 466293.0,
307
+ "step": 33
308
+ },
309
+ {
310
+ "epoch": 0.2176,
311
+ "grad_norm": 3.615352153778076,
312
+ "learning_rate": 5e-06,
313
+ "loss": 1.3252,
314
+ "mean_token_accuracy": 0.6587846502661705,
315
+ "num_tokens": 482281.0,
316
+ "step": 34
317
+ },
318
+ {
319
+ "epoch": 0.224,
320
+ "grad_norm": 3.8253462314605713,
321
+ "learning_rate": 5e-06,
322
+ "loss": 1.284,
323
+ "mean_token_accuracy": 0.6593102514743805,
324
+ "num_tokens": 494866.0,
325
+ "step": 35
326
+ },
327
+ {
328
+ "epoch": 0.2304,
329
+ "grad_norm": 3.8242146968841553,
330
+ "learning_rate": 5e-06,
331
+ "loss": 1.4445,
332
+ "mean_token_accuracy": 0.6356016099452972,
333
+ "num_tokens": 509737.0,
334
+ "step": 36
335
+ },
336
+ {
337
+ "epoch": 0.2368,
338
+ "grad_norm": 3.3142733573913574,
339
+ "learning_rate": 5e-06,
340
+ "loss": 1.3622,
341
+ "mean_token_accuracy": 0.6442921012639999,
342
+ "num_tokens": 524987.0,
343
+ "step": 37
344
+ },
345
+ {
346
+ "epoch": 0.2432,
347
+ "grad_norm": 3.3738458156585693,
348
+ "learning_rate": 5e-06,
349
+ "loss": 1.5074,
350
+ "mean_token_accuracy": 0.611025758087635,
351
+ "num_tokens": 540515.0,
352
+ "step": 38
353
+ },
354
+ {
355
+ "epoch": 0.2496,
356
+ "grad_norm": 3.5873780250549316,
357
+ "learning_rate": 5e-06,
358
+ "loss": 1.2594,
359
+ "mean_token_accuracy": 0.672322042286396,
360
+ "num_tokens": 554108.0,
361
+ "step": 39
362
+ },
363
+ {
364
+ "epoch": 0.256,
365
+ "grad_norm": 3.3734376430511475,
366
+ "learning_rate": 5e-06,
367
+ "loss": 1.386,
368
+ "mean_token_accuracy": 0.6393688693642616,
369
+ "num_tokens": 570067.0,
370
+ "step": 40
371
+ },
372
+ {
373
+ "epoch": 0.2624,
374
+ "grad_norm": 3.8531699180603027,
375
+ "learning_rate": 5e-06,
376
+ "loss": 1.2987,
377
+ "mean_token_accuracy": 0.6596419736742973,
378
+ "num_tokens": 583435.0,
379
+ "step": 41
380
+ },
381
+ {
382
+ "epoch": 0.2688,
383
+ "grad_norm": 3.7436106204986572,
384
+ "learning_rate": 5e-06,
385
+ "loss": 1.3517,
386
+ "mean_token_accuracy": 0.6706470102071762,
387
+ "num_tokens": 596105.0,
388
+ "step": 42
389
+ },
390
+ {
391
+ "epoch": 0.2752,
392
+ "grad_norm": 3.717897653579712,
393
+ "learning_rate": 5e-06,
394
+ "loss": 1.147,
395
+ "mean_token_accuracy": 0.7020372673869133,
396
+ "num_tokens": 610310.0,
397
+ "step": 43
398
+ },
399
+ {
400
+ "epoch": 0.2816,
401
+ "grad_norm": 3.6992413997650146,
402
+ "learning_rate": 5e-06,
403
+ "loss": 1.19,
404
+ "mean_token_accuracy": 0.6930102109909058,
405
+ "num_tokens": 624612.0,
406
+ "step": 44
407
+ },
408
+ {
409
+ "epoch": 0.288,
410
+ "grad_norm": 3.5740365982055664,
411
+ "learning_rate": 5e-06,
412
+ "loss": 1.368,
413
+ "mean_token_accuracy": 0.6361823007464409,
414
+ "num_tokens": 637714.0,
415
+ "step": 45
416
+ },
417
+ {
418
+ "epoch": 0.2944,
419
+ "grad_norm": 3.959993600845337,
420
+ "learning_rate": 5e-06,
421
+ "loss": 1.2512,
422
+ "mean_token_accuracy": 0.6959773004055023,
423
+ "num_tokens": 650333.0,
424
+ "step": 46
425
+ },
426
+ {
427
+ "epoch": 0.3008,
428
+ "grad_norm": 3.906562089920044,
429
+ "learning_rate": 5e-06,
430
+ "loss": 1.2867,
431
+ "mean_token_accuracy": 0.6637914702296257,
432
+ "num_tokens": 664887.0,
433
+ "step": 47
434
+ },
435
+ {
436
+ "epoch": 0.3072,
437
+ "grad_norm": 3.9186768531799316,
438
+ "learning_rate": 5e-06,
439
+ "loss": 1.3578,
440
+ "mean_token_accuracy": 0.6380120068788528,
441
+ "num_tokens": 676568.0,
442
+ "step": 48
443
+ },
444
+ {
445
+ "epoch": 0.3136,
446
+ "grad_norm": 3.360166549682617,
447
+ "learning_rate": 5e-06,
448
+ "loss": 1.0643,
449
+ "mean_token_accuracy": 0.7167638763785362,
450
+ "num_tokens": 690799.0,
451
+ "step": 49
452
+ },
453
+ {
454
+ "epoch": 0.32,
455
+ "grad_norm": 3.504518508911133,
456
+ "learning_rate": 5e-06,
457
+ "loss": 1.2447,
458
+ "mean_token_accuracy": 0.6620298400521278,
459
+ "num_tokens": 706489.0,
460
+ "step": 50
461
+ },
462
+ {
463
+ "epoch": 0.3264,
464
+ "grad_norm": 3.65311861038208,
465
+ "learning_rate": 5e-06,
466
+ "loss": 1.4979,
467
+ "mean_token_accuracy": 0.6288957595825195,
468
+ "num_tokens": 721499.0,
469
+ "step": 51
470
+ },
471
+ {
472
+ "epoch": 0.3328,
473
+ "grad_norm": 3.348665475845337,
474
+ "learning_rate": 5e-06,
475
+ "loss": 1.3043,
476
+ "mean_token_accuracy": 0.6755396276712418,
477
+ "num_tokens": 738238.0,
478
+ "step": 52
479
+ },
480
+ {
481
+ "epoch": 0.3392,
482
+ "grad_norm": 3.7475781440734863,
483
+ "learning_rate": 5e-06,
484
+ "loss": 1.3655,
485
+ "mean_token_accuracy": 0.6580836474895477,
486
+ "num_tokens": 750933.0,
487
+ "step": 53
488
+ },
489
+ {
490
+ "epoch": 0.3456,
491
+ "grad_norm": 3.65537691116333,
492
+ "learning_rate": 5e-06,
493
+ "loss": 1.428,
494
+ "mean_token_accuracy": 0.6435032188892365,
495
+ "num_tokens": 766190.0,
496
+ "step": 54
497
+ },
498
+ {
499
+ "epoch": 0.352,
500
+ "grad_norm": 3.4737765789031982,
501
+ "learning_rate": 5e-06,
502
+ "loss": 1.3459,
503
+ "mean_token_accuracy": 0.6618407666683197,
504
+ "num_tokens": 781058.0,
505
+ "step": 55
506
+ },
507
+ {
508
+ "epoch": 0.3584,
509
+ "grad_norm": 3.7944328784942627,
510
+ "learning_rate": 5e-06,
511
+ "loss": 1.4812,
512
+ "mean_token_accuracy": 0.6265006363391876,
513
+ "num_tokens": 793591.0,
514
+ "step": 56
515
+ },
516
+ {
517
+ "epoch": 0.3648,
518
+ "grad_norm": 3.4754698276519775,
519
+ "learning_rate": 5e-06,
520
+ "loss": 1.1157,
521
+ "mean_token_accuracy": 0.6880814656615257,
522
+ "num_tokens": 807562.0,
523
+ "step": 57
524
+ },
525
+ {
526
+ "epoch": 0.3712,
527
+ "grad_norm": 3.345503568649292,
528
+ "learning_rate": 5e-06,
529
+ "loss": 1.2552,
530
+ "mean_token_accuracy": 0.6683396399021149,
531
+ "num_tokens": 822029.0,
532
+ "step": 58
533
+ },
534
+ {
535
+ "epoch": 0.3776,
536
+ "grad_norm": 3.702852487564087,
537
+ "learning_rate": 5e-06,
538
+ "loss": 1.2624,
539
+ "mean_token_accuracy": 0.6913647577166557,
540
+ "num_tokens": 833999.0,
541
+ "step": 59
542
+ },
543
+ {
544
+ "epoch": 0.384,
545
+ "grad_norm": 3.169661045074463,
546
+ "learning_rate": 5e-06,
547
+ "loss": 1.3735,
548
+ "mean_token_accuracy": 0.6340024396777153,
549
+ "num_tokens": 848952.0,
550
+ "step": 60
551
+ },
552
+ {
553
+ "epoch": 0.3904,
554
+ "grad_norm": 3.8648996353149414,
555
+ "learning_rate": 5e-06,
556
+ "loss": 1.2589,
557
+ "mean_token_accuracy": 0.6712265312671661,
558
+ "num_tokens": 865078.0,
559
+ "step": 61
560
+ },
561
+ {
562
+ "epoch": 0.3968,
563
+ "grad_norm": 3.6297190189361572,
564
+ "learning_rate": 5e-06,
565
+ "loss": 1.228,
566
+ "mean_token_accuracy": 0.6683308631181717,
567
+ "num_tokens": 877815.0,
568
+ "step": 62
569
+ },
570
+ {
571
+ "epoch": 0.4032,
572
+ "grad_norm": 3.3031067848205566,
573
+ "learning_rate": 5e-06,
574
+ "loss": 1.3284,
575
+ "mean_token_accuracy": 0.6419568583369255,
576
+ "num_tokens": 891433.0,
577
+ "step": 63
578
+ },
579
+ {
580
+ "epoch": 0.4096,
581
+ "grad_norm": 3.6126585006713867,
582
+ "learning_rate": 5e-06,
583
+ "loss": 1.327,
584
+ "mean_token_accuracy": 0.6622776314616203,
585
+ "num_tokens": 904268.0,
586
+ "step": 64
587
+ },
588
+ {
589
+ "epoch": 0.416,
590
+ "grad_norm": 3.407878875732422,
591
+ "learning_rate": 5e-06,
592
+ "loss": 1.3261,
593
+ "mean_token_accuracy": 0.6527406051754951,
594
+ "num_tokens": 919814.0,
595
+ "step": 65
596
+ },
597
+ {
598
+ "epoch": 0.4224,
599
+ "grad_norm": 3.381817102432251,
600
+ "learning_rate": 5e-06,
601
+ "loss": 1.2016,
602
+ "mean_token_accuracy": 0.674388200044632,
603
+ "num_tokens": 933076.0,
604
+ "step": 66
605
+ },
606
+ {
607
+ "epoch": 0.4288,
608
+ "grad_norm": 3.4750583171844482,
609
+ "learning_rate": 5e-06,
610
+ "loss": 1.2737,
611
+ "mean_token_accuracy": 0.6750325560569763,
612
+ "num_tokens": 947949.0,
613
+ "step": 67
614
+ },
615
+ {
616
+ "epoch": 0.4352,
617
+ "grad_norm": 3.401789426803589,
618
+ "learning_rate": 5e-06,
619
+ "loss": 1.443,
620
+ "mean_token_accuracy": 0.6314026713371277,
621
+ "num_tokens": 963300.0,
622
+ "step": 68
623
+ },
624
+ {
625
+ "epoch": 0.4416,
626
+ "grad_norm": 3.7318997383117676,
627
+ "learning_rate": 5e-06,
628
+ "loss": 1.3147,
629
+ "mean_token_accuracy": 0.6670505329966545,
630
+ "num_tokens": 978591.0,
631
+ "step": 69
632
+ },
633
+ {
634
+ "epoch": 0.448,
635
+ "grad_norm": 3.5171058177948,
636
+ "learning_rate": 5e-06,
637
+ "loss": 1.2636,
638
+ "mean_token_accuracy": 0.6640981435775757,
639
+ "num_tokens": 995710.0,
640
+ "step": 70
641
+ },
642
+ {
643
+ "epoch": 0.4544,
644
+ "grad_norm": 3.5986971855163574,
645
+ "learning_rate": 5e-06,
646
+ "loss": 1.4008,
647
+ "mean_token_accuracy": 0.6460948511958122,
648
+ "num_tokens": 1008761.0,
649
+ "step": 71
650
+ },
651
+ {
652
+ "epoch": 0.4608,
653
+ "grad_norm": 3.790513753890991,
654
+ "learning_rate": 5e-06,
655
+ "loss": 1.0588,
656
+ "mean_token_accuracy": 0.7220645099878311,
657
+ "num_tokens": 1022118.0,
658
+ "step": 72
659
+ },
660
+ {
661
+ "epoch": 0.4672,
662
+ "grad_norm": 3.2633607387542725,
663
+ "learning_rate": 5e-06,
664
+ "loss": 1.4292,
665
+ "mean_token_accuracy": 0.633380301296711,
666
+ "num_tokens": 1036809.0,
667
+ "step": 73
668
+ },
669
+ {
670
+ "epoch": 0.4736,
671
+ "grad_norm": 4.51404333114624,
672
+ "learning_rate": 5e-06,
673
+ "loss": 1.2447,
674
+ "mean_token_accuracy": 0.6805145666003227,
675
+ "num_tokens": 1051135.0,
676
+ "step": 74
677
+ },
678
+ {
679
+ "epoch": 0.48,
680
+ "grad_norm": 3.9093401432037354,
681
+ "learning_rate": 5e-06,
682
+ "loss": 1.2158,
683
+ "mean_token_accuracy": 0.6765370890498161,
684
+ "num_tokens": 1066566.0,
685
+ "step": 75
686
+ },
687
+ {
688
+ "epoch": 0.4864,
689
+ "grad_norm": 3.798604726791382,
690
+ "learning_rate": 5e-06,
691
+ "loss": 1.2877,
692
+ "mean_token_accuracy": 0.6586913466453552,
693
+ "num_tokens": 1080474.0,
694
+ "step": 76
695
+ },
696
+ {
697
+ "epoch": 0.4928,
698
+ "grad_norm": 3.8617875576019287,
699
+ "learning_rate": 5e-06,
700
+ "loss": 1.3414,
701
+ "mean_token_accuracy": 0.6375375539064407,
702
+ "num_tokens": 1092414.0,
703
+ "step": 77
704
+ },
705
+ {
706
+ "epoch": 0.4992,
707
+ "grad_norm": 3.4888036251068115,
708
+ "learning_rate": 5e-06,
709
+ "loss": 1.3153,
710
+ "mean_token_accuracy": 0.6624401807785034,
711
+ "num_tokens": 1105750.0,
712
+ "step": 78
713
+ },
714
+ {
715
+ "epoch": 0.5056,
716
+ "grad_norm": 4.31488037109375,
717
+ "learning_rate": 5e-06,
718
+ "loss": 1.5732,
719
+ "mean_token_accuracy": 0.631282813847065,
720
+ "num_tokens": 1119046.0,
721
+ "step": 79
722
+ },
723
+ {
724
+ "epoch": 0.512,
725
+ "grad_norm": 3.053265333175659,
726
+ "learning_rate": 5e-06,
727
+ "loss": 1.1599,
728
+ "mean_token_accuracy": 0.6818074658513069,
729
+ "num_tokens": 1135088.0,
730
+ "step": 80
731
+ },
732
+ {
733
+ "epoch": 0.5184,
734
+ "grad_norm": 3.616725444793701,
735
+ "learning_rate": 5e-06,
736
+ "loss": 1.43,
737
+ "mean_token_accuracy": 0.6406295001506805,
738
+ "num_tokens": 1151156.0,
739
+ "step": 81
740
+ },
741
+ {
742
+ "epoch": 0.5248,
743
+ "grad_norm": 4.136235237121582,
744
+ "learning_rate": 5e-06,
745
+ "loss": 1.0219,
746
+ "mean_token_accuracy": 0.730844259262085,
747
+ "num_tokens": 1167434.0,
748
+ "step": 82
749
+ },
750
+ {
751
+ "epoch": 0.5312,
752
+ "grad_norm": 3.338418483734131,
753
+ "learning_rate": 5e-06,
754
+ "loss": 1.2563,
755
+ "mean_token_accuracy": 0.6696086227893829,
756
+ "num_tokens": 1180703.0,
757
+ "step": 83
758
+ },
759
+ {
760
+ "epoch": 0.5376,
761
+ "grad_norm": 4.320113658905029,
762
+ "learning_rate": 5e-06,
763
+ "loss": 1.3359,
764
+ "mean_token_accuracy": 0.6449860632419586,
765
+ "num_tokens": 1192466.0,
766
+ "step": 84
767
+ },
768
+ {
769
+ "epoch": 0.544,
770
+ "grad_norm": 3.6109604835510254,
771
+ "learning_rate": 5e-06,
772
+ "loss": 1.3403,
773
+ "mean_token_accuracy": 0.6509855315089226,
774
+ "num_tokens": 1207465.0,
775
+ "step": 85
776
+ },
777
+ {
778
+ "epoch": 0.5504,
779
+ "grad_norm": 3.9164552688598633,
780
+ "learning_rate": 5e-06,
781
+ "loss": 1.2705,
782
+ "mean_token_accuracy": 0.657896488904953,
783
+ "num_tokens": 1218611.0,
784
+ "step": 86
785
+ },
786
+ {
787
+ "epoch": 0.5568,
788
+ "grad_norm": 3.731842279434204,
789
+ "learning_rate": 5e-06,
790
+ "loss": 1.2964,
791
+ "mean_token_accuracy": 0.6592853665351868,
792
+ "num_tokens": 1231942.0,
793
+ "step": 87
794
+ },
795
+ {
796
+ "epoch": 0.5632,
797
+ "grad_norm": 3.651313304901123,
798
+ "learning_rate": 5e-06,
799
+ "loss": 1.36,
800
+ "mean_token_accuracy": 0.6576117277145386,
801
+ "num_tokens": 1246108.0,
802
+ "step": 88
803
+ },
804
+ {
805
+ "epoch": 0.5696,
806
+ "grad_norm": 3.089268445968628,
807
+ "learning_rate": 5e-06,
808
+ "loss": 1.2009,
809
+ "mean_token_accuracy": 0.6747561246156693,
810
+ "num_tokens": 1261356.0,
811
+ "step": 89
812
+ },
813
+ {
814
+ "epoch": 0.576,
815
+ "grad_norm": 3.238022565841675,
816
+ "learning_rate": 5e-06,
817
+ "loss": 1.3602,
818
+ "mean_token_accuracy": 0.6399379670619965,
819
+ "num_tokens": 1277200.0,
820
+ "step": 90
821
+ },
822
+ {
823
+ "epoch": 0.5824,
824
+ "grad_norm": 3.467700481414795,
825
+ "learning_rate": 5e-06,
826
+ "loss": 1.3403,
827
+ "mean_token_accuracy": 0.6624496802687645,
828
+ "num_tokens": 1292186.0,
829
+ "step": 91
830
+ },
831
+ {
832
+ "epoch": 0.5888,
833
+ "grad_norm": 3.325700283050537,
834
+ "learning_rate": 5e-06,
835
+ "loss": 1.2627,
836
+ "mean_token_accuracy": 0.6559372171759605,
837
+ "num_tokens": 1304719.0,
838
+ "step": 92
839
+ },
840
+ {
841
+ "epoch": 0.5952,
842
+ "grad_norm": 3.621305227279663,
843
+ "learning_rate": 5e-06,
844
+ "loss": 1.1766,
845
+ "mean_token_accuracy": 0.6822386905550957,
846
+ "num_tokens": 1318790.0,
847
+ "step": 93
848
+ },
849
+ {
850
+ "epoch": 0.6016,
851
+ "grad_norm": 3.729935884475708,
852
+ "learning_rate": 5e-06,
853
+ "loss": 1.3674,
854
+ "mean_token_accuracy": 0.645961195230484,
855
+ "num_tokens": 1329771.0,
856
+ "step": 94
857
+ },
858
+ {
859
+ "epoch": 0.608,
860
+ "grad_norm": 3.6264734268188477,
861
+ "learning_rate": 5e-06,
862
+ "loss": 1.1152,
863
+ "mean_token_accuracy": 0.687596932053566,
864
+ "num_tokens": 1342886.0,
865
+ "step": 95
866
+ },
867
+ {
868
+ "epoch": 0.6144,
869
+ "grad_norm": 3.485283136367798,
870
+ "learning_rate": 5e-06,
871
+ "loss": 1.4425,
872
+ "mean_token_accuracy": 0.6406073719263077,
873
+ "num_tokens": 1358395.0,
874
+ "step": 96
875
+ },
876
+ {
877
+ "epoch": 0.6208,
878
+ "grad_norm": 3.9530715942382812,
879
+ "learning_rate": 5e-06,
880
+ "loss": 1.2683,
881
+ "mean_token_accuracy": 0.6643862873315811,
882
+ "num_tokens": 1371799.0,
883
+ "step": 97
884
+ },
885
+ {
886
+ "epoch": 0.6272,
887
+ "grad_norm": 4.173793315887451,
888
+ "learning_rate": 5e-06,
889
+ "loss": 1.2914,
890
+ "mean_token_accuracy": 0.660434328019619,
891
+ "num_tokens": 1383470.0,
892
+ "step": 98
893
+ },
894
+ {
895
+ "epoch": 0.6336,
896
+ "grad_norm": 3.614288330078125,
897
+ "learning_rate": 5e-06,
898
+ "loss": 1.2038,
899
+ "mean_token_accuracy": 0.6892227753996849,
900
+ "num_tokens": 1400378.0,
901
+ "step": 99
902
+ },
903
+ {
904
+ "epoch": 0.64,
905
+ "grad_norm": 2.9820358753204346,
906
+ "learning_rate": 5e-06,
907
+ "loss": 1.1759,
908
+ "mean_token_accuracy": 0.6901198551058769,
909
+ "num_tokens": 1417726.0,
910
+ "step": 100
911
+ },
912
+ {
913
+ "epoch": 0.6464,
914
+ "grad_norm": 3.3349997997283936,
915
+ "learning_rate": 5e-06,
916
+ "loss": 1.3748,
917
+ "mean_token_accuracy": 0.6374614164233208,
918
+ "num_tokens": 1433643.0,
919
+ "step": 101
920
+ },
921
+ {
922
+ "epoch": 0.6528,
923
+ "grad_norm": 3.2100603580474854,
924
+ "learning_rate": 5e-06,
925
+ "loss": 1.3996,
926
+ "mean_token_accuracy": 0.6497422009706497,
927
+ "num_tokens": 1449613.0,
928
+ "step": 102
929
+ },
930
+ {
931
+ "epoch": 0.6592,
932
+ "grad_norm": 3.778851270675659,
933
+ "learning_rate": 5e-06,
934
+ "loss": 1.0333,
935
+ "mean_token_accuracy": 0.7377151399850845,
936
+ "num_tokens": 1462451.0,
937
+ "step": 103
938
+ },
939
+ {
940
+ "epoch": 0.6656,
941
+ "grad_norm": 3.1992948055267334,
942
+ "learning_rate": 5e-06,
943
+ "loss": 1.3104,
944
+ "mean_token_accuracy": 0.6540233194828033,
945
+ "num_tokens": 1476702.0,
946
+ "step": 104
947
+ },
948
+ {
949
+ "epoch": 0.672,
950
+ "grad_norm": 3.510753631591797,
951
+ "learning_rate": 5e-06,
952
+ "loss": 1.4638,
953
+ "mean_token_accuracy": 0.6315533146262169,
954
+ "num_tokens": 1491797.0,
955
+ "step": 105
956
+ },
957
+ {
958
+ "epoch": 0.6784,
959
+ "grad_norm": 3.976029396057129,
960
+ "learning_rate": 5e-06,
961
+ "loss": 1.5173,
962
+ "mean_token_accuracy": 0.6174216568470001,
963
+ "num_tokens": 1506937.0,
964
+ "step": 106
965
+ },
966
+ {
967
+ "epoch": 0.6848,
968
+ "grad_norm": 4.135453224182129,
969
+ "learning_rate": 5e-06,
970
+ "loss": 1.2671,
971
+ "mean_token_accuracy": 0.6688078567385674,
972
+ "num_tokens": 1517912.0,
973
+ "step": 107
974
+ },
975
+ {
976
+ "epoch": 0.6912,
977
+ "grad_norm": 3.417698621749878,
978
+ "learning_rate": 5e-06,
979
+ "loss": 1.3911,
980
+ "mean_token_accuracy": 0.653182253241539,
981
+ "num_tokens": 1534828.0,
982
+ "step": 108
983
+ },
984
+ {
985
+ "epoch": 0.6976,
986
+ "grad_norm": 3.1860220432281494,
987
+ "learning_rate": 5e-06,
988
+ "loss": 1.1702,
989
+ "mean_token_accuracy": 0.6850942298769951,
990
+ "num_tokens": 1549388.0,
991
+ "step": 109
992
+ },
993
+ {
994
+ "epoch": 0.704,
995
+ "grad_norm": 4.109538555145264,
996
+ "learning_rate": 5e-06,
997
+ "loss": 0.9738,
998
+ "mean_token_accuracy": 0.7242774963378906,
999
+ "num_tokens": 1561336.0,
1000
+ "step": 110
1001
+ },
1002
+ {
1003
+ "epoch": 0.7104,
1004
+ "grad_norm": 3.6836867332458496,
1005
+ "learning_rate": 5e-06,
1006
+ "loss": 1.1749,
1007
+ "mean_token_accuracy": 0.6976689174771309,
1008
+ "num_tokens": 1576039.0,
1009
+ "step": 111
1010
+ },
1011
+ {
1012
+ "epoch": 0.7168,
1013
+ "grad_norm": 3.6196529865264893,
1014
+ "learning_rate": 5e-06,
1015
+ "loss": 1.2509,
1016
+ "mean_token_accuracy": 0.6862899139523506,
1017
+ "num_tokens": 1590356.0,
1018
+ "step": 112
1019
+ },
1020
+ {
1021
+ "epoch": 0.7232,
1022
+ "grad_norm": 3.6636641025543213,
1023
+ "learning_rate": 5e-06,
1024
+ "loss": 1.222,
1025
+ "mean_token_accuracy": 0.6742222532629967,
1026
+ "num_tokens": 1603828.0,
1027
+ "step": 113
1028
+ },
1029
+ {
1030
+ "epoch": 0.7296,
1031
+ "grad_norm": 3.1560022830963135,
1032
+ "learning_rate": 5e-06,
1033
+ "loss": 1.238,
1034
+ "mean_token_accuracy": 0.6736825779080391,
1035
+ "num_tokens": 1621001.0,
1036
+ "step": 114
1037
+ },
1038
+ {
1039
+ "epoch": 0.736,
1040
+ "grad_norm": 3.1845784187316895,
1041
+ "learning_rate": 5e-06,
1042
+ "loss": 1.3708,
1043
+ "mean_token_accuracy": 0.6455038785934448,
1044
+ "num_tokens": 1636499.0,
1045
+ "step": 115
1046
+ },
1047
+ {
1048
+ "epoch": 0.7424,
1049
+ "grad_norm": 3.5999317169189453,
1050
+ "learning_rate": 5e-06,
1051
+ "loss": 1.1897,
1052
+ "mean_token_accuracy": 0.6913699880242348,
1053
+ "num_tokens": 1649043.0,
1054
+ "step": 116
1055
+ },
1056
+ {
1057
+ "epoch": 0.7488,
1058
+ "grad_norm": 3.9774956703186035,
1059
+ "learning_rate": 5e-06,
1060
+ "loss": 1.0401,
1061
+ "mean_token_accuracy": 0.7169342339038849,
1062
+ "num_tokens": 1663873.0,
1063
+ "step": 117
1064
+ },
1065
+ {
1066
+ "epoch": 0.7552,
1067
+ "grad_norm": 3.3667006492614746,
1068
+ "learning_rate": 5e-06,
1069
+ "loss": 1.389,
1070
+ "mean_token_accuracy": 0.6400957256555557,
1071
+ "num_tokens": 1679667.0,
1072
+ "step": 118
1073
+ },
1074
+ {
1075
+ "epoch": 0.7616,
1076
+ "grad_norm": 3.5805838108062744,
1077
+ "learning_rate": 5e-06,
1078
+ "loss": 1.3896,
1079
+ "mean_token_accuracy": 0.6595878526568413,
1080
+ "num_tokens": 1695115.0,
1081
+ "step": 119
1082
+ },
1083
+ {
1084
+ "epoch": 0.768,
1085
+ "grad_norm": 3.8038456439971924,
1086
+ "learning_rate": 5e-06,
1087
+ "loss": 1.4199,
1088
+ "mean_token_accuracy": 0.6296796873211861,
1089
+ "num_tokens": 1707134.0,
1090
+ "step": 120
1091
+ },
1092
+ {
1093
+ "epoch": 0.7744,
1094
+ "grad_norm": 3.626406669616699,
1095
+ "learning_rate": 5e-06,
1096
+ "loss": 1.2602,
1097
+ "mean_token_accuracy": 0.6764451712369919,
1098
+ "num_tokens": 1720062.0,
1099
+ "step": 121
1100
+ },
1101
+ {
1102
+ "epoch": 0.7808,
1103
+ "grad_norm": 4.185618877410889,
1104
+ "learning_rate": 5e-06,
1105
+ "loss": 1.4208,
1106
+ "mean_token_accuracy": 0.6561515182256699,
1107
+ "num_tokens": 1731658.0,
1108
+ "step": 122
1109
+ },
1110
+ {
1111
+ "epoch": 0.7872,
1112
+ "grad_norm": 3.6125693321228027,
1113
+ "learning_rate": 5e-06,
1114
+ "loss": 1.1589,
1115
+ "mean_token_accuracy": 0.6781378239393234,
1116
+ "num_tokens": 1747033.0,
1117
+ "step": 123
1118
+ },
1119
+ {
1120
+ "epoch": 0.7936,
1121
+ "grad_norm": 4.1037373542785645,
1122
+ "learning_rate": 5e-06,
1123
+ "loss": 1.2256,
1124
+ "mean_token_accuracy": 0.6652502194046974,
1125
+ "num_tokens": 1758454.0,
1126
+ "step": 124
1127
+ },
1128
+ {
1129
+ "epoch": 0.8,
1130
+ "grad_norm": 3.3770744800567627,
1131
+ "learning_rate": 5e-06,
1132
+ "loss": 1.2673,
1133
+ "mean_token_accuracy": 0.6793160140514374,
1134
+ "num_tokens": 1773549.0,
1135
+ "step": 125
1136
+ },
1137
+ {
1138
+ "epoch": 0.8064,
1139
+ "grad_norm": 3.3607754707336426,
1140
+ "learning_rate": 5e-06,
1141
+ "loss": 1.2752,
1142
+ "mean_token_accuracy": 0.6853690519928932,
1143
+ "num_tokens": 1789080.0,
1144
+ "step": 126
1145
+ },
1146
+ {
1147
+ "epoch": 0.8128,
1148
+ "grad_norm": 3.806748390197754,
1149
+ "learning_rate": 5e-06,
1150
+ "loss": 1.1944,
1151
+ "mean_token_accuracy": 0.6872068718075752,
1152
+ "num_tokens": 1801796.0,
1153
+ "step": 127
1154
+ },
1155
+ {
1156
+ "epoch": 0.8192,
1157
+ "grad_norm": 3.8479700088500977,
1158
+ "learning_rate": 5e-06,
1159
+ "loss": 1.3977,
1160
+ "mean_token_accuracy": 0.636077530682087,
1161
+ "num_tokens": 1814242.0,
1162
+ "step": 128
1163
+ },
1164
+ {
1165
+ "epoch": 0.8256,
1166
+ "grad_norm": 3.4976117610931396,
1167
+ "learning_rate": 5e-06,
1168
+ "loss": 1.1139,
1169
+ "mean_token_accuracy": 0.697039358317852,
1170
+ "num_tokens": 1826839.0,
1171
+ "step": 129
1172
+ },
1173
+ {
1174
+ "epoch": 0.832,
1175
+ "grad_norm": 3.545767307281494,
1176
+ "learning_rate": 5e-06,
1177
+ "loss": 1.4587,
1178
+ "mean_token_accuracy": 0.6343350037932396,
1179
+ "num_tokens": 1840704.0,
1180
+ "step": 130
1181
+ },
1182
+ {
1183
+ "epoch": 0.8384,
1184
+ "grad_norm": 3.3037948608398438,
1185
+ "learning_rate": 5e-06,
1186
+ "loss": 1.051,
1187
+ "mean_token_accuracy": 0.7136509492993355,
1188
+ "num_tokens": 1854753.0,
1189
+ "step": 131
1190
+ },
1191
+ {
1192
+ "epoch": 0.8448,
1193
+ "grad_norm": 3.749946117401123,
1194
+ "learning_rate": 5e-06,
1195
+ "loss": 1.5613,
1196
+ "mean_token_accuracy": 0.6120798960328102,
1197
+ "num_tokens": 1868075.0,
1198
+ "step": 132
1199
+ },
1200
+ {
1201
+ "epoch": 0.8512,
1202
+ "grad_norm": 3.2564704418182373,
1203
+ "learning_rate": 5e-06,
1204
+ "loss": 1.5152,
1205
+ "mean_token_accuracy": 0.6301259547472,
1206
+ "num_tokens": 1883671.0,
1207
+ "step": 133
1208
+ },
1209
+ {
1210
+ "epoch": 0.8576,
1211
+ "grad_norm": 3.6490490436553955,
1212
+ "learning_rate": 5e-06,
1213
+ "loss": 1.3408,
1214
+ "mean_token_accuracy": 0.657665990293026,
1215
+ "num_tokens": 1895662.0,
1216
+ "step": 134
1217
+ },
1218
+ {
1219
+ "epoch": 0.864,
1220
+ "grad_norm": 3.407827615737915,
1221
+ "learning_rate": 5e-06,
1222
+ "loss": 1.6605,
1223
+ "mean_token_accuracy": 0.5928493365645409,
1224
+ "num_tokens": 1910847.0,
1225
+ "step": 135
1226
+ },
1227
+ {
1228
+ "epoch": 0.8704,
1229
+ "grad_norm": 4.415482521057129,
1230
+ "learning_rate": 5e-06,
1231
+ "loss": 1.1946,
1232
+ "mean_token_accuracy": 0.6526624895632267,
1233
+ "num_tokens": 1925173.0,
1234
+ "step": 136
1235
+ },
1236
+ {
1237
+ "epoch": 0.8768,
1238
+ "grad_norm": 4.364404201507568,
1239
+ "learning_rate": 5e-06,
1240
+ "loss": 1.3597,
1241
+ "mean_token_accuracy": 0.6597945764660835,
1242
+ "num_tokens": 1938507.0,
1243
+ "step": 137
1244
+ },
1245
+ {
1246
+ "epoch": 0.8832,
1247
+ "grad_norm": 3.498443603515625,
1248
+ "learning_rate": 5e-06,
1249
+ "loss": 1.2335,
1250
+ "mean_token_accuracy": 0.6793266087770462,
1251
+ "num_tokens": 1955459.0,
1252
+ "step": 138
1253
+ },
1254
+ {
1255
+ "epoch": 0.8896,
1256
+ "grad_norm": 3.321516275405884,
1257
+ "learning_rate": 5e-06,
1258
+ "loss": 1.1991,
1259
+ "mean_token_accuracy": 0.6588682383298874,
1260
+ "num_tokens": 1970350.0,
1261
+ "step": 139
1262
+ },
1263
+ {
1264
+ "epoch": 0.896,
1265
+ "grad_norm": 3.556004524230957,
1266
+ "learning_rate": 5e-06,
1267
+ "loss": 1.1146,
1268
+ "mean_token_accuracy": 0.7137203514575958,
1269
+ "num_tokens": 1984042.0,
1270
+ "step": 140
1271
+ },
1272
+ {
1273
+ "epoch": 0.9024,
1274
+ "grad_norm": 3.075051784515381,
1275
+ "learning_rate": 5e-06,
1276
+ "loss": 1.2796,
1277
+ "mean_token_accuracy": 0.6648645251989365,
1278
+ "num_tokens": 2000795.0,
1279
+ "step": 141
1280
+ },
1281
+ {
1282
+ "epoch": 0.9088,
1283
+ "grad_norm": 3.5248730182647705,
1284
+ "learning_rate": 5e-06,
1285
+ "loss": 1.2528,
1286
+ "mean_token_accuracy": 0.6755008921027184,
1287
+ "num_tokens": 2015306.0,
1288
+ "step": 142
1289
+ },
1290
+ {
1291
+ "epoch": 0.9152,
1292
+ "grad_norm": 3.545891284942627,
1293
+ "learning_rate": 5e-06,
1294
+ "loss": 1.5914,
1295
+ "mean_token_accuracy": 0.6275190636515617,
1296
+ "num_tokens": 2030477.0,
1297
+ "step": 143
1298
+ },
1299
+ {
1300
+ "epoch": 0.9216,
1301
+ "grad_norm": 3.4129698276519775,
1302
+ "learning_rate": 5e-06,
1303
+ "loss": 1.1794,
1304
+ "mean_token_accuracy": 0.6718302294611931,
1305
+ "num_tokens": 2045094.0,
1306
+ "step": 144
1307
+ },
1308
+ {
1309
+ "epoch": 0.928,
1310
+ "grad_norm": 4.104367733001709,
1311
+ "learning_rate": 5e-06,
1312
+ "loss": 1.2672,
1313
+ "mean_token_accuracy": 0.680908165872097,
1314
+ "num_tokens": 2059572.0,
1315
+ "step": 145
1316
+ },
1317
+ {
1318
+ "epoch": 0.9344,
1319
+ "grad_norm": 3.316425323486328,
1320
+ "learning_rate": 5e-06,
1321
+ "loss": 1.4394,
1322
+ "mean_token_accuracy": 0.6278339251875877,
1323
+ "num_tokens": 2074507.0,
1324
+ "step": 146
1325
+ },
1326
+ {
1327
+ "epoch": 0.9408,
1328
+ "grad_norm": 3.6546125411987305,
1329
+ "learning_rate": 5e-06,
1330
+ "loss": 1.1124,
1331
+ "mean_token_accuracy": 0.6924281641840935,
1332
+ "num_tokens": 2087040.0,
1333
+ "step": 147
1334
+ },
1335
+ {
1336
+ "epoch": 0.9472,
1337
+ "grad_norm": 3.6782755851745605,
1338
+ "learning_rate": 5e-06,
1339
+ "loss": 1.3804,
1340
+ "mean_token_accuracy": 0.6391346678137779,
1341
+ "num_tokens": 2101730.0,
1342
+ "step": 148
1343
+ },
1344
+ {
1345
+ "epoch": 0.9536,
1346
+ "grad_norm": 3.386892080307007,
1347
+ "learning_rate": 5e-06,
1348
+ "loss": 1.2437,
1349
+ "mean_token_accuracy": 0.6677525192499161,
1350
+ "num_tokens": 2116581.0,
1351
+ "step": 149
1352
+ },
1353
+ {
1354
+ "epoch": 0.96,
1355
+ "grad_norm": 3.5542984008789062,
1356
+ "learning_rate": 5e-06,
1357
+ "loss": 1.0982,
1358
+ "mean_token_accuracy": 0.7068964689970016,
1359
+ "num_tokens": 2131452.0,
1360
+ "step": 150
1361
+ },
1362
+ {
1363
+ "epoch": 0.9664,
1364
+ "grad_norm": 3.5495564937591553,
1365
+ "learning_rate": 5e-06,
1366
+ "loss": 1.4026,
1367
+ "mean_token_accuracy": 0.6545470431447029,
1368
+ "num_tokens": 2147428.0,
1369
+ "step": 151
1370
+ },
1371
+ {
1372
+ "epoch": 0.9728,
1373
+ "grad_norm": 3.4057626724243164,
1374
+ "learning_rate": 5e-06,
1375
+ "loss": 1.3493,
1376
+ "mean_token_accuracy": 0.6586249023675919,
1377
+ "num_tokens": 2164711.0,
1378
+ "step": 152
1379
+ },
1380
+ {
1381
+ "epoch": 0.9792,
1382
+ "grad_norm": 3.815967559814453,
1383
+ "learning_rate": 5e-06,
1384
+ "loss": 1.3319,
1385
+ "mean_token_accuracy": 0.6602702289819717,
1386
+ "num_tokens": 2177075.0,
1387
+ "step": 153
1388
+ },
1389
+ {
1390
+ "epoch": 0.9856,
1391
+ "grad_norm": 3.8833365440368652,
1392
+ "learning_rate": 5e-06,
1393
+ "loss": 1.3236,
1394
+ "mean_token_accuracy": 0.6528388634324074,
1395
+ "num_tokens": 2190015.0,
1396
+ "step": 154
1397
+ },
1398
+ {
1399
+ "epoch": 0.992,
1400
+ "grad_norm": 3.010931968688965,
1401
+ "learning_rate": 5e-06,
1402
+ "loss": 1.3359,
1403
+ "mean_token_accuracy": 0.6606268882751465,
1404
+ "num_tokens": 2208904.0,
1405
+ "step": 155
1406
+ },
1407
+ {
1408
+ "epoch": 0.9984,
1409
+ "grad_norm": 3.366220474243164,
1410
+ "learning_rate": 5e-06,
1411
+ "loss": 1.2807,
1412
+ "mean_token_accuracy": 0.6731663048267365,
1413
+ "num_tokens": 2223354.0,
1414
+ "step": 156
1415
+ }
1416
+ ],
1417
+ "logging_steps": 1,
1418
+ "max_steps": 156,
1419
+ "num_input_tokens_seen": 0,
1420
+ "num_train_epochs": 1,
1421
+ "save_steps": 250,
1422
+ "stateful_callbacks": {
1423
+ "TrainerControl": {
1424
+ "args": {
1425
+ "should_epoch_stop": false,
1426
+ "should_evaluate": false,
1427
+ "should_log": false,
1428
+ "should_save": true,
1429
+ "should_training_stop": true
1430
+ },
1431
+ "attributes": {}
1432
+ }
1433
+ },
1434
+ "total_flos": 6046380687360.0,
1435
+ "train_batch_size": 4,
1436
+ "trial_name": null,
1437
+ "trial_params": null
1438
+ }