MarkProMaster229 commited on
Commit
d1efaad
·
verified ·
1 Parent(s): 66499e1

Delete trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +0 -412
trainer_state.json DELETED
@@ -1,412 +0,0 @@
1
- {
2
- "best_global_step": null,
3
- "best_metric": null,
4
- "best_model_checkpoint": null,
5
- "epoch": 0.6827233074151337,
6
- "eval_steps": 500,
7
- "global_step": 5400,
8
- "is_hyper_param_search": false,
9
- "is_local_process_zero": true,
10
- "is_world_process_zero": true,
11
- "log_history": [
12
- {
13
- "epoch": 0.012643024211391365,
14
- "grad_norm": 0.8759542107582092,
15
- "learning_rate": 0.0004987484197218711,
16
- "loss": 3.8722,
17
- "step": 100
18
- },
19
- {
20
- "epoch": 0.02528604842278273,
21
- "grad_norm": 0.9236809015274048,
22
- "learning_rate": 0.0004974841972187105,
23
- "loss": 3.4858,
24
- "step": 200
25
- },
26
- {
27
- "epoch": 0.037929072634174096,
28
- "grad_norm": 0.8505849242210388,
29
- "learning_rate": 0.0004962199747155499,
30
- "loss": 3.3438,
31
- "step": 300
32
- },
33
- {
34
- "epoch": 0.05057209684556546,
35
- "grad_norm": 0.8044902682304382,
36
- "learning_rate": 0.0004949557522123893,
37
- "loss": 3.2837,
38
- "step": 400
39
- },
40
- {
41
- "epoch": 0.06321512105695683,
42
- "grad_norm": 0.7873915433883667,
43
- "learning_rate": 0.0004936915297092288,
44
- "loss": 3.194,
45
- "step": 500
46
- },
47
- {
48
- "epoch": 0.07585814526834819,
49
- "grad_norm": 0.7622674107551575,
50
- "learning_rate": 0.0004924273072060682,
51
- "loss": 3.126,
52
- "step": 600
53
- },
54
- {
55
- "epoch": 0.08850116947973956,
56
- "grad_norm": 0.8418383002281189,
57
- "learning_rate": 0.0004911630847029077,
58
- "loss": 3.0518,
59
- "step": 700
60
- },
61
- {
62
- "epoch": 0.10114419369113092,
63
- "grad_norm": 0.7434802055358887,
64
- "learning_rate": 0.0004898988621997471,
65
- "loss": 3.0434,
66
- "step": 800
67
- },
68
- {
69
- "epoch": 0.11378721790252229,
70
- "grad_norm": 0.8024940490722656,
71
- "learning_rate": 0.0004886346396965867,
72
- "loss": 2.9942,
73
- "step": 900
74
- },
75
- {
76
- "epoch": 0.12643024211391365,
77
- "grad_norm": 0.8081286549568176,
78
- "learning_rate": 0.00048737041719342606,
79
- "loss": 2.9878,
80
- "step": 1000
81
- },
82
- {
83
- "epoch": 0.139073266325305,
84
- "grad_norm": 0.7084025144577026,
85
- "learning_rate": 0.0004861061946902655,
86
- "loss": 2.9314,
87
- "step": 1100
88
- },
89
- {
90
- "epoch": 0.15171629053669639,
91
- "grad_norm": 0.7388598322868347,
92
- "learning_rate": 0.000484841972187105,
93
- "loss": 2.9152,
94
- "step": 1200
95
- },
96
- {
97
- "epoch": 0.16435931474808774,
98
- "grad_norm": 0.7991167306900024,
99
- "learning_rate": 0.0004835777496839444,
100
- "loss": 2.917,
101
- "step": 1300
102
- },
103
- {
104
- "epoch": 0.17700233895947912,
105
- "grad_norm": 0.7912219762802124,
106
- "learning_rate": 0.0004823135271807838,
107
- "loss": 2.8725,
108
- "step": 1400
109
- },
110
- {
111
- "epoch": 0.18964536317087047,
112
- "grad_norm": 0.8445726633071899,
113
- "learning_rate": 0.00048104930467762324,
114
- "loss": 2.8843,
115
- "step": 1500
116
- },
117
- {
118
- "epoch": 0.20228838738226185,
119
- "grad_norm": 0.7209933400154114,
120
- "learning_rate": 0.0004797850821744627,
121
- "loss": 2.8298,
122
- "step": 1600
123
- },
124
- {
125
- "epoch": 0.2149314115936532,
126
- "grad_norm": 0.7905689477920532,
127
- "learning_rate": 0.00047852085967130215,
128
- "loss": 2.862,
129
- "step": 1700
130
- },
131
- {
132
- "epoch": 0.22757443580504458,
133
- "grad_norm": 0.745158314704895,
134
- "learning_rate": 0.0004772566371681416,
135
- "loss": 2.781,
136
- "step": 1800
137
- },
138
- {
139
- "epoch": 0.24021746001643593,
140
- "grad_norm": 0.7118976712226868,
141
- "learning_rate": 0.00047599241466498107,
142
- "loss": 2.7783,
143
- "step": 1900
144
- },
145
- {
146
- "epoch": 0.2528604842278273,
147
- "grad_norm": 0.7946869730949402,
148
- "learning_rate": 0.0004747281921618205,
149
- "loss": 2.7825,
150
- "step": 2000
151
- },
152
- {
153
- "epoch": 0.26550350843921866,
154
- "grad_norm": 0.7247060537338257,
155
- "learning_rate": 0.00047346396965865993,
156
- "loss": 2.7839,
157
- "step": 2100
158
- },
159
- {
160
- "epoch": 0.27814653265061,
161
- "grad_norm": 0.7256483435630798,
162
- "learning_rate": 0.0004721997471554994,
163
- "loss": 2.7731,
164
- "step": 2200
165
- },
166
- {
167
- "epoch": 0.29078955686200136,
168
- "grad_norm": 0.7218326926231384,
169
- "learning_rate": 0.0004709355246523388,
170
- "loss": 2.8133,
171
- "step": 2300
172
- },
173
- {
174
- "epoch": 0.30343258107339277,
175
- "grad_norm": 0.7010550498962402,
176
- "learning_rate": 0.00046967130214917825,
177
- "loss": 2.7432,
178
- "step": 2400
179
- },
180
- {
181
- "epoch": 0.3160756052847841,
182
- "grad_norm": 0.7964794635772705,
183
- "learning_rate": 0.0004684070796460177,
184
- "loss": 2.7811,
185
- "step": 2500
186
- },
187
- {
188
- "epoch": 0.3287186294961755,
189
- "grad_norm": 0.8072954416275024,
190
- "learning_rate": 0.00046714285714285716,
191
- "loss": 2.7089,
192
- "step": 2600
193
- },
194
- {
195
- "epoch": 0.3413616537075668,
196
- "grad_norm": 0.6594070196151733,
197
- "learning_rate": 0.0004658786346396966,
198
- "loss": 2.7161,
199
- "step": 2700
200
- },
201
- {
202
- "epoch": 0.35400467791895823,
203
- "grad_norm": 0.704298734664917,
204
- "learning_rate": 0.000464614412136536,
205
- "loss": 2.698,
206
- "step": 2800
207
- },
208
- {
209
- "epoch": 0.3666477021303496,
210
- "grad_norm": 0.7253355383872986,
211
- "learning_rate": 0.0004633501896333755,
212
- "loss": 2.696,
213
- "step": 2900
214
- },
215
- {
216
- "epoch": 0.37929072634174094,
217
- "grad_norm": 0.7043545246124268,
218
- "learning_rate": 0.00046208596713021493,
219
- "loss": 2.6807,
220
- "step": 3000
221
- },
222
- {
223
- "epoch": 0.3919337505531323,
224
- "grad_norm": 0.6532794237136841,
225
- "learning_rate": 0.0004608217446270544,
226
- "loss": 2.6985,
227
- "step": 3100
228
- },
229
- {
230
- "epoch": 0.4045767747645237,
231
- "grad_norm": 0.7272788286209106,
232
- "learning_rate": 0.0004595575221238938,
233
- "loss": 2.6767,
234
- "step": 3200
235
- },
236
- {
237
- "epoch": 0.41721979897591505,
238
- "grad_norm": 0.695071280002594,
239
- "learning_rate": 0.00045829329962073325,
240
- "loss": 2.6609,
241
- "step": 3300
242
- },
243
- {
244
- "epoch": 0.4298628231873064,
245
- "grad_norm": 0.7230761051177979,
246
- "learning_rate": 0.0004570290771175727,
247
- "loss": 2.6488,
248
- "step": 3400
249
- },
250
- {
251
- "epoch": 0.44250584739869775,
252
- "grad_norm": 0.7420136332511902,
253
- "learning_rate": 0.00045576485461441217,
254
- "loss": 2.6507,
255
- "step": 3500
256
- },
257
- {
258
- "epoch": 0.45514887161008916,
259
- "grad_norm": 0.7115824222564697,
260
- "learning_rate": 0.00045450063211125157,
261
- "loss": 2.644,
262
- "step": 3600
263
- },
264
- {
265
- "epoch": 0.4677918958214805,
266
- "grad_norm": 0.6667810678482056,
267
- "learning_rate": 0.000453236409608091,
268
- "loss": 2.6841,
269
- "step": 3700
270
- },
271
- {
272
- "epoch": 0.48043492003287186,
273
- "grad_norm": 0.6836283802986145,
274
- "learning_rate": 0.0004519721871049305,
275
- "loss": 2.6462,
276
- "step": 3800
277
- },
278
- {
279
- "epoch": 0.4930779442442632,
280
- "grad_norm": 0.7117214202880859,
281
- "learning_rate": 0.00045070796460176994,
282
- "loss": 2.6201,
283
- "step": 3900
284
- },
285
- {
286
- "epoch": 0.5057209684556546,
287
- "grad_norm": 0.6085230708122253,
288
- "learning_rate": 0.0004494437420986094,
289
- "loss": 2.6198,
290
- "step": 4000
291
- },
292
- {
293
- "epoch": 0.5183639926670459,
294
- "grad_norm": 0.663446843624115,
295
- "learning_rate": 0.0004481795195954488,
296
- "loss": 2.5972,
297
- "step": 4100
298
- },
299
- {
300
- "epoch": 0.5310070168784373,
301
- "grad_norm": 0.670093297958374,
302
- "learning_rate": 0.00044691529709228826,
303
- "loss": 2.6052,
304
- "step": 4200
305
- },
306
- {
307
- "epoch": 0.5436500410898287,
308
- "grad_norm": 0.6052363514900208,
309
- "learning_rate": 0.00044565107458912766,
310
- "loss": 2.6038,
311
- "step": 4300
312
- },
313
- {
314
- "epoch": 0.55629306530122,
315
- "grad_norm": 0.6686172485351562,
316
- "learning_rate": 0.0004443868520859671,
317
- "loss": 2.5484,
318
- "step": 4400
319
- },
320
- {
321
- "epoch": 0.5689360895126114,
322
- "grad_norm": 0.6228762865066528,
323
- "learning_rate": 0.0004431226295828066,
324
- "loss": 2.6119,
325
- "step": 4500
326
- },
327
- {
328
- "epoch": 0.5815791137240027,
329
- "grad_norm": 0.6712014079093933,
330
- "learning_rate": 0.00044185840707964603,
331
- "loss": 2.581,
332
- "step": 4600
333
- },
334
- {
335
- "epoch": 0.5942221379353941,
336
- "grad_norm": 0.6657222509384155,
337
- "learning_rate": 0.0004405941845764855,
338
- "loss": 2.5822,
339
- "step": 4700
340
- },
341
- {
342
- "epoch": 0.6068651621467855,
343
- "grad_norm": 0.639202356338501,
344
- "learning_rate": 0.00043932996207332494,
345
- "loss": 2.5736,
346
- "step": 4800
347
- },
348
- {
349
- "epoch": 0.6195081863581768,
350
- "grad_norm": 0.654742419719696,
351
- "learning_rate": 0.0004380657395701644,
352
- "loss": 2.5515,
353
- "step": 4900
354
- },
355
- {
356
- "epoch": 0.6321512105695682,
357
- "grad_norm": 0.704134464263916,
358
- "learning_rate": 0.0004368015170670038,
359
- "loss": 2.5499,
360
- "step": 5000
361
- },
362
- {
363
- "epoch": 0.6447942347809597,
364
- "grad_norm": 0.6817001104354858,
365
- "learning_rate": 0.0004355372945638432,
366
- "loss": 2.611,
367
- "step": 5100
368
- },
369
- {
370
- "epoch": 0.657437258992351,
371
- "grad_norm": 0.6351118087768555,
372
- "learning_rate": 0.00043427307206068266,
373
- "loss": 2.566,
374
- "step": 5200
375
- },
376
- {
377
- "epoch": 0.6700802832037424,
378
- "grad_norm": 0.6755563020706177,
379
- "learning_rate": 0.0004330088495575221,
380
- "loss": 2.5771,
381
- "step": 5300
382
- },
383
- {
384
- "epoch": 0.6827233074151337,
385
- "grad_norm": 0.6010642647743225,
386
- "learning_rate": 0.0004317446270543616,
387
- "loss": 2.5216,
388
- "step": 5400
389
- }
390
- ],
391
- "logging_steps": 100,
392
- "max_steps": 39550,
393
- "num_input_tokens_seen": 0,
394
- "num_train_epochs": 5,
395
- "save_steps": 300,
396
- "stateful_callbacks": {
397
- "TrainerControl": {
398
- "args": {
399
- "should_epoch_stop": false,
400
- "should_evaluate": false,
401
- "should_log": false,
402
- "should_save": true,
403
- "should_training_stop": false
404
- },
405
- "attributes": {}
406
- }
407
- },
408
- "total_flos": 4.2329309184e+16,
409
- "train_batch_size": 15,
410
- "trial_name": null,
411
- "trial_params": null
412
- }