Rakancorle1 commited on
Commit
0ff0063
·
verified ·
1 Parent(s): e6366ba

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 +931 -0
  5. training_loss.png +0 -0
README.md CHANGED
@@ -4,6 +4,7 @@ license: llama3.1
4
  base_model: meta-llama/Llama-Guard-3-8B
5
  tags:
6
  - llama-factory
 
7
  - generated_from_trainer
8
  model-index:
9
  - name: ThinkGuard_27k_3epochs_gas16_1.5e
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # ThinkGuard_27k_3epochs_gas16_1.5e
17
 
18
- This model is a fine-tuned version of [meta-llama/Llama-Guard-3-8B](https://huggingface.co/meta-llama/Llama-Guard-3-8B) on an unknown dataset.
19
 
20
  ## Model description
21
 
 
4
  base_model: meta-llama/Llama-Guard-3-8B
5
  tags:
6
  - llama-factory
7
+ - full
8
  - generated_from_trainer
9
  model-index:
10
  - name: ThinkGuard_27k_3epochs_gas16_1.5e
 
16
 
17
  # ThinkGuard_27k_3epochs_gas16_1.5e
18
 
19
+ This model is a fine-tuned version of [meta-llama/Llama-Guard-3-8B](https://huggingface.co/meta-llama/Llama-Guard-3-8B) on the beavertails_27k_GenExplanation dataset.
20
 
21
  ## Model description
22
 
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.994262174488745,
3
+ "total_flos": 76511791349760.0,
4
+ "train_loss": 0.29241873632015297,
5
+ "train_runtime": 70414.7871,
6
+ "train_samples_per_second": 1.158,
7
+ "train_steps_per_second": 0.018
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.994262174488745,
3
+ "total_flos": 76511791349760.0,
4
+ "train_loss": 0.29241873632015297,
5
+ "train_runtime": 70414.7871,
6
+ "train_samples_per_second": 1.158,
7
+ "train_steps_per_second": 0.018
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,931 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.994262174488745,
5
+ "eval_steps": 500,
6
+ "global_step": 1272,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02353979696925114,
13
+ "grad_norm": 8.782275747399924,
14
+ "learning_rate": 3.846153846153846e-06,
15
+ "loss": 1.5306,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.04707959393850228,
20
+ "grad_norm": 4.3210560861300085,
21
+ "learning_rate": 7.692307692307692e-06,
22
+ "loss": 0.8302,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.07061939090775342,
27
+ "grad_norm": 3.023277309045724,
28
+ "learning_rate": 1.153846153846154e-05,
29
+ "loss": 0.6217,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.09415918787700456,
34
+ "grad_norm": 2.654024929474284,
35
+ "learning_rate": 1.4999975655288908e-05,
36
+ "loss": 0.5967,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.1176989848462557,
41
+ "grad_norm": 2.8603296276880723,
42
+ "learning_rate": 1.4997054481186253e-05,
43
+ "loss": 0.5732,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.14123878181550684,
48
+ "grad_norm": 2.4252498205382347,
49
+ "learning_rate": 1.4989266537738068e-05,
50
+ "loss": 0.5573,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.164778578784758,
55
+ "grad_norm": 2.0008212035639787,
56
+ "learning_rate": 1.4976616880546462e-05,
57
+ "loss": 0.5488,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.18831837575400912,
62
+ "grad_norm": 2.1468051355877065,
63
+ "learning_rate": 1.4959113721231613e-05,
64
+ "loss": 0.5378,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.21185817272326027,
69
+ "grad_norm": 2.09758107733118,
70
+ "learning_rate": 1.4936768422101134e-05,
71
+ "loss": 0.524,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.2353979696925114,
76
+ "grad_norm": 1.9891368777701572,
77
+ "learning_rate": 1.4909595488774145e-05,
78
+ "loss": 0.5238,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.25893776666176255,
83
+ "grad_norm": 1.8856649935655327,
84
+ "learning_rate": 1.487761256076484e-05,
85
+ "loss": 0.51,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.2824775636310137,
90
+ "grad_norm": 1.9946450768120385,
91
+ "learning_rate": 1.4840840400031667e-05,
92
+ "loss": 0.5045,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.3060173606002648,
97
+ "grad_norm": 2.02887277527658,
98
+ "learning_rate": 1.4799302877499543e-05,
99
+ "loss": 0.4993,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.329557157569516,
104
+ "grad_norm": 1.7846741494593932,
105
+ "learning_rate": 1.475302695756387e-05,
106
+ "loss": 0.4959,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.3530969545387671,
111
+ "grad_norm": 1.7910316946805966,
112
+ "learning_rate": 1.4702042680586378e-05,
113
+ "loss": 0.4903,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.37663675150801823,
118
+ "grad_norm": 1.583559495202325,
119
+ "learning_rate": 1.4646383143394222e-05,
120
+ "loss": 0.4852,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.40017654847726936,
125
+ "grad_norm": 1.8258988651876884,
126
+ "learning_rate": 1.458608447779491e-05,
127
+ "loss": 0.4835,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.42371634544652054,
132
+ "grad_norm": 1.7923596937902033,
133
+ "learning_rate": 1.4521185827121071e-05,
134
+ "loss": 0.474,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.44725614241577166,
139
+ "grad_norm": 1.8160198247756183,
140
+ "learning_rate": 1.4451729320820273e-05,
141
+ "loss": 0.4721,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.4707959393850228,
146
+ "grad_norm": 1.7150789518558318,
147
+ "learning_rate": 1.437776004710637e-05,
148
+ "loss": 0.4729,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.49433573635427397,
153
+ "grad_norm": 1.7400033525862848,
154
+ "learning_rate": 1.4299326023690146e-05,
155
+ "loss": 0.4607,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.5178755333235251,
160
+ "grad_norm": 1.7002289663882648,
161
+ "learning_rate": 1.4216478166608262e-05,
162
+ "loss": 0.4613,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.5414153302927762,
167
+ "grad_norm": 1.534972777417761,
168
+ "learning_rate": 1.4129270257170722e-05,
169
+ "loss": 0.4567,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.5649551272620273,
174
+ "grad_norm": 1.6489960881148158,
175
+ "learning_rate": 1.403775890704834e-05,
176
+ "loss": 0.4559,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.5884949242312785,
181
+ "grad_norm": 1.7846842786992185,
182
+ "learning_rate": 1.3942003521522837e-05,
183
+ "loss": 0.4624,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.6120347212005296,
188
+ "grad_norm": 1.4351473554704604,
189
+ "learning_rate": 1.3842066260923462e-05,
190
+ "loss": 0.4496,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.6355745181697808,
195
+ "grad_norm": 1.541683257911224,
196
+ "learning_rate": 1.3738012000275142e-05,
197
+ "loss": 0.4516,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.659114315139032,
202
+ "grad_norm": 1.5879015034427175,
203
+ "learning_rate": 1.3629908287184381e-05,
204
+ "loss": 0.4475,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.6826541121082831,
209
+ "grad_norm": 1.728309878285744,
210
+ "learning_rate": 1.3517825297990205e-05,
211
+ "loss": 0.4446,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.7061939090775342,
216
+ "grad_norm": 1.4690146169046678,
217
+ "learning_rate": 1.340183579220868e-05,
218
+ "loss": 0.4491,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.7297337060467853,
223
+ "grad_norm": 1.4969895208337947,
224
+ "learning_rate": 1.328201506530052e-05,
225
+ "loss": 0.4436,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.7532735030160365,
230
+ "grad_norm": 1.4385449709842155,
231
+ "learning_rate": 1.3158440899792465e-05,
232
+ "loss": 0.4344,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.7768132999852876,
237
+ "grad_norm": 1.4819229027772451,
238
+ "learning_rate": 1.3031193514784178e-05,
239
+ "loss": 0.4387,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.8003530969545387,
244
+ "grad_norm": 1.5138328167674073,
245
+ "learning_rate": 1.2900355513873408e-05,
246
+ "loss": 0.4374,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.82389289392379,
251
+ "grad_norm": 1.5486731806719658,
252
+ "learning_rate": 1.276601183153324e-05,
253
+ "loss": 0.4268,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.8474326908930411,
258
+ "grad_norm": 1.3966287731811042,
259
+ "learning_rate": 1.2628249677976246e-05,
260
+ "loss": 0.4222,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.8709724878622922,
265
+ "grad_norm": 1.4851117277296138,
266
+ "learning_rate": 1.2487158482541324e-05,
267
+ "loss": 0.4304,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.8945122848315433,
272
+ "grad_norm": 1.4886042812511937,
273
+ "learning_rate": 1.2342829835639957e-05,
274
+ "loss": 0.4327,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 0.9180520818007945,
279
+ "grad_norm": 1.5916975841301912,
280
+ "learning_rate": 1.219535742929962e-05,
281
+ "loss": 0.4312,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 0.9415918787700456,
286
+ "grad_norm": 1.4359537454947529,
287
+ "learning_rate": 1.204483699634289e-05,
288
+ "loss": 0.418,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 0.9651316757392967,
293
+ "grad_norm": 1.5039520247360716,
294
+ "learning_rate": 1.1891366248241758e-05,
295
+ "loss": 0.4169,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 0.9886714727085479,
300
+ "grad_norm": 1.4396432279411446,
301
+ "learning_rate": 1.1735044811687508e-05,
302
+ "loss": 0.4165,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 1.012211269677799,
307
+ "grad_norm": 1.4837321963480583,
308
+ "learning_rate": 1.1575974163917282e-05,
309
+ "loss": 0.3505,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 1.0357510666470502,
314
+ "grad_norm": 1.438403278814582,
315
+ "learning_rate": 1.1414257566839376e-05,
316
+ "loss": 0.2795,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 1.0592908636163012,
321
+ "grad_norm": 1.510877062877757,
322
+ "learning_rate": 1.125e-05,
323
+ "loss": 0.2814,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 1.0828306605855524,
328
+ "grad_norm": 1.519349451859131,
329
+ "learning_rate": 1.1083308092435006e-05,
330
+ "loss": 0.2773,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 1.1063704575548037,
335
+ "grad_norm": 1.3106222683180944,
336
+ "learning_rate": 1.091429005345085e-05,
337
+ "loss": 0.2706,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 1.1299102545240547,
342
+ "grad_norm": 1.3984600885535579,
343
+ "learning_rate": 1.0743055602379712e-05,
344
+ "loss": 0.2754,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 1.153450051493306,
349
+ "grad_norm": 1.4833593619381575,
350
+ "learning_rate": 1.0569715897354354e-05,
351
+ "loss": 0.2756,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 1.176989848462557,
356
+ "grad_norm": 1.3932629119237105,
357
+ "learning_rate": 1.0394383463148993e-05,
358
+ "loss": 0.27,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 1.2005296454318082,
363
+ "grad_norm": 1.4062493836827707,
364
+ "learning_rate": 1.0217172118132994e-05,
365
+ "loss": 0.2703,
366
+ "step": 510
367
+ },
368
+ {
369
+ "epoch": 1.2240694424010594,
370
+ "grad_norm": 1.4874549162201474,
371
+ "learning_rate": 1.003819690038481e-05,
372
+ "loss": 0.2739,
373
+ "step": 520
374
+ },
375
+ {
376
+ "epoch": 1.2476092393703104,
377
+ "grad_norm": 1.3708361425521698,
378
+ "learning_rate": 9.857573993014161e-06,
379
+ "loss": 0.2754,
380
+ "step": 530
381
+ },
382
+ {
383
+ "epoch": 1.2711490363395614,
384
+ "grad_norm": 1.3978327948674458,
385
+ "learning_rate": 9.675420648740886e-06,
386
+ "loss": 0.2692,
387
+ "step": 540
388
+ },
389
+ {
390
+ "epoch": 1.2946888333088127,
391
+ "grad_norm": 1.4715538000650579,
392
+ "learning_rate": 9.491855113779456e-06,
393
+ "loss": 0.2694,
394
+ "step": 550
395
+ },
396
+ {
397
+ "epoch": 1.318228630278064,
398
+ "grad_norm": 1.3468394374097452,
399
+ "learning_rate": 9.306996551078556e-06,
400
+ "loss": 0.2678,
401
+ "step": 560
402
+ },
403
+ {
404
+ "epoch": 1.341768427247315,
405
+ "grad_norm": 1.5175300555093134,
406
+ "learning_rate": 9.120964962965565e-06,
407
+ "loss": 0.266,
408
+ "step": 570
409
+ },
410
+ {
411
+ "epoch": 1.3653082242165662,
412
+ "grad_norm": 1.317724218624981,
413
+ "learning_rate": 8.933881113246134e-06,
414
+ "loss": 0.2666,
415
+ "step": 580
416
+ },
417
+ {
418
+ "epoch": 1.3888480211858174,
419
+ "grad_norm": 1.3809839276376867,
420
+ "learning_rate": 8.74586644880946e-06,
421
+ "loss": 0.2672,
422
+ "step": 590
423
+ },
424
+ {
425
+ "epoch": 1.4123878181550684,
426
+ "grad_norm": 1.3120307514577167,
427
+ "learning_rate": 8.557043020790113e-06,
428
+ "loss": 0.2671,
429
+ "step": 600
430
+ },
431
+ {
432
+ "epoch": 1.4359276151243194,
433
+ "grad_norm": 1.5865169519570839,
434
+ "learning_rate": 8.367533405337635e-06,
435
+ "loss": 0.2669,
436
+ "step": 610
437
+ },
438
+ {
439
+ "epoch": 1.4594674120935707,
440
+ "grad_norm": 1.315530128518467,
441
+ "learning_rate": 8.177460624045303e-06,
442
+ "loss": 0.2619,
443
+ "step": 620
444
+ },
445
+ {
446
+ "epoch": 1.483007209062822,
447
+ "grad_norm": 1.2729315130214236,
448
+ "learning_rate": 7.986948064089741e-06,
449
+ "loss": 0.2625,
450
+ "step": 630
451
+ },
452
+ {
453
+ "epoch": 1.506547006032073,
454
+ "grad_norm": 1.2890769997858393,
455
+ "learning_rate": 7.7961193981332e-06,
456
+ "loss": 0.2624,
457
+ "step": 640
458
+ },
459
+ {
460
+ "epoch": 1.5300868030013242,
461
+ "grad_norm": 1.344250246429933,
462
+ "learning_rate": 7.605098504040519e-06,
463
+ "loss": 0.2571,
464
+ "step": 650
465
+ },
466
+ {
467
+ "epoch": 1.5536265999705754,
468
+ "grad_norm": 1.3112449299655193,
469
+ "learning_rate": 7.414009384462882e-06,
470
+ "loss": 0.2603,
471
+ "step": 660
472
+ },
473
+ {
474
+ "epoch": 1.5771663969398264,
475
+ "grad_norm": 1.3722682469514345,
476
+ "learning_rate": 7.22297608634056e-06,
477
+ "loss": 0.2592,
478
+ "step": 670
479
+ },
480
+ {
481
+ "epoch": 1.6007061939090774,
482
+ "grad_norm": 1.3170860077023523,
483
+ "learning_rate": 7.032122620376899e-06,
484
+ "loss": 0.2535,
485
+ "step": 680
486
+ },
487
+ {
488
+ "epoch": 1.6242459908783287,
489
+ "grad_norm": 1.3106743856940115,
490
+ "learning_rate": 6.841572880535854e-06,
491
+ "loss": 0.2531,
492
+ "step": 690
493
+ },
494
+ {
495
+ "epoch": 1.64778578784758,
496
+ "grad_norm": 1.2798600011553958,
497
+ "learning_rate": 6.6514505636152814e-06,
498
+ "loss": 0.2548,
499
+ "step": 700
500
+ },
501
+ {
502
+ "epoch": 1.671325584816831,
503
+ "grad_norm": 1.3968866134114841,
504
+ "learning_rate": 6.461879088948227e-06,
505
+ "loss": 0.2507,
506
+ "step": 710
507
+ },
508
+ {
509
+ "epoch": 1.6948653817860821,
510
+ "grad_norm": 1.378986621794313,
511
+ "learning_rate": 6.272981518284342e-06,
512
+ "loss": 0.249,
513
+ "step": 720
514
+ },
515
+ {
516
+ "epoch": 1.7184051787553334,
517
+ "grad_norm": 1.2776388271568753,
518
+ "learning_rate": 6.084880475903424e-06,
519
+ "loss": 0.2451,
520
+ "step": 730
521
+ },
522
+ {
523
+ "epoch": 1.7419449757245844,
524
+ "grad_norm": 1.317908633074014,
525
+ "learning_rate": 5.897698069012938e-06,
526
+ "loss": 0.2484,
527
+ "step": 740
528
+ },
529
+ {
530
+ "epoch": 1.7654847726938354,
531
+ "grad_norm": 1.2702586608772435,
532
+ "learning_rate": 5.711555808481213e-06,
533
+ "loss": 0.2486,
534
+ "step": 750
535
+ },
536
+ {
537
+ "epoch": 1.7890245696630867,
538
+ "grad_norm": 1.3790950940287556,
539
+ "learning_rate": 5.526574529957729e-06,
540
+ "loss": 0.2451,
541
+ "step": 760
542
+ },
543
+ {
544
+ "epoch": 1.8125643666323379,
545
+ "grad_norm": 1.417794065597115,
546
+ "learning_rate": 5.342874315431765e-06,
547
+ "loss": 0.2442,
548
+ "step": 770
549
+ },
550
+ {
551
+ "epoch": 1.836104163601589,
552
+ "grad_norm": 1.3118057182766703,
553
+ "learning_rate": 5.160574415280257e-06,
554
+ "loss": 0.2438,
555
+ "step": 780
556
+ },
557
+ {
558
+ "epoch": 1.85964396057084,
559
+ "grad_norm": 1.380491802365371,
560
+ "learning_rate": 4.979793170855537e-06,
561
+ "loss": 0.2368,
562
+ "step": 790
563
+ },
564
+ {
565
+ "epoch": 1.8831837575400914,
566
+ "grad_norm": 1.3404717450970274,
567
+ "learning_rate": 4.800647937663126e-06,
568
+ "loss": 0.2393,
569
+ "step": 800
570
+ },
571
+ {
572
+ "epoch": 1.9067235545093424,
573
+ "grad_norm": 1.3084085289097462,
574
+ "learning_rate": 4.623255009179547e-06,
575
+ "loss": 0.2402,
576
+ "step": 810
577
+ },
578
+ {
579
+ "epoch": 1.9302633514785934,
580
+ "grad_norm": 1.2819422568439027,
581
+ "learning_rate": 4.4477295413595385e-06,
582
+ "loss": 0.2359,
583
+ "step": 820
584
+ },
585
+ {
586
+ "epoch": 1.9538031484478446,
587
+ "grad_norm": 1.2506810831080168,
588
+ "learning_rate": 4.274185477881681e-06,
589
+ "loss": 0.2365,
590
+ "step": 830
591
+ },
592
+ {
593
+ "epoch": 1.9773429454170959,
594
+ "grad_norm": 1.236732854007973,
595
+ "learning_rate": 4.102735476181019e-06,
596
+ "loss": 0.2308,
597
+ "step": 840
598
+ },
599
+ {
600
+ "epoch": 2.000882742386347,
601
+ "grad_norm": 1.2110215418076606,
602
+ "learning_rate": 3.933490834316633e-06,
603
+ "loss": 0.2284,
604
+ "step": 850
605
+ },
606
+ {
607
+ "epoch": 2.024422539355598,
608
+ "grad_norm": 1.4341116240165481,
609
+ "learning_rate": 3.7665614187216603e-06,
610
+ "loss": 0.128,
611
+ "step": 860
612
+ },
613
+ {
614
+ "epoch": 2.0479623363248494,
615
+ "grad_norm": 1.1448959546708688,
616
+ "learning_rate": 3.6020555928826935e-06,
617
+ "loss": 0.1201,
618
+ "step": 870
619
+ },
620
+ {
621
+ "epoch": 2.0715021332941004,
622
+ "grad_norm": 1.1891044799932367,
623
+ "learning_rate": 3.4400801469947734e-06,
624
+ "loss": 0.1196,
625
+ "step": 880
626
+ },
627
+ {
628
+ "epoch": 2.0950419302633514,
629
+ "grad_norm": 1.2226413834941532,
630
+ "learning_rate": 3.2807402286377675e-06,
631
+ "loss": 0.1183,
632
+ "step": 890
633
+ },
634
+ {
635
+ "epoch": 2.1185817272326024,
636
+ "grad_norm": 1.3069856648534286,
637
+ "learning_rate": 3.1241392745189985e-06,
638
+ "loss": 0.1181,
639
+ "step": 900
640
+ },
641
+ {
642
+ "epoch": 2.142121524201854,
643
+ "grad_norm": 1.1702804083458354,
644
+ "learning_rate": 2.970378943326543e-06,
645
+ "loss": 0.1166,
646
+ "step": 910
647
+ },
648
+ {
649
+ "epoch": 2.165661321171105,
650
+ "grad_norm": 1.2124612400300803,
651
+ "learning_rate": 2.8195590497367222e-06,
652
+ "loss": 0.1155,
653
+ "step": 920
654
+ },
655
+ {
656
+ "epoch": 2.189201118140356,
657
+ "grad_norm": 1.1611863697964633,
658
+ "learning_rate": 2.671777499618652e-06,
659
+ "loss": 0.1159,
660
+ "step": 930
661
+ },
662
+ {
663
+ "epoch": 2.2127409151096074,
664
+ "grad_norm": 1.1087717944945865,
665
+ "learning_rate": 2.527130226477926e-06,
666
+ "loss": 0.1168,
667
+ "step": 940
668
+ },
669
+ {
670
+ "epoch": 2.2362807120788584,
671
+ "grad_norm": 1.2482164891383611,
672
+ "learning_rate": 2.3857111291806348e-06,
673
+ "loss": 0.114,
674
+ "step": 950
675
+ },
676
+ {
677
+ "epoch": 2.2598205090481094,
678
+ "grad_norm": 1.1704503748862236,
679
+ "learning_rate": 2.2476120109982267e-06,
680
+ "loss": 0.1146,
681
+ "step": 960
682
+ },
683
+ {
684
+ "epoch": 2.283360306017361,
685
+ "grad_norm": 1.1831598264423941,
686
+ "learning_rate": 2.1129225200127196e-06,
687
+ "loss": 0.1142,
688
+ "step": 970
689
+ },
690
+ {
691
+ "epoch": 2.306900102986612,
692
+ "grad_norm": 1.1923513771478957,
693
+ "learning_rate": 1.981730090920969e-06,
694
+ "loss": 0.1121,
695
+ "step": 980
696
+ },
697
+ {
698
+ "epoch": 2.330439899955863,
699
+ "grad_norm": 1.1650306486853608,
700
+ "learning_rate": 1.8541198882757892e-06,
701
+ "loss": 0.1103,
702
+ "step": 990
703
+ },
704
+ {
705
+ "epoch": 2.353979696925114,
706
+ "grad_norm": 1.2474184378379198,
707
+ "learning_rate": 1.73017475120072e-06,
708
+ "loss": 0.1111,
709
+ "step": 1000
710
+ },
711
+ {
712
+ "epoch": 2.3775194938943653,
713
+ "grad_norm": 1.1410121185645685,
714
+ "learning_rate": 1.6099751396144142e-06,
715
+ "loss": 0.1101,
716
+ "step": 1010
717
+ },
718
+ {
719
+ "epoch": 2.4010592908636164,
720
+ "grad_norm": 1.1781449006163716,
721
+ "learning_rate": 1.493599081999453e-06,
722
+ "loss": 0.1085,
723
+ "step": 1020
724
+ },
725
+ {
726
+ "epoch": 2.4245990878328674,
727
+ "grad_norm": 1.095164172281657,
728
+ "learning_rate": 1.3811221247495784e-06,
729
+ "loss": 0.109,
730
+ "step": 1030
731
+ },
732
+ {
733
+ "epoch": 2.448138884802119,
734
+ "grad_norm": 1.1290274158456068,
735
+ "learning_rate": 1.2726172831281898e-06,
736
+ "loss": 0.1083,
737
+ "step": 1040
738
+ },
739
+ {
740
+ "epoch": 2.47167868177137,
741
+ "grad_norm": 1.1205826114027153,
742
+ "learning_rate": 1.1681549938699157e-06,
743
+ "loss": 0.1079,
744
+ "step": 1050
745
+ },
746
+ {
747
+ "epoch": 2.495218478740621,
748
+ "grad_norm": 1.237936841602485,
749
+ "learning_rate": 1.0678030694561044e-06,
750
+ "loss": 0.1084,
751
+ "step": 1060
752
+ },
753
+ {
754
+ "epoch": 2.518758275709872,
755
+ "grad_norm": 1.2748079747165908,
756
+ "learning_rate": 9.716266540938183e-07,
757
+ "loss": 0.1084,
758
+ "step": 1070
759
+ },
760
+ {
761
+ "epoch": 2.542298072679123,
762
+ "grad_norm": 1.1620787715820426,
763
+ "learning_rate": 8.796881814269941e-07,
764
+ "loss": 0.106,
765
+ "step": 1080
766
+ },
767
+ {
768
+ "epoch": 2.5658378696483743,
769
+ "grad_norm": 1.1814024146181927,
770
+ "learning_rate": 7.92047334007169e-07,
771
+ "loss": 0.1037,
772
+ "step": 1090
773
+ },
774
+ {
775
+ "epoch": 2.5893776666176254,
776
+ "grad_norm": 1.1411929584644598,
777
+ "learning_rate": 7.087610045501023e-07,
778
+ "loss": 0.1044,
779
+ "step": 1100
780
+ },
781
+ {
782
+ "epoch": 2.612917463586877,
783
+ "grad_norm": 1.1770255689497389,
784
+ "learning_rate": 6.29883259003445e-07,
785
+ "loss": 0.1051,
786
+ "step": 1110
787
+ },
788
+ {
789
+ "epoch": 2.636457260556128,
790
+ "grad_norm": 1.1667654713458508,
791
+ "learning_rate": 5.554653014494176e-07,
792
+ "loss": 0.1031,
793
+ "step": 1120
794
+ },
795
+ {
796
+ "epoch": 2.659997057525379,
797
+ "grad_norm": 1.238846596793772,
798
+ "learning_rate": 4.855554408652985e-07,
799
+ "loss": 0.1039,
800
+ "step": 1130
801
+ },
802
+ {
803
+ "epoch": 2.68353685449463,
804
+ "grad_norm": 1.0962748439341052,
805
+ "learning_rate": 4.201990597632832e-07,
806
+ "loss": 0.1032,
807
+ "step": 1140
808
+ },
809
+ {
810
+ "epoch": 2.707076651463881,
811
+ "grad_norm": 1.1228066231093363,
812
+ "learning_rate": 3.5943858473009237e-07,
813
+ "loss": 0.105,
814
+ "step": 1150
815
+ },
816
+ {
817
+ "epoch": 2.7306164484331323,
818
+ "grad_norm": 1.1681919060053207,
819
+ "learning_rate": 3.033134588854289e-07,
820
+ "loss": 0.1029,
821
+ "step": 1160
822
+ },
823
+ {
824
+ "epoch": 2.7541562454023834,
825
+ "grad_norm": 1.107535834090668,
826
+ "learning_rate": 2.518601162771755e-07,
827
+ "loss": 0.1048,
828
+ "step": 1170
829
+ },
830
+ {
831
+ "epoch": 2.777696042371635,
832
+ "grad_norm": 1.0985797806813495,
833
+ "learning_rate": 2.0511195822997013e-07,
834
+ "loss": 0.102,
835
+ "step": 1180
836
+ },
837
+ {
838
+ "epoch": 2.801235839340886,
839
+ "grad_norm": 1.1563052202999995,
840
+ "learning_rate": 1.6309933166247403e-07,
841
+ "loss": 0.1028,
842
+ "step": 1190
843
+ },
844
+ {
845
+ "epoch": 2.824775636310137,
846
+ "grad_norm": 1.1299722934234167,
847
+ "learning_rate": 1.258495093874454e-07,
848
+ "loss": 0.1006,
849
+ "step": 1200
850
+ },
851
+ {
852
+ "epoch": 2.848315433279388,
853
+ "grad_norm": 1.0731524173833442,
854
+ "learning_rate": 9.338667240738619e-08,
855
+ "loss": 0.1029,
856
+ "step": 1210
857
+ },
858
+ {
859
+ "epoch": 2.871855230248639,
860
+ "grad_norm": 1.0991693593720597,
861
+ "learning_rate": 6.573189421726466e-08,
862
+ "loss": 0.102,
863
+ "step": 1220
864
+ },
865
+ {
866
+ "epoch": 2.8953950272178903,
867
+ "grad_norm": 1.1501966264186627,
868
+ "learning_rate": 4.2903127124496454e-08,
869
+ "loss": 0.1032,
870
+ "step": 1230
871
+ },
872
+ {
873
+ "epoch": 2.9189348241871413,
874
+ "grad_norm": 1.1357187349706255,
875
+ "learning_rate": 2.4915190595068415e-08,
876
+ "loss": 0.1031,
877
+ "step": 1240
878
+ },
879
+ {
880
+ "epoch": 2.942474621156393,
881
+ "grad_norm": 1.1136651822369734,
882
+ "learning_rate": 1.1779761633370633e-08,
883
+ "loss": 0.0998,
884
+ "step": 1250
885
+ },
886
+ {
887
+ "epoch": 2.966014418125644,
888
+ "grad_norm": 1.1212278780119938,
889
+ "learning_rate": 3.50536720197997e-09,
890
+ "loss": 0.1005,
891
+ "step": 1260
892
+ },
893
+ {
894
+ "epoch": 2.989554215094895,
895
+ "grad_norm": 1.1503104789107468,
896
+ "learning_rate": 9.73786863237014e-11,
897
+ "loss": 0.1019,
898
+ "step": 1270
899
+ },
900
+ {
901
+ "epoch": 2.994262174488745,
902
+ "step": 1272,
903
+ "total_flos": 76511791349760.0,
904
+ "train_loss": 0.29241873632015297,
905
+ "train_runtime": 70414.7871,
906
+ "train_samples_per_second": 1.158,
907
+ "train_steps_per_second": 0.018
908
+ }
909
+ ],
910
+ "logging_steps": 10,
911
+ "max_steps": 1272,
912
+ "num_input_tokens_seen": 0,
913
+ "num_train_epochs": 3,
914
+ "save_steps": 500,
915
+ "stateful_callbacks": {
916
+ "TrainerControl": {
917
+ "args": {
918
+ "should_epoch_stop": false,
919
+ "should_evaluate": false,
920
+ "should_log": false,
921
+ "should_save": true,
922
+ "should_training_stop": true
923
+ },
924
+ "attributes": {}
925
+ }
926
+ },
927
+ "total_flos": 76511791349760.0,
928
+ "train_batch_size": 1,
929
+ "trial_name": null,
930
+ "trial_params": null
931
+ }
training_loss.png ADDED