config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "bigscience/bloom-560m",
3
+ "apply_residual_connection_post_layernorm": false,
4
+ "architectures": [
5
+ "BloomForCausalLM"
6
+ ],
7
+ "attention_dropout": 0.0,
8
+ "attention_softmax_in_fp32": true,
9
+ "bias_dropout_fusion": true,
10
+ "bos_token_id": 1,
11
+ "eos_token_id": 2,
12
+ "hidden_dropout": 0.0,
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "layer_norm_epsilon": 1e-05,
16
+ "masked_softmax_fusion": true,
17
+ "model_type": "bloom",
18
+ "n_head": 16,
19
+ "n_inner": null,
20
+ "n_layer": 24,
21
+ "offset_alibi": 100,
22
+ "pad_token_id": 3,
23
+ "pretraining_tp": 1,
24
+ "skip_bias_add": true,
25
+ "skip_bias_add_qkv": false,
26
+ "slow_but_exact": false,
27
+ "torch_dtype": "float32",
28
+ "transformers_version": "4.27.1",
29
+ "unk_token_id": 0,
30
+ "use_cache": true,
31
+ "vocab_size": 250880
32
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 3,
6
+ "transformers_version": "4.27.1"
7
+ }
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b5fa02ce1f2057ffdcccd86019817387a38e6b835bf0ecf5d2277b489b598e7
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43b53b9f3902cdc0cb20261e0ca2072411cb0f630ebf92e5e4b4caeb3994fd1f
3
+ size 1064
trainer_state.json ADDED
@@ -0,0 +1,916 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.893939393939394,
5
+ "global_step": 1500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1.9873737373737378e-05,
13
+ "loss": 8.938,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 1.9747474747474747e-05,
19
+ "loss": 0.3506,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 1.9621212121212123e-05,
25
+ "loss": 0.1109,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 1.9494949494949496e-05,
31
+ "loss": 0.1104,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.06,
36
+ "learning_rate": 1.936868686868687e-05,
37
+ "loss": 0.1353,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 1.9242424242424244e-05,
43
+ "loss": 0.1138,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.09,
48
+ "learning_rate": 1.9116161616161617e-05,
49
+ "loss": 0.1159,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.1,
54
+ "learning_rate": 1.8989898989898993e-05,
55
+ "loss": 0.1157,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.11,
60
+ "learning_rate": 1.8863636363636366e-05,
61
+ "loss": 0.1018,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.13,
66
+ "learning_rate": 1.873737373737374e-05,
67
+ "loss": 0.1111,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.14,
72
+ "learning_rate": 1.8611111111111114e-05,
73
+ "loss": 0.1432,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.15,
78
+ "learning_rate": 1.8484848484848487e-05,
79
+ "loss": 0.1059,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.16,
84
+ "learning_rate": 1.835858585858586e-05,
85
+ "loss": 0.1213,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.18,
90
+ "learning_rate": 1.8232323232323236e-05,
91
+ "loss": 0.1344,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.19,
96
+ "learning_rate": 1.810606060606061e-05,
97
+ "loss": 0.1174,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.2,
102
+ "learning_rate": 1.797979797979798e-05,
103
+ "loss": 0.1579,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.21,
108
+ "learning_rate": 1.7853535353535354e-05,
109
+ "loss": 0.109,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.23,
114
+ "learning_rate": 1.772727272727273e-05,
115
+ "loss": 0.1852,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.24,
120
+ "learning_rate": 1.7601010101010102e-05,
121
+ "loss": 0.0766,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.25,
126
+ "learning_rate": 1.7474747474747475e-05,
127
+ "loss": 0.1577,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.27,
132
+ "learning_rate": 1.734848484848485e-05,
133
+ "loss": 0.1063,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.28,
138
+ "learning_rate": 1.7222222222222224e-05,
139
+ "loss": 0.1002,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.29,
144
+ "learning_rate": 1.7095959595959596e-05,
145
+ "loss": 0.1122,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.3,
150
+ "learning_rate": 1.6969696969696972e-05,
151
+ "loss": 0.1283,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.32,
156
+ "learning_rate": 1.6843434343434345e-05,
157
+ "loss": 0.1163,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.33,
162
+ "learning_rate": 1.6717171717171718e-05,
163
+ "loss": 0.1092,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.34,
168
+ "learning_rate": 1.6590909090909094e-05,
169
+ "loss": 0.114,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.35,
174
+ "learning_rate": 1.6464646464646466e-05,
175
+ "loss": 0.0944,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.37,
180
+ "learning_rate": 1.633838383838384e-05,
181
+ "loss": 0.1325,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.38,
186
+ "learning_rate": 1.6212121212121212e-05,
187
+ "loss": 0.0885,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.39,
192
+ "learning_rate": 1.6085858585858588e-05,
193
+ "loss": 0.1479,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.4,
198
+ "learning_rate": 1.595959595959596e-05,
199
+ "loss": 0.135,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.42,
204
+ "learning_rate": 1.5833333333333333e-05,
205
+ "loss": 0.1261,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.43,
210
+ "learning_rate": 1.570707070707071e-05,
211
+ "loss": 0.0947,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.44,
216
+ "learning_rate": 1.5580808080808082e-05,
217
+ "loss": 0.1241,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.45,
222
+ "learning_rate": 1.5454545454545454e-05,
223
+ "loss": 0.1229,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.47,
228
+ "learning_rate": 1.532828282828283e-05,
229
+ "loss": 0.1149,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.48,
234
+ "learning_rate": 1.5202020202020203e-05,
235
+ "loss": 0.133,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.49,
240
+ "learning_rate": 1.5075757575757577e-05,
241
+ "loss": 0.1303,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.51,
246
+ "learning_rate": 1.4949494949494952e-05,
247
+ "loss": 0.0846,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.52,
252
+ "learning_rate": 1.4823232323232324e-05,
253
+ "loss": 0.0893,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.53,
258
+ "learning_rate": 1.4696969696969699e-05,
259
+ "loss": 0.0996,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.54,
264
+ "learning_rate": 1.4570707070707071e-05,
265
+ "loss": 0.0987,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.56,
270
+ "learning_rate": 1.4444444444444446e-05,
271
+ "loss": 0.1045,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.57,
276
+ "learning_rate": 1.431818181818182e-05,
277
+ "loss": 0.1004,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.58,
282
+ "learning_rate": 1.4191919191919193e-05,
283
+ "loss": 0.0917,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.59,
288
+ "learning_rate": 1.4065656565656567e-05,
289
+ "loss": 0.15,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.61,
294
+ "learning_rate": 1.3939393939393942e-05,
295
+ "loss": 0.0974,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.62,
300
+ "learning_rate": 1.3813131313131314e-05,
301
+ "loss": 0.1036,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.63,
306
+ "learning_rate": 1.3686868686868689e-05,
307
+ "loss": 0.1123,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.64,
312
+ "learning_rate": 1.3560606060606061e-05,
313
+ "loss": 0.0806,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.66,
318
+ "learning_rate": 1.3434343434343436e-05,
319
+ "loss": 0.1059,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.67,
324
+ "learning_rate": 1.330808080808081e-05,
325
+ "loss": 0.0984,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.68,
330
+ "learning_rate": 1.3181818181818183e-05,
331
+ "loss": 0.1285,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 0.69,
336
+ "learning_rate": 1.3055555555555557e-05,
337
+ "loss": 0.1089,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 0.71,
342
+ "learning_rate": 1.2929292929292931e-05,
343
+ "loss": 0.0928,
344
+ "step": 560
345
+ },
346
+ {
347
+ "epoch": 0.72,
348
+ "learning_rate": 1.2803030303030304e-05,
349
+ "loss": 0.0948,
350
+ "step": 570
351
+ },
352
+ {
353
+ "epoch": 0.73,
354
+ "learning_rate": 1.2676767676767678e-05,
355
+ "loss": 0.0934,
356
+ "step": 580
357
+ },
358
+ {
359
+ "epoch": 0.74,
360
+ "learning_rate": 1.2550505050505051e-05,
361
+ "loss": 0.094,
362
+ "step": 590
363
+ },
364
+ {
365
+ "epoch": 0.76,
366
+ "learning_rate": 1.2424242424242425e-05,
367
+ "loss": 0.0862,
368
+ "step": 600
369
+ },
370
+ {
371
+ "epoch": 0.77,
372
+ "learning_rate": 1.22979797979798e-05,
373
+ "loss": 0.0951,
374
+ "step": 610
375
+ },
376
+ {
377
+ "epoch": 0.78,
378
+ "learning_rate": 1.2171717171717172e-05,
379
+ "loss": 0.1,
380
+ "step": 620
381
+ },
382
+ {
383
+ "epoch": 0.8,
384
+ "learning_rate": 1.2045454545454547e-05,
385
+ "loss": 0.0789,
386
+ "step": 630
387
+ },
388
+ {
389
+ "epoch": 0.81,
390
+ "learning_rate": 1.191919191919192e-05,
391
+ "loss": 0.0726,
392
+ "step": 640
393
+ },
394
+ {
395
+ "epoch": 0.82,
396
+ "learning_rate": 1.1792929292929294e-05,
397
+ "loss": 0.1093,
398
+ "step": 650
399
+ },
400
+ {
401
+ "epoch": 0.83,
402
+ "learning_rate": 1.1666666666666668e-05,
403
+ "loss": 0.1078,
404
+ "step": 660
405
+ },
406
+ {
407
+ "epoch": 0.85,
408
+ "learning_rate": 1.154040404040404e-05,
409
+ "loss": 0.1226,
410
+ "step": 670
411
+ },
412
+ {
413
+ "epoch": 0.86,
414
+ "learning_rate": 1.1414141414141415e-05,
415
+ "loss": 0.0753,
416
+ "step": 680
417
+ },
418
+ {
419
+ "epoch": 0.87,
420
+ "learning_rate": 1.128787878787879e-05,
421
+ "loss": 0.0905,
422
+ "step": 690
423
+ },
424
+ {
425
+ "epoch": 0.88,
426
+ "learning_rate": 1.1161616161616162e-05,
427
+ "loss": 0.0887,
428
+ "step": 700
429
+ },
430
+ {
431
+ "epoch": 0.9,
432
+ "learning_rate": 1.1035353535353536e-05,
433
+ "loss": 0.1082,
434
+ "step": 710
435
+ },
436
+ {
437
+ "epoch": 0.91,
438
+ "learning_rate": 1.0909090909090909e-05,
439
+ "loss": 0.0809,
440
+ "step": 720
441
+ },
442
+ {
443
+ "epoch": 0.92,
444
+ "learning_rate": 1.0782828282828283e-05,
445
+ "loss": 0.1048,
446
+ "step": 730
447
+ },
448
+ {
449
+ "epoch": 0.93,
450
+ "learning_rate": 1.0656565656565658e-05,
451
+ "loss": 0.1034,
452
+ "step": 740
453
+ },
454
+ {
455
+ "epoch": 0.95,
456
+ "learning_rate": 1.053030303030303e-05,
457
+ "loss": 0.1425,
458
+ "step": 750
459
+ },
460
+ {
461
+ "epoch": 0.96,
462
+ "learning_rate": 1.0404040404040405e-05,
463
+ "loss": 0.0966,
464
+ "step": 760
465
+ },
466
+ {
467
+ "epoch": 0.97,
468
+ "learning_rate": 1.0277777777777777e-05,
469
+ "loss": 0.118,
470
+ "step": 770
471
+ },
472
+ {
473
+ "epoch": 0.98,
474
+ "learning_rate": 1.0151515151515152e-05,
475
+ "loss": 0.1258,
476
+ "step": 780
477
+ },
478
+ {
479
+ "epoch": 1.0,
480
+ "learning_rate": 1.0025252525252526e-05,
481
+ "loss": 0.1023,
482
+ "step": 790
483
+ },
484
+ {
485
+ "epoch": 1.01,
486
+ "learning_rate": 9.8989898989899e-06,
487
+ "loss": 0.0791,
488
+ "step": 800
489
+ },
490
+ {
491
+ "epoch": 1.02,
492
+ "learning_rate": 9.772727272727273e-06,
493
+ "loss": 0.0656,
494
+ "step": 810
495
+ },
496
+ {
497
+ "epoch": 1.04,
498
+ "learning_rate": 9.646464646464647e-06,
499
+ "loss": 0.0656,
500
+ "step": 820
501
+ },
502
+ {
503
+ "epoch": 1.05,
504
+ "learning_rate": 9.52020202020202e-06,
505
+ "loss": 0.0741,
506
+ "step": 830
507
+ },
508
+ {
509
+ "epoch": 1.06,
510
+ "learning_rate": 9.393939393939396e-06,
511
+ "loss": 0.0925,
512
+ "step": 840
513
+ },
514
+ {
515
+ "epoch": 1.07,
516
+ "learning_rate": 9.267676767676769e-06,
517
+ "loss": 0.0753,
518
+ "step": 850
519
+ },
520
+ {
521
+ "epoch": 1.09,
522
+ "learning_rate": 9.141414141414143e-06,
523
+ "loss": 0.0735,
524
+ "step": 860
525
+ },
526
+ {
527
+ "epoch": 1.1,
528
+ "learning_rate": 9.015151515151516e-06,
529
+ "loss": 0.0843,
530
+ "step": 870
531
+ },
532
+ {
533
+ "epoch": 1.11,
534
+ "learning_rate": 8.888888888888888e-06,
535
+ "loss": 0.0641,
536
+ "step": 880
537
+ },
538
+ {
539
+ "epoch": 1.12,
540
+ "learning_rate": 8.762626262626264e-06,
541
+ "loss": 0.078,
542
+ "step": 890
543
+ },
544
+ {
545
+ "epoch": 1.14,
546
+ "learning_rate": 8.636363636363637e-06,
547
+ "loss": 0.0819,
548
+ "step": 900
549
+ },
550
+ {
551
+ "epoch": 1.15,
552
+ "learning_rate": 8.510101010101011e-06,
553
+ "loss": 0.079,
554
+ "step": 910
555
+ },
556
+ {
557
+ "epoch": 1.16,
558
+ "learning_rate": 8.383838383838384e-06,
559
+ "loss": 0.0794,
560
+ "step": 920
561
+ },
562
+ {
563
+ "epoch": 1.17,
564
+ "learning_rate": 8.257575757575758e-06,
565
+ "loss": 0.0684,
566
+ "step": 930
567
+ },
568
+ {
569
+ "epoch": 1.19,
570
+ "learning_rate": 8.131313131313133e-06,
571
+ "loss": 0.0833,
572
+ "step": 940
573
+ },
574
+ {
575
+ "epoch": 1.2,
576
+ "learning_rate": 8.005050505050505e-06,
577
+ "loss": 0.079,
578
+ "step": 950
579
+ },
580
+ {
581
+ "epoch": 1.21,
582
+ "learning_rate": 7.87878787878788e-06,
583
+ "loss": 0.0675,
584
+ "step": 960
585
+ },
586
+ {
587
+ "epoch": 1.22,
588
+ "learning_rate": 7.752525252525254e-06,
589
+ "loss": 0.1101,
590
+ "step": 970
591
+ },
592
+ {
593
+ "epoch": 1.24,
594
+ "learning_rate": 7.6262626262626275e-06,
595
+ "loss": 0.072,
596
+ "step": 980
597
+ },
598
+ {
599
+ "epoch": 1.25,
600
+ "learning_rate": 7.500000000000001e-06,
601
+ "loss": 0.0754,
602
+ "step": 990
603
+ },
604
+ {
605
+ "epoch": 1.26,
606
+ "learning_rate": 7.373737373737374e-06,
607
+ "loss": 0.0807,
608
+ "step": 1000
609
+ },
610
+ {
611
+ "epoch": 1.28,
612
+ "learning_rate": 7.247474747474747e-06,
613
+ "loss": 0.0794,
614
+ "step": 1010
615
+ },
616
+ {
617
+ "epoch": 1.29,
618
+ "learning_rate": 7.121212121212122e-06,
619
+ "loss": 0.0866,
620
+ "step": 1020
621
+ },
622
+ {
623
+ "epoch": 1.3,
624
+ "learning_rate": 6.994949494949496e-06,
625
+ "loss": 0.0901,
626
+ "step": 1030
627
+ },
628
+ {
629
+ "epoch": 1.31,
630
+ "learning_rate": 6.868686868686869e-06,
631
+ "loss": 0.0676,
632
+ "step": 1040
633
+ },
634
+ {
635
+ "epoch": 1.33,
636
+ "learning_rate": 6.742424242424243e-06,
637
+ "loss": 0.0816,
638
+ "step": 1050
639
+ },
640
+ {
641
+ "epoch": 1.34,
642
+ "learning_rate": 6.616161616161617e-06,
643
+ "loss": 0.089,
644
+ "step": 1060
645
+ },
646
+ {
647
+ "epoch": 1.35,
648
+ "learning_rate": 6.489898989898991e-06,
649
+ "loss": 0.0831,
650
+ "step": 1070
651
+ },
652
+ {
653
+ "epoch": 1.36,
654
+ "learning_rate": 6.363636363636364e-06,
655
+ "loss": 0.0721,
656
+ "step": 1080
657
+ },
658
+ {
659
+ "epoch": 1.38,
660
+ "learning_rate": 6.237373737373738e-06,
661
+ "loss": 0.0743,
662
+ "step": 1090
663
+ },
664
+ {
665
+ "epoch": 1.39,
666
+ "learning_rate": 6.111111111111112e-06,
667
+ "loss": 0.0833,
668
+ "step": 1100
669
+ },
670
+ {
671
+ "epoch": 1.4,
672
+ "learning_rate": 5.984848484848486e-06,
673
+ "loss": 0.1008,
674
+ "step": 1110
675
+ },
676
+ {
677
+ "epoch": 1.41,
678
+ "learning_rate": 5.858585858585859e-06,
679
+ "loss": 0.1181,
680
+ "step": 1120
681
+ },
682
+ {
683
+ "epoch": 1.43,
684
+ "learning_rate": 5.732323232323233e-06,
685
+ "loss": 0.0634,
686
+ "step": 1130
687
+ },
688
+ {
689
+ "epoch": 1.44,
690
+ "learning_rate": 5.606060606060606e-06,
691
+ "loss": 0.0656,
692
+ "step": 1140
693
+ },
694
+ {
695
+ "epoch": 1.45,
696
+ "learning_rate": 5.4797979797979804e-06,
697
+ "loss": 0.0797,
698
+ "step": 1150
699
+ },
700
+ {
701
+ "epoch": 1.46,
702
+ "learning_rate": 5.353535353535354e-06,
703
+ "loss": 0.0644,
704
+ "step": 1160
705
+ },
706
+ {
707
+ "epoch": 1.48,
708
+ "learning_rate": 5.2272727272727274e-06,
709
+ "loss": 0.0605,
710
+ "step": 1170
711
+ },
712
+ {
713
+ "epoch": 1.49,
714
+ "learning_rate": 5.101010101010101e-06,
715
+ "loss": 0.0668,
716
+ "step": 1180
717
+ },
718
+ {
719
+ "epoch": 1.5,
720
+ "learning_rate": 4.974747474747475e-06,
721
+ "loss": 0.0638,
722
+ "step": 1190
723
+ },
724
+ {
725
+ "epoch": 1.52,
726
+ "learning_rate": 4.848484848484849e-06,
727
+ "loss": 0.0797,
728
+ "step": 1200
729
+ },
730
+ {
731
+ "epoch": 1.53,
732
+ "learning_rate": 4.722222222222222e-06,
733
+ "loss": 0.0941,
734
+ "step": 1210
735
+ },
736
+ {
737
+ "epoch": 1.54,
738
+ "learning_rate": 4.595959595959597e-06,
739
+ "loss": 0.0784,
740
+ "step": 1220
741
+ },
742
+ {
743
+ "epoch": 1.55,
744
+ "learning_rate": 4.46969696969697e-06,
745
+ "loss": 0.0561,
746
+ "step": 1230
747
+ },
748
+ {
749
+ "epoch": 1.57,
750
+ "learning_rate": 4.343434343434344e-06,
751
+ "loss": 0.0661,
752
+ "step": 1240
753
+ },
754
+ {
755
+ "epoch": 1.58,
756
+ "learning_rate": 4.217171717171717e-06,
757
+ "loss": 0.071,
758
+ "step": 1250
759
+ },
760
+ {
761
+ "epoch": 1.59,
762
+ "learning_rate": 4.0909090909090915e-06,
763
+ "loss": 0.0702,
764
+ "step": 1260
765
+ },
766
+ {
767
+ "epoch": 1.6,
768
+ "learning_rate": 3.964646464646465e-06,
769
+ "loss": 0.0702,
770
+ "step": 1270
771
+ },
772
+ {
773
+ "epoch": 1.62,
774
+ "learning_rate": 3.8383838383838385e-06,
775
+ "loss": 0.0795,
776
+ "step": 1280
777
+ },
778
+ {
779
+ "epoch": 1.63,
780
+ "learning_rate": 3.7121212121212124e-06,
781
+ "loss": 0.0627,
782
+ "step": 1290
783
+ },
784
+ {
785
+ "epoch": 1.64,
786
+ "learning_rate": 3.585858585858586e-06,
787
+ "loss": 0.086,
788
+ "step": 1300
789
+ },
790
+ {
791
+ "epoch": 1.65,
792
+ "learning_rate": 3.45959595959596e-06,
793
+ "loss": 0.0567,
794
+ "step": 1310
795
+ },
796
+ {
797
+ "epoch": 1.67,
798
+ "learning_rate": 3.3333333333333333e-06,
799
+ "loss": 0.0764,
800
+ "step": 1320
801
+ },
802
+ {
803
+ "epoch": 1.68,
804
+ "learning_rate": 3.2070707070707072e-06,
805
+ "loss": 0.0576,
806
+ "step": 1330
807
+ },
808
+ {
809
+ "epoch": 1.69,
810
+ "learning_rate": 3.0808080808080807e-06,
811
+ "loss": 0.0783,
812
+ "step": 1340
813
+ },
814
+ {
815
+ "epoch": 1.7,
816
+ "learning_rate": 2.954545454545455e-06,
817
+ "loss": 0.0649,
818
+ "step": 1350
819
+ },
820
+ {
821
+ "epoch": 1.72,
822
+ "learning_rate": 2.8282828282828286e-06,
823
+ "loss": 0.0756,
824
+ "step": 1360
825
+ },
826
+ {
827
+ "epoch": 1.73,
828
+ "learning_rate": 2.7020202020202025e-06,
829
+ "loss": 0.0655,
830
+ "step": 1370
831
+ },
832
+ {
833
+ "epoch": 1.74,
834
+ "learning_rate": 2.575757575757576e-06,
835
+ "loss": 0.0664,
836
+ "step": 1380
837
+ },
838
+ {
839
+ "epoch": 1.76,
840
+ "learning_rate": 2.4494949494949495e-06,
841
+ "loss": 0.0814,
842
+ "step": 1390
843
+ },
844
+ {
845
+ "epoch": 1.77,
846
+ "learning_rate": 2.3232323232323234e-06,
847
+ "loss": 0.0612,
848
+ "step": 1400
849
+ },
850
+ {
851
+ "epoch": 1.78,
852
+ "learning_rate": 2.196969696969697e-06,
853
+ "loss": 0.0664,
854
+ "step": 1410
855
+ },
856
+ {
857
+ "epoch": 1.79,
858
+ "learning_rate": 2.070707070707071e-06,
859
+ "loss": 0.0484,
860
+ "step": 1420
861
+ },
862
+ {
863
+ "epoch": 1.81,
864
+ "learning_rate": 1.944444444444445e-06,
865
+ "loss": 0.0733,
866
+ "step": 1430
867
+ },
868
+ {
869
+ "epoch": 1.82,
870
+ "learning_rate": 1.8181818181818183e-06,
871
+ "loss": 0.0644,
872
+ "step": 1440
873
+ },
874
+ {
875
+ "epoch": 1.83,
876
+ "learning_rate": 1.6919191919191922e-06,
877
+ "loss": 0.0675,
878
+ "step": 1450
879
+ },
880
+ {
881
+ "epoch": 1.84,
882
+ "learning_rate": 1.565656565656566e-06,
883
+ "loss": 0.0749,
884
+ "step": 1460
885
+ },
886
+ {
887
+ "epoch": 1.86,
888
+ "learning_rate": 1.4393939393939396e-06,
889
+ "loss": 0.0604,
890
+ "step": 1470
891
+ },
892
+ {
893
+ "epoch": 1.87,
894
+ "learning_rate": 1.3131313131313134e-06,
895
+ "loss": 0.074,
896
+ "step": 1480
897
+ },
898
+ {
899
+ "epoch": 1.88,
900
+ "learning_rate": 1.186868686868687e-06,
901
+ "loss": 0.0808,
902
+ "step": 1490
903
+ },
904
+ {
905
+ "epoch": 1.89,
906
+ "learning_rate": 1.0606060606060608e-06,
907
+ "loss": 0.0812,
908
+ "step": 1500
909
+ }
910
+ ],
911
+ "max_steps": 1584,
912
+ "num_train_epochs": 2,
913
+ "total_flos": 2785192250966016.0,
914
+ "trial_name": null,
915
+ "trial_params": null
916
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:267b4d9e52d87b26a96cb7c13b7efeb8b4f292335ab2f8d8de64583b64cf5683
3
+ size 3960