sedrickkeh commited on
Commit
d3f7745
·
verified ·
1 Parent(s): 996e80d

End of training

Browse files
Files changed (5) hide show
  1. README.md +2 -1
  2. all_results.json +8 -0
  3. train_results.json +8 -0
  4. trainer_state.json +1092 -0
  5. training_loss.png +0 -0
README.md CHANGED
@@ -4,6 +4,7 @@ license: apache-2.0
4
  base_model: Qwen/Qwen2.5-7B-Instruct
5
  tags:
6
  - llama-factory
 
7
  - generated_from_trainer
8
  model-index:
9
  - name: 316_globalbatchsize32_lr1e5_epochs15
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # 316_globalbatchsize32_lr1e5_epochs15
17
 
18
- This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on an unknown dataset.
19
 
20
  ## Model description
21
 
 
4
  base_model: Qwen/Qwen2.5-7B-Instruct
5
  tags:
6
  - llama-factory
7
+ - full
8
  - generated_from_trainer
9
  model-index:
10
  - name: 316_globalbatchsize32_lr1e5_epochs15
 
16
 
17
  # 316_globalbatchsize32_lr1e5_epochs15
18
 
19
+ This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on the mlfoundations-dev/openthoughts_316 dataset.
20
 
21
  ## Model description
22
 
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 8.481547119440691e+16,
4
+ "train_loss": 0.2379360094014555,
5
+ "train_runtime": 6253.349,
6
+ "train_samples_per_second": 0.758,
7
+ "train_steps_per_second": 0.024
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 8.481547119440691e+16,
4
+ "train_loss": 0.2379360094014555,
5
+ "train_runtime": 6253.349,
6
+ "train_samples_per_second": 0.758,
7
+ "train_steps_per_second": 0.024
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1092 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 15.0,
5
+ "eval_steps": 500,
6
+ "global_step": 150,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.1,
13
+ "grad_norm": 5.45644890068779,
14
+ "learning_rate": 6.666666666666667e-07,
15
+ "loss": 0.8054,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.2,
20
+ "grad_norm": 5.760022753217485,
21
+ "learning_rate": 1.3333333333333334e-06,
22
+ "loss": 0.8566,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.3,
27
+ "grad_norm": 5.703638267473252,
28
+ "learning_rate": 2.0000000000000003e-06,
29
+ "loss": 0.8626,
30
+ "step": 3
31
+ },
32
+ {
33
+ "epoch": 0.4,
34
+ "grad_norm": 6.110963969550015,
35
+ "learning_rate": 2.666666666666667e-06,
36
+ "loss": 0.9238,
37
+ "step": 4
38
+ },
39
+ {
40
+ "epoch": 0.5,
41
+ "grad_norm": 4.761643788756697,
42
+ "learning_rate": 3.3333333333333333e-06,
43
+ "loss": 0.8581,
44
+ "step": 5
45
+ },
46
+ {
47
+ "epoch": 0.6,
48
+ "grad_norm": 4.29771328391321,
49
+ "learning_rate": 4.000000000000001e-06,
50
+ "loss": 0.8681,
51
+ "step": 6
52
+ },
53
+ {
54
+ "epoch": 0.7,
55
+ "grad_norm": 2.4756801117534217,
56
+ "learning_rate": 4.666666666666667e-06,
57
+ "loss": 0.8289,
58
+ "step": 7
59
+ },
60
+ {
61
+ "epoch": 0.8,
62
+ "grad_norm": 2.2144200581350812,
63
+ "learning_rate": 5.333333333333334e-06,
64
+ "loss": 0.8244,
65
+ "step": 8
66
+ },
67
+ {
68
+ "epoch": 0.9,
69
+ "grad_norm": 3.912901318058458,
70
+ "learning_rate": 6e-06,
71
+ "loss": 0.7957,
72
+ "step": 9
73
+ },
74
+ {
75
+ "epoch": 1.0,
76
+ "grad_norm": 4.165778131324685,
77
+ "learning_rate": 6.666666666666667e-06,
78
+ "loss": 0.7969,
79
+ "step": 10
80
+ },
81
+ {
82
+ "epoch": 1.1,
83
+ "grad_norm": 4.093578392857469,
84
+ "learning_rate": 7.333333333333333e-06,
85
+ "loss": 0.7222,
86
+ "step": 11
87
+ },
88
+ {
89
+ "epoch": 1.2,
90
+ "grad_norm": 4.585686741760957,
91
+ "learning_rate": 8.000000000000001e-06,
92
+ "loss": 0.7576,
93
+ "step": 12
94
+ },
95
+ {
96
+ "epoch": 1.3,
97
+ "grad_norm": 4.227643424601122,
98
+ "learning_rate": 8.666666666666668e-06,
99
+ "loss": 0.7494,
100
+ "step": 13
101
+ },
102
+ {
103
+ "epoch": 1.4,
104
+ "grad_norm": 3.0915284393682256,
105
+ "learning_rate": 9.333333333333334e-06,
106
+ "loss": 0.7361,
107
+ "step": 14
108
+ },
109
+ {
110
+ "epoch": 1.5,
111
+ "grad_norm": 2.012031527587278,
112
+ "learning_rate": 1e-05,
113
+ "loss": 0.6901,
114
+ "step": 15
115
+ },
116
+ {
117
+ "epoch": 1.6,
118
+ "grad_norm": 1.825988497662613,
119
+ "learning_rate": 9.99864620589731e-06,
120
+ "loss": 0.6738,
121
+ "step": 16
122
+ },
123
+ {
124
+ "epoch": 1.7,
125
+ "grad_norm": 2.0220122087316508,
126
+ "learning_rate": 9.994585556692624e-06,
127
+ "loss": 0.6613,
128
+ "step": 17
129
+ },
130
+ {
131
+ "epoch": 1.8,
132
+ "grad_norm": 1.9971123511586695,
133
+ "learning_rate": 9.987820251299121e-06,
134
+ "loss": 0.6795,
135
+ "step": 18
136
+ },
137
+ {
138
+ "epoch": 1.9,
139
+ "grad_norm": 1.3354562527771812,
140
+ "learning_rate": 9.978353953249023e-06,
141
+ "loss": 0.636,
142
+ "step": 19
143
+ },
144
+ {
145
+ "epoch": 2.0,
146
+ "grad_norm": 1.2093863766066717,
147
+ "learning_rate": 9.966191788709716e-06,
148
+ "loss": 0.6093,
149
+ "step": 20
150
+ },
151
+ {
152
+ "epoch": 2.1,
153
+ "grad_norm": 1.4696690338986078,
154
+ "learning_rate": 9.951340343707852e-06,
155
+ "loss": 0.6346,
156
+ "step": 21
157
+ },
158
+ {
159
+ "epoch": 2.2,
160
+ "grad_norm": 1.3392011179138974,
161
+ "learning_rate": 9.933807660562898e-06,
162
+ "loss": 0.5936,
163
+ "step": 22
164
+ },
165
+ {
166
+ "epoch": 2.3,
167
+ "grad_norm": 0.9312569898589178,
168
+ "learning_rate": 9.913603233532067e-06,
169
+ "loss": 0.5628,
170
+ "step": 23
171
+ },
172
+ {
173
+ "epoch": 2.4,
174
+ "grad_norm": 0.9584514112385119,
175
+ "learning_rate": 9.890738003669029e-06,
176
+ "loss": 0.5589,
177
+ "step": 24
178
+ },
179
+ {
180
+ "epoch": 2.5,
181
+ "grad_norm": 1.1235166057418648,
182
+ "learning_rate": 9.86522435289912e-06,
183
+ "loss": 0.5685,
184
+ "step": 25
185
+ },
186
+ {
187
+ "epoch": 2.6,
188
+ "grad_norm": 1.0208228551871161,
189
+ "learning_rate": 9.83707609731432e-06,
190
+ "loss": 0.5398,
191
+ "step": 26
192
+ },
193
+ {
194
+ "epoch": 2.7,
195
+ "grad_norm": 0.9055516223006299,
196
+ "learning_rate": 9.806308479691595e-06,
197
+ "loss": 0.5764,
198
+ "step": 27
199
+ },
200
+ {
201
+ "epoch": 2.8,
202
+ "grad_norm": 0.9758425594133711,
203
+ "learning_rate": 9.77293816123866e-06,
204
+ "loss": 0.5818,
205
+ "step": 28
206
+ },
207
+ {
208
+ "epoch": 2.9,
209
+ "grad_norm": 1.0854268157351894,
210
+ "learning_rate": 9.736983212571646e-06,
211
+ "loss": 0.5074,
212
+ "step": 29
213
+ },
214
+ {
215
+ "epoch": 3.0,
216
+ "grad_norm": 1.0002476303764538,
217
+ "learning_rate": 9.698463103929542e-06,
218
+ "loss": 0.5713,
219
+ "step": 30
220
+ },
221
+ {
222
+ "epoch": 3.1,
223
+ "grad_norm": 0.7994843300609084,
224
+ "learning_rate": 9.657398694630713e-06,
225
+ "loss": 0.4694,
226
+ "step": 31
227
+ },
228
+ {
229
+ "epoch": 3.2,
230
+ "grad_norm": 0.7551036719358174,
231
+ "learning_rate": 9.613812221777212e-06,
232
+ "loss": 0.4375,
233
+ "step": 32
234
+ },
235
+ {
236
+ "epoch": 3.3,
237
+ "grad_norm": 0.7907253429782667,
238
+ "learning_rate": 9.567727288213005e-06,
239
+ "loss": 0.4821,
240
+ "step": 33
241
+ },
242
+ {
243
+ "epoch": 3.4,
244
+ "grad_norm": 0.98432961387585,
245
+ "learning_rate": 9.519168849742603e-06,
246
+ "loss": 0.5182,
247
+ "step": 34
248
+ },
249
+ {
250
+ "epoch": 3.5,
251
+ "grad_norm": 0.7323725774105448,
252
+ "learning_rate": 9.468163201617063e-06,
253
+ "loss": 0.4632,
254
+ "step": 35
255
+ },
256
+ {
257
+ "epoch": 3.6,
258
+ "grad_norm": 0.843682270627834,
259
+ "learning_rate": 9.414737964294636e-06,
260
+ "loss": 0.4471,
261
+ "step": 36
262
+ },
263
+ {
264
+ "epoch": 3.7,
265
+ "grad_norm": 0.818957704445535,
266
+ "learning_rate": 9.358922068483813e-06,
267
+ "loss": 0.4481,
268
+ "step": 37
269
+ },
270
+ {
271
+ "epoch": 3.8,
272
+ "grad_norm": 0.8434619149903689,
273
+ "learning_rate": 9.30074573947683e-06,
274
+ "loss": 0.447,
275
+ "step": 38
276
+ },
277
+ {
278
+ "epoch": 3.9,
279
+ "grad_norm": 0.6875715680565743,
280
+ "learning_rate": 9.24024048078213e-06,
281
+ "loss": 0.3888,
282
+ "step": 39
283
+ },
284
+ {
285
+ "epoch": 4.0,
286
+ "grad_norm": 0.6709858781048713,
287
+ "learning_rate": 9.177439057064684e-06,
288
+ "loss": 0.3943,
289
+ "step": 40
290
+ },
291
+ {
292
+ "epoch": 4.1,
293
+ "grad_norm": 0.8064840728512959,
294
+ "learning_rate": 9.112375476403313e-06,
295
+ "loss": 0.3847,
296
+ "step": 41
297
+ },
298
+ {
299
+ "epoch": 4.2,
300
+ "grad_norm": 0.6969019084071135,
301
+ "learning_rate": 9.045084971874738e-06,
302
+ "loss": 0.3941,
303
+ "step": 42
304
+ },
305
+ {
306
+ "epoch": 4.3,
307
+ "grad_norm": 0.6320772903365803,
308
+ "learning_rate": 8.97560398247424e-06,
309
+ "loss": 0.3552,
310
+ "step": 43
311
+ },
312
+ {
313
+ "epoch": 4.4,
314
+ "grad_norm": 0.7417550318411654,
315
+ "learning_rate": 8.903970133383297e-06,
316
+ "loss": 0.3118,
317
+ "step": 44
318
+ },
319
+ {
320
+ "epoch": 4.5,
321
+ "grad_norm": 0.7792874741084093,
322
+ "learning_rate": 8.83022221559489e-06,
323
+ "loss": 0.3311,
324
+ "step": 45
325
+ },
326
+ {
327
+ "epoch": 4.6,
328
+ "grad_norm": 0.7077969057934511,
329
+ "learning_rate": 8.754400164907496e-06,
330
+ "loss": 0.3236,
331
+ "step": 46
332
+ },
333
+ {
334
+ "epoch": 4.7,
335
+ "grad_norm": 0.7353627300120197,
336
+ "learning_rate": 8.676545040299145e-06,
337
+ "loss": 0.3548,
338
+ "step": 47
339
+ },
340
+ {
341
+ "epoch": 4.8,
342
+ "grad_norm": 0.7124912659874838,
343
+ "learning_rate": 8.596699001693257e-06,
344
+ "loss": 0.3308,
345
+ "step": 48
346
+ },
347
+ {
348
+ "epoch": 4.9,
349
+ "grad_norm": 0.7137026642495771,
350
+ "learning_rate": 8.51490528712831e-06,
351
+ "loss": 0.3279,
352
+ "step": 49
353
+ },
354
+ {
355
+ "epoch": 5.0,
356
+ "grad_norm": 0.6560846796125491,
357
+ "learning_rate": 8.43120818934367e-06,
358
+ "loss": 0.3142,
359
+ "step": 50
360
+ },
361
+ {
362
+ "epoch": 5.1,
363
+ "grad_norm": 0.7317802945089542,
364
+ "learning_rate": 8.345653031794292e-06,
365
+ "loss": 0.3004,
366
+ "step": 51
367
+ },
368
+ {
369
+ "epoch": 5.2,
370
+ "grad_norm": 0.7830121121881166,
371
+ "learning_rate": 8.258286144107277e-06,
372
+ "loss": 0.2591,
373
+ "step": 52
374
+ },
375
+ {
376
+ "epoch": 5.3,
377
+ "grad_norm": 0.980706179145974,
378
+ "learning_rate": 8.16915483699355e-06,
379
+ "loss": 0.2398,
380
+ "step": 53
381
+ },
382
+ {
383
+ "epoch": 5.4,
384
+ "grad_norm": 0.8575321977363193,
385
+ "learning_rate": 8.078307376628292e-06,
386
+ "loss": 0.2489,
387
+ "step": 54
388
+ },
389
+ {
390
+ "epoch": 5.5,
391
+ "grad_norm": 0.6115709560784736,
392
+ "learning_rate": 7.985792958513932e-06,
393
+ "loss": 0.2515,
394
+ "step": 55
395
+ },
396
+ {
397
+ "epoch": 5.6,
398
+ "grad_norm": 0.7740033959246835,
399
+ "learning_rate": 7.891661680839932e-06,
400
+ "loss": 0.2566,
401
+ "step": 56
402
+ },
403
+ {
404
+ "epoch": 5.7,
405
+ "grad_norm": 0.714778018580984,
406
+ "learning_rate": 7.795964517353734e-06,
407
+ "loss": 0.2252,
408
+ "step": 57
409
+ },
410
+ {
411
+ "epoch": 5.8,
412
+ "grad_norm": 0.7247493394530297,
413
+ "learning_rate": 7.698753289757565e-06,
414
+ "loss": 0.2198,
415
+ "step": 58
416
+ },
417
+ {
418
+ "epoch": 5.9,
419
+ "grad_norm": 0.8011021603411893,
420
+ "learning_rate": 7.600080639646077e-06,
421
+ "loss": 0.2478,
422
+ "step": 59
423
+ },
424
+ {
425
+ "epoch": 6.0,
426
+ "grad_norm": 0.6822418140115627,
427
+ "learning_rate": 7.500000000000001e-06,
428
+ "loss": 0.1873,
429
+ "step": 60
430
+ },
431
+ {
432
+ "epoch": 6.1,
433
+ "grad_norm": 0.884809770268133,
434
+ "learning_rate": 7.398565566251232e-06,
435
+ "loss": 0.1718,
436
+ "step": 61
437
+ },
438
+ {
439
+ "epoch": 6.2,
440
+ "grad_norm": 0.833226691162642,
441
+ "learning_rate": 7.295832266935059e-06,
442
+ "loss": 0.1794,
443
+ "step": 62
444
+ },
445
+ {
446
+ "epoch": 6.3,
447
+ "grad_norm": 0.7859639404871255,
448
+ "learning_rate": 7.191855733945388e-06,
449
+ "loss": 0.1628,
450
+ "step": 63
451
+ },
452
+ {
453
+ "epoch": 6.4,
454
+ "grad_norm": 1.4787261653683548,
455
+ "learning_rate": 7.08669227240909e-06,
456
+ "loss": 0.2054,
457
+ "step": 64
458
+ },
459
+ {
460
+ "epoch": 6.5,
461
+ "grad_norm": 0.7182243781352645,
462
+ "learning_rate": 6.980398830195785e-06,
463
+ "loss": 0.1307,
464
+ "step": 65
465
+ },
466
+ {
467
+ "epoch": 6.6,
468
+ "grad_norm": 0.7417671974325057,
469
+ "learning_rate": 6.873032967079562e-06,
470
+ "loss": 0.1609,
471
+ "step": 66
472
+ },
473
+ {
474
+ "epoch": 6.7,
475
+ "grad_norm": 0.678112033468826,
476
+ "learning_rate": 6.7646528235693445e-06,
477
+ "loss": 0.154,
478
+ "step": 67
479
+ },
480
+ {
481
+ "epoch": 6.8,
482
+ "grad_norm": 0.6849273910905006,
483
+ "learning_rate": 6.655317089424791e-06,
484
+ "loss": 0.1467,
485
+ "step": 68
486
+ },
487
+ {
488
+ "epoch": 6.9,
489
+ "grad_norm": 0.6133050858596855,
490
+ "learning_rate": 6.545084971874738e-06,
491
+ "loss": 0.1433,
492
+ "step": 69
493
+ },
494
+ {
495
+ "epoch": 7.0,
496
+ "grad_norm": 0.7025492720750506,
497
+ "learning_rate": 6.434016163555452e-06,
498
+ "loss": 0.1558,
499
+ "step": 70
500
+ },
501
+ {
502
+ "epoch": 7.1,
503
+ "grad_norm": 0.5939241743382766,
504
+ "learning_rate": 6.322170810186013e-06,
505
+ "loss": 0.1154,
506
+ "step": 71
507
+ },
508
+ {
509
+ "epoch": 7.2,
510
+ "grad_norm": 0.6551743814202161,
511
+ "learning_rate": 6.209609477998339e-06,
512
+ "loss": 0.089,
513
+ "step": 72
514
+ },
515
+ {
516
+ "epoch": 7.3,
517
+ "grad_norm": 0.6738211444755293,
518
+ "learning_rate": 6.0963931209395165e-06,
519
+ "loss": 0.1305,
520
+ "step": 73
521
+ },
522
+ {
523
+ "epoch": 7.4,
524
+ "grad_norm": 0.7807788497212088,
525
+ "learning_rate": 5.982583047664151e-06,
526
+ "loss": 0.0925,
527
+ "step": 74
528
+ },
529
+ {
530
+ "epoch": 7.5,
531
+ "grad_norm": 0.7093670961487814,
532
+ "learning_rate": 5.8682408883346535e-06,
533
+ "loss": 0.0856,
534
+ "step": 75
535
+ },
536
+ {
537
+ "epoch": 7.6,
538
+ "grad_norm": 0.6308197479344778,
539
+ "learning_rate": 5.753428561247416e-06,
540
+ "loss": 0.105,
541
+ "step": 76
542
+ },
543
+ {
544
+ "epoch": 7.7,
545
+ "grad_norm": 0.5760024857690129,
546
+ "learning_rate": 5.638208239302975e-06,
547
+ "loss": 0.1059,
548
+ "step": 77
549
+ },
550
+ {
551
+ "epoch": 7.8,
552
+ "grad_norm": 0.5842094074834091,
553
+ "learning_rate": 5.522642316338268e-06,
554
+ "loss": 0.102,
555
+ "step": 78
556
+ },
557
+ {
558
+ "epoch": 7.9,
559
+ "grad_norm": 0.6001565299169008,
560
+ "learning_rate": 5.406793373339292e-06,
561
+ "loss": 0.0931,
562
+ "step": 79
563
+ },
564
+ {
565
+ "epoch": 8.0,
566
+ "grad_norm": 0.6582634642872089,
567
+ "learning_rate": 5.290724144552379e-06,
568
+ "loss": 0.101,
569
+ "step": 80
570
+ },
571
+ {
572
+ "epoch": 8.1,
573
+ "grad_norm": 0.5439026193832287,
574
+ "learning_rate": 5.174497483512506e-06,
575
+ "loss": 0.0647,
576
+ "step": 81
577
+ },
578
+ {
579
+ "epoch": 8.2,
580
+ "grad_norm": 0.6060347168628608,
581
+ "learning_rate": 5.0581763290069865e-06,
582
+ "loss": 0.0601,
583
+ "step": 82
584
+ },
585
+ {
586
+ "epoch": 8.3,
587
+ "grad_norm": 0.6611443276006925,
588
+ "learning_rate": 4.941823670993016e-06,
589
+ "loss": 0.0678,
590
+ "step": 83
591
+ },
592
+ {
593
+ "epoch": 8.4,
594
+ "grad_norm": 0.7605219891730366,
595
+ "learning_rate": 4.825502516487497e-06,
596
+ "loss": 0.0783,
597
+ "step": 84
598
+ },
599
+ {
600
+ "epoch": 8.5,
601
+ "grad_norm": 0.6045012675156527,
602
+ "learning_rate": 4.7092758554476215e-06,
603
+ "loss": 0.0584,
604
+ "step": 85
605
+ },
606
+ {
607
+ "epoch": 8.6,
608
+ "grad_norm": 0.7233963933340125,
609
+ "learning_rate": 4.59320662666071e-06,
610
+ "loss": 0.0666,
611
+ "step": 86
612
+ },
613
+ {
614
+ "epoch": 8.7,
615
+ "grad_norm": 0.5569053489022656,
616
+ "learning_rate": 4.477357683661734e-06,
617
+ "loss": 0.0547,
618
+ "step": 87
619
+ },
620
+ {
621
+ "epoch": 8.8,
622
+ "grad_norm": 0.5559039385553084,
623
+ "learning_rate": 4.361791760697027e-06,
624
+ "loss": 0.0785,
625
+ "step": 88
626
+ },
627
+ {
628
+ "epoch": 8.9,
629
+ "grad_norm": 0.5556629188201956,
630
+ "learning_rate": 4.246571438752585e-06,
631
+ "loss": 0.0642,
632
+ "step": 89
633
+ },
634
+ {
635
+ "epoch": 9.0,
636
+ "grad_norm": 0.5790897185039965,
637
+ "learning_rate": 4.131759111665349e-06,
638
+ "loss": 0.0522,
639
+ "step": 90
640
+ },
641
+ {
642
+ "epoch": 9.1,
643
+ "grad_norm": 0.6027745406131993,
644
+ "learning_rate": 4.017416952335849e-06,
645
+ "loss": 0.0524,
646
+ "step": 91
647
+ },
648
+ {
649
+ "epoch": 9.2,
650
+ "grad_norm": 0.463255876056378,
651
+ "learning_rate": 3.903606879060483e-06,
652
+ "loss": 0.0312,
653
+ "step": 92
654
+ },
655
+ {
656
+ "epoch": 9.3,
657
+ "grad_norm": 0.4952411302704621,
658
+ "learning_rate": 3.790390522001662e-06,
659
+ "loss": 0.0325,
660
+ "step": 93
661
+ },
662
+ {
663
+ "epoch": 9.4,
664
+ "grad_norm": 0.6078404485080328,
665
+ "learning_rate": 3.6778291898139907e-06,
666
+ "loss": 0.042,
667
+ "step": 94
668
+ },
669
+ {
670
+ "epoch": 9.5,
671
+ "grad_norm": 0.5902634839005834,
672
+ "learning_rate": 3.5659838364445505e-06,
673
+ "loss": 0.0392,
674
+ "step": 95
675
+ },
676
+ {
677
+ "epoch": 9.6,
678
+ "grad_norm": 0.49447046111646803,
679
+ "learning_rate": 3.4549150281252635e-06,
680
+ "loss": 0.0315,
681
+ "step": 96
682
+ },
683
+ {
684
+ "epoch": 9.7,
685
+ "grad_norm": 0.5605316793296448,
686
+ "learning_rate": 3.3446829105752103e-06,
687
+ "loss": 0.04,
688
+ "step": 97
689
+ },
690
+ {
691
+ "epoch": 9.8,
692
+ "grad_norm": 0.6375500669013182,
693
+ "learning_rate": 3.2353471764306567e-06,
694
+ "loss": 0.0337,
695
+ "step": 98
696
+ },
697
+ {
698
+ "epoch": 9.9,
699
+ "grad_norm": 0.4472086442084951,
700
+ "learning_rate": 3.12696703292044e-06,
701
+ "loss": 0.0317,
702
+ "step": 99
703
+ },
704
+ {
705
+ "epoch": 10.0,
706
+ "grad_norm": 0.46602982260632675,
707
+ "learning_rate": 3.019601169804216e-06,
708
+ "loss": 0.0402,
709
+ "step": 100
710
+ },
711
+ {
712
+ "epoch": 10.1,
713
+ "grad_norm": 0.3813466284022515,
714
+ "learning_rate": 2.9133077275909112e-06,
715
+ "loss": 0.0216,
716
+ "step": 101
717
+ },
718
+ {
719
+ "epoch": 10.2,
720
+ "grad_norm": 0.455818698598771,
721
+ "learning_rate": 2.8081442660546126e-06,
722
+ "loss": 0.0222,
723
+ "step": 102
724
+ },
725
+ {
726
+ "epoch": 10.3,
727
+ "grad_norm": 0.3501312765063474,
728
+ "learning_rate": 2.7041677330649408e-06,
729
+ "loss": 0.0198,
730
+ "step": 103
731
+ },
732
+ {
733
+ "epoch": 10.4,
734
+ "grad_norm": 0.3757425281553194,
735
+ "learning_rate": 2.601434433748771e-06,
736
+ "loss": 0.0289,
737
+ "step": 104
738
+ },
739
+ {
740
+ "epoch": 10.5,
741
+ "grad_norm": 0.41788768541144317,
742
+ "learning_rate": 2.5000000000000015e-06,
743
+ "loss": 0.0224,
744
+ "step": 105
745
+ },
746
+ {
747
+ "epoch": 10.6,
748
+ "grad_norm": 0.4744129849743114,
749
+ "learning_rate": 2.3999193603539234e-06,
750
+ "loss": 0.0294,
751
+ "step": 106
752
+ },
753
+ {
754
+ "epoch": 10.7,
755
+ "grad_norm": 0.44511151611382244,
756
+ "learning_rate": 2.3012467102424373e-06,
757
+ "loss": 0.0159,
758
+ "step": 107
759
+ },
760
+ {
761
+ "epoch": 10.8,
762
+ "grad_norm": 0.5747679046158326,
763
+ "learning_rate": 2.204035482646267e-06,
764
+ "loss": 0.024,
765
+ "step": 108
766
+ },
767
+ {
768
+ "epoch": 10.9,
769
+ "grad_norm": 0.48309056921628807,
770
+ "learning_rate": 2.1083383191600676e-06,
771
+ "loss": 0.0165,
772
+ "step": 109
773
+ },
774
+ {
775
+ "epoch": 11.0,
776
+ "grad_norm": 0.3678481549024087,
777
+ "learning_rate": 2.0142070414860704e-06,
778
+ "loss": 0.0208,
779
+ "step": 110
780
+ },
781
+ {
782
+ "epoch": 11.1,
783
+ "grad_norm": 0.359419218455559,
784
+ "learning_rate": 1.9216926233717087e-06,
785
+ "loss": 0.0189,
786
+ "step": 111
787
+ },
788
+ {
789
+ "epoch": 11.2,
790
+ "grad_norm": 0.23198184996326746,
791
+ "learning_rate": 1.8308451630064484e-06,
792
+ "loss": 0.0106,
793
+ "step": 112
794
+ },
795
+ {
796
+ "epoch": 11.3,
797
+ "grad_norm": 0.32404279906347383,
798
+ "learning_rate": 1.7417138558927244e-06,
799
+ "loss": 0.0163,
800
+ "step": 113
801
+ },
802
+ {
803
+ "epoch": 11.4,
804
+ "grad_norm": 0.6427059152860045,
805
+ "learning_rate": 1.6543469682057105e-06,
806
+ "loss": 0.0114,
807
+ "step": 114
808
+ },
809
+ {
810
+ "epoch": 11.5,
811
+ "grad_norm": 0.25041123953375244,
812
+ "learning_rate": 1.5687918106563326e-06,
813
+ "loss": 0.0126,
814
+ "step": 115
815
+ },
816
+ {
817
+ "epoch": 11.6,
818
+ "grad_norm": 0.28895936106203457,
819
+ "learning_rate": 1.4850947128716914e-06,
820
+ "loss": 0.0148,
821
+ "step": 116
822
+ },
823
+ {
824
+ "epoch": 11.7,
825
+ "grad_norm": 0.29393456592529227,
826
+ "learning_rate": 1.4033009983067454e-06,
827
+ "loss": 0.016,
828
+ "step": 117
829
+ },
830
+ {
831
+ "epoch": 11.8,
832
+ "grad_norm": 0.2636158141060736,
833
+ "learning_rate": 1.3234549597008572e-06,
834
+ "loss": 0.0142,
835
+ "step": 118
836
+ },
837
+ {
838
+ "epoch": 11.9,
839
+ "grad_norm": 0.2638963903471292,
840
+ "learning_rate": 1.2455998350925042e-06,
841
+ "loss": 0.0173,
842
+ "step": 119
843
+ },
844
+ {
845
+ "epoch": 12.0,
846
+ "grad_norm": 0.2902755362049078,
847
+ "learning_rate": 1.1697777844051105e-06,
848
+ "loss": 0.0117,
849
+ "step": 120
850
+ },
851
+ {
852
+ "epoch": 12.1,
853
+ "grad_norm": 0.2395169915153762,
854
+ "learning_rate": 1.096029866616704e-06,
855
+ "loss": 0.011,
856
+ "step": 121
857
+ },
858
+ {
859
+ "epoch": 12.2,
860
+ "grad_norm": 0.22844426778585733,
861
+ "learning_rate": 1.0243960175257605e-06,
862
+ "loss": 0.0139,
863
+ "step": 122
864
+ },
865
+ {
866
+ "epoch": 12.3,
867
+ "grad_norm": 0.22125598333566673,
868
+ "learning_rate": 9.549150281252633e-07,
869
+ "loss": 0.0138,
870
+ "step": 123
871
+ },
872
+ {
873
+ "epoch": 12.4,
874
+ "grad_norm": 0.2456761278799785,
875
+ "learning_rate": 8.876245235966884e-07,
876
+ "loss": 0.0112,
877
+ "step": 124
878
+ },
879
+ {
880
+ "epoch": 12.5,
881
+ "grad_norm": 0.19964265632582723,
882
+ "learning_rate": 8.225609429353187e-07,
883
+ "loss": 0.0106,
884
+ "step": 125
885
+ },
886
+ {
887
+ "epoch": 12.6,
888
+ "grad_norm": 0.23420390579285358,
889
+ "learning_rate": 7.597595192178702e-07,
890
+ "loss": 0.0146,
891
+ "step": 126
892
+ },
893
+ {
894
+ "epoch": 12.7,
895
+ "grad_norm": 0.20632816171101767,
896
+ "learning_rate": 6.992542605231739e-07,
897
+ "loss": 0.0091,
898
+ "step": 127
899
+ },
900
+ {
901
+ "epoch": 12.8,
902
+ "grad_norm": 0.23775375009146965,
903
+ "learning_rate": 6.410779315161885e-07,
904
+ "loss": 0.0107,
905
+ "step": 128
906
+ },
907
+ {
908
+ "epoch": 12.9,
909
+ "grad_norm": 0.2118725012300239,
910
+ "learning_rate": 5.852620357053651e-07,
911
+ "loss": 0.0084,
912
+ "step": 129
913
+ },
914
+ {
915
+ "epoch": 13.0,
916
+ "grad_norm": 0.26953272537694395,
917
+ "learning_rate": 5.318367983829393e-07,
918
+ "loss": 0.0096,
919
+ "step": 130
920
+ },
921
+ {
922
+ "epoch": 13.1,
923
+ "grad_norm": 0.19519736941796972,
924
+ "learning_rate": 4.808311502573976e-07,
925
+ "loss": 0.009,
926
+ "step": 131
927
+ },
928
+ {
929
+ "epoch": 13.2,
930
+ "grad_norm": 0.17411684232028857,
931
+ "learning_rate": 4.322727117869951e-07,
932
+ "loss": 0.0112,
933
+ "step": 132
934
+ },
935
+ {
936
+ "epoch": 13.3,
937
+ "grad_norm": 0.15926392868818823,
938
+ "learning_rate": 3.8618777822278854e-07,
939
+ "loss": 0.0097,
940
+ "step": 133
941
+ },
942
+ {
943
+ "epoch": 13.4,
944
+ "grad_norm": 0.1405271281556315,
945
+ "learning_rate": 3.426013053692878e-07,
946
+ "loss": 0.0069,
947
+ "step": 134
948
+ },
949
+ {
950
+ "epoch": 13.5,
951
+ "grad_norm": 0.16600768538701607,
952
+ "learning_rate": 3.015368960704584e-07,
953
+ "loss": 0.0108,
954
+ "step": 135
955
+ },
956
+ {
957
+ "epoch": 13.6,
958
+ "grad_norm": 0.17721046661374376,
959
+ "learning_rate": 2.63016787428354e-07,
960
+ "loss": 0.01,
961
+ "step": 136
962
+ },
963
+ {
964
+ "epoch": 13.7,
965
+ "grad_norm": 0.14170979926112712,
966
+ "learning_rate": 2.2706183876134047e-07,
967
+ "loss": 0.0073,
968
+ "step": 137
969
+ },
970
+ {
971
+ "epoch": 13.8,
972
+ "grad_norm": 0.14708132869236895,
973
+ "learning_rate": 1.9369152030840553e-07,
974
+ "loss": 0.0076,
975
+ "step": 138
976
+ },
977
+ {
978
+ "epoch": 13.9,
979
+ "grad_norm": 0.14194896166176377,
980
+ "learning_rate": 1.6292390268568103e-07,
981
+ "loss": 0.0059,
982
+ "step": 139
983
+ },
984
+ {
985
+ "epoch": 14.0,
986
+ "grad_norm": 0.1672952260346504,
987
+ "learning_rate": 1.3477564710088097e-07,
988
+ "loss": 0.0087,
989
+ "step": 140
990
+ },
991
+ {
992
+ "epoch": 14.1,
993
+ "grad_norm": 0.14790941756728795,
994
+ "learning_rate": 1.0926199633097156e-07,
995
+ "loss": 0.0065,
996
+ "step": 141
997
+ },
998
+ {
999
+ "epoch": 14.2,
1000
+ "grad_norm": 0.1543104570325743,
1001
+ "learning_rate": 8.639676646793382e-08,
1002
+ "loss": 0.0084,
1003
+ "step": 142
1004
+ },
1005
+ {
1006
+ "epoch": 14.3,
1007
+ "grad_norm": 0.18074780754737557,
1008
+ "learning_rate": 6.61923394371039e-08,
1009
+ "loss": 0.0121,
1010
+ "step": 143
1011
+ },
1012
+ {
1013
+ "epoch": 14.4,
1014
+ "grad_norm": 0.11318793125638911,
1015
+ "learning_rate": 4.865965629214819e-08,
1016
+ "loss": 0.0051,
1017
+ "step": 144
1018
+ },
1019
+ {
1020
+ "epoch": 14.5,
1021
+ "grad_norm": 0.1398402355020413,
1022
+ "learning_rate": 3.3808211290284886e-08,
1023
+ "loss": 0.0065,
1024
+ "step": 145
1025
+ },
1026
+ {
1027
+ "epoch": 14.6,
1028
+ "grad_norm": 0.13844915147008274,
1029
+ "learning_rate": 2.1646046750978255e-08,
1030
+ "loss": 0.0075,
1031
+ "step": 146
1032
+ },
1033
+ {
1034
+ "epoch": 14.7,
1035
+ "grad_norm": 0.16875064849806554,
1036
+ "learning_rate": 1.2179748700879013e-08,
1037
+ "loss": 0.0109,
1038
+ "step": 147
1039
+ },
1040
+ {
1041
+ "epoch": 14.8,
1042
+ "grad_norm": 0.13630194717892072,
1043
+ "learning_rate": 5.414443307377171e-09,
1044
+ "loss": 0.0077,
1045
+ "step": 148
1046
+ },
1047
+ {
1048
+ "epoch": 14.9,
1049
+ "grad_norm": 0.16493416397590105,
1050
+ "learning_rate": 1.3537941026914302e-09,
1051
+ "loss": 0.0107,
1052
+ "step": 149
1053
+ },
1054
+ {
1055
+ "epoch": 15.0,
1056
+ "grad_norm": 0.13222454006520384,
1057
+ "learning_rate": 0.0,
1058
+ "loss": 0.0081,
1059
+ "step": 150
1060
+ },
1061
+ {
1062
+ "epoch": 15.0,
1063
+ "step": 150,
1064
+ "total_flos": 8.481547119440691e+16,
1065
+ "train_loss": 0.2379360094014555,
1066
+ "train_runtime": 6253.349,
1067
+ "train_samples_per_second": 0.758,
1068
+ "train_steps_per_second": 0.024
1069
+ }
1070
+ ],
1071
+ "logging_steps": 1.0,
1072
+ "max_steps": 150,
1073
+ "num_input_tokens_seen": 0,
1074
+ "num_train_epochs": 15,
1075
+ "save_steps": 500,
1076
+ "stateful_callbacks": {
1077
+ "TrainerControl": {
1078
+ "args": {
1079
+ "should_epoch_stop": false,
1080
+ "should_evaluate": false,
1081
+ "should_log": false,
1082
+ "should_save": true,
1083
+ "should_training_stop": true
1084
+ },
1085
+ "attributes": {}
1086
+ }
1087
+ },
1088
+ "total_flos": 8.481547119440691e+16,
1089
+ "train_batch_size": 1,
1090
+ "trial_name": null,
1091
+ "trial_params": null
1092
+ }
training_loss.png ADDED