reinhardh commited on
Commit
b02b76b
·
verified ·
1 Parent(s): c23421e

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 +917 -0
  5. training_loss.png +0 -0
README.md CHANGED
@@ -4,6 +4,7 @@ license: apache-2.0
4
  base_model: Qwen/Qwen2.5-32B-Instruct
5
  tags:
6
  - llama-factory
 
7
  - generated_from_trainer
8
  model-index:
9
  - name: model-name-or-path_Qwen25-32B-Instruct_fig1_scaling_openthoughts_30k
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # model-name-or-path_Qwen25-32B-Instruct_fig1_scaling_openthoughts_30k
17
 
18
- This model is a fine-tuned version of [Qwen/Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) on an unknown dataset.
19
 
20
  ## Model description
21
 
 
4
  base_model: Qwen/Qwen2.5-32B-Instruct
5
  tags:
6
  - llama-factory
7
+ - full
8
  - generated_from_trainer
9
  model-index:
10
  - name: model-name-or-path_Qwen25-32B-Instruct_fig1_scaling_openthoughts_30k
 
16
 
17
  # model-name-or-path_Qwen25-32B-Instruct_fig1_scaling_openthoughts_30k
18
 
19
+ This model is a fine-tuned version of [Qwen/Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) on the mlfoundations-dev/fig1_scaling_openthoughts_30k dataset.
20
 
21
  ## Model description
22
 
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.9655172413793105,
3
+ "total_flos": 4.29306832130723e+18,
4
+ "train_loss": 0.40043015813827515,
5
+ "train_runtime": 25003.3066,
6
+ "train_samples_per_second": 2.597,
7
+ "train_steps_per_second": 0.005
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.9655172413793105,
3
+ "total_flos": 4.29306832130723e+18,
4
+ "train_loss": 0.40043015813827515,
5
+ "train_runtime": 25003.3066,
6
+ "train_samples_per_second": 2.597,
7
+ "train_steps_per_second": 0.005
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,917 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 4.9655172413793105,
5
+ "eval_steps": 500,
6
+ "global_step": 125,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.03940886699507389,
13
+ "grad_norm": 2.5584669687578936,
14
+ "learning_rate": 6.153846153846155e-06,
15
+ "loss": 0.7558,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.07881773399014778,
20
+ "grad_norm": 2.565466980202395,
21
+ "learning_rate": 1.230769230769231e-05,
22
+ "loss": 0.7528,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.11822660098522167,
27
+ "grad_norm": 1.8753957272537682,
28
+ "learning_rate": 1.8461538461538465e-05,
29
+ "loss": 0.7189,
30
+ "step": 3
31
+ },
32
+ {
33
+ "epoch": 0.15763546798029557,
34
+ "grad_norm": 1.5499772451700504,
35
+ "learning_rate": 2.461538461538462e-05,
36
+ "loss": 0.6998,
37
+ "step": 4
38
+ },
39
+ {
40
+ "epoch": 0.19704433497536947,
41
+ "grad_norm": 1.629703652668965,
42
+ "learning_rate": 3.0769230769230774e-05,
43
+ "loss": 0.639,
44
+ "step": 5
45
+ },
46
+ {
47
+ "epoch": 0.23645320197044334,
48
+ "grad_norm": 1.351435580335104,
49
+ "learning_rate": 3.692307692307693e-05,
50
+ "loss": 0.6071,
51
+ "step": 6
52
+ },
53
+ {
54
+ "epoch": 0.27586206896551724,
55
+ "grad_norm": 2.1491351662060074,
56
+ "learning_rate": 4.307692307692308e-05,
57
+ "loss": 0.6034,
58
+ "step": 7
59
+ },
60
+ {
61
+ "epoch": 0.31527093596059114,
62
+ "grad_norm": 1.1876259973559116,
63
+ "learning_rate": 4.923076923076924e-05,
64
+ "loss": 0.5724,
65
+ "step": 8
66
+ },
67
+ {
68
+ "epoch": 0.35467980295566504,
69
+ "grad_norm": 1.371610184643724,
70
+ "learning_rate": 5.538461538461539e-05,
71
+ "loss": 0.5544,
72
+ "step": 9
73
+ },
74
+ {
75
+ "epoch": 0.39408866995073893,
76
+ "grad_norm": 1.032169030947674,
77
+ "learning_rate": 6.153846153846155e-05,
78
+ "loss": 0.5503,
79
+ "step": 10
80
+ },
81
+ {
82
+ "epoch": 0.43349753694581283,
83
+ "grad_norm": 1.1747270299234402,
84
+ "learning_rate": 6.76923076923077e-05,
85
+ "loss": 0.5314,
86
+ "step": 11
87
+ },
88
+ {
89
+ "epoch": 0.4729064039408867,
90
+ "grad_norm": 0.9851373875214015,
91
+ "learning_rate": 7.384615384615386e-05,
92
+ "loss": 0.5293,
93
+ "step": 12
94
+ },
95
+ {
96
+ "epoch": 0.5123152709359606,
97
+ "grad_norm": 0.9498199365678511,
98
+ "learning_rate": 8e-05,
99
+ "loss": 0.5197,
100
+ "step": 13
101
+ },
102
+ {
103
+ "epoch": 0.5517241379310345,
104
+ "grad_norm": 0.6850321797374413,
105
+ "learning_rate": 7.998426505532213e-05,
106
+ "loss": 0.5108,
107
+ "step": 14
108
+ },
109
+ {
110
+ "epoch": 0.5911330049261084,
111
+ "grad_norm": 0.741883724345909,
112
+ "learning_rate": 7.993707260071268e-05,
113
+ "loss": 0.5039,
114
+ "step": 15
115
+ },
116
+ {
117
+ "epoch": 0.6305418719211823,
118
+ "grad_norm": 0.6201096559420001,
119
+ "learning_rate": 7.985845976470478e-05,
120
+ "loss": 0.492,
121
+ "step": 16
122
+ },
123
+ {
124
+ "epoch": 0.6699507389162561,
125
+ "grad_norm": 0.5810333918069511,
126
+ "learning_rate": 7.974848839572971e-05,
127
+ "loss": 0.4894,
128
+ "step": 17
129
+ },
130
+ {
131
+ "epoch": 0.7093596059113301,
132
+ "grad_norm": 0.538664331289255,
133
+ "learning_rate": 7.960724501345783e-05,
134
+ "loss": 0.478,
135
+ "step": 18
136
+ },
137
+ {
138
+ "epoch": 0.7487684729064039,
139
+ "grad_norm": 0.48351034605589627,
140
+ "learning_rate": 7.943484074072943e-05,
141
+ "loss": 0.4758,
142
+ "step": 19
143
+ },
144
+ {
145
+ "epoch": 0.7881773399014779,
146
+ "grad_norm": 0.5588522296796855,
147
+ "learning_rate": 7.923141121612922e-05,
148
+ "loss": 0.4755,
149
+ "step": 20
150
+ },
151
+ {
152
+ "epoch": 0.8275862068965517,
153
+ "grad_norm": 0.5099414849429755,
154
+ "learning_rate": 7.899711648727294e-05,
155
+ "loss": 0.4669,
156
+ "step": 21
157
+ },
158
+ {
159
+ "epoch": 0.8669950738916257,
160
+ "grad_norm": 0.5111926450042357,
161
+ "learning_rate": 7.873214088489047e-05,
162
+ "loss": 0.4634,
163
+ "step": 22
164
+ },
165
+ {
166
+ "epoch": 0.9064039408866995,
167
+ "grad_norm": 0.42484145211363433,
168
+ "learning_rate": 7.843669287780399e-05,
169
+ "loss": 0.4539,
170
+ "step": 23
171
+ },
172
+ {
173
+ "epoch": 0.9458128078817734,
174
+ "grad_norm": 0.4063070610232536,
175
+ "learning_rate": 7.811100490891586e-05,
176
+ "loss": 0.4545,
177
+ "step": 24
178
+ },
179
+ {
180
+ "epoch": 0.9852216748768473,
181
+ "grad_norm": 0.34890570659951814,
182
+ "learning_rate": 7.775533321233471e-05,
183
+ "loss": 0.4472,
184
+ "step": 25
185
+ },
186
+ {
187
+ "epoch": 1.0344827586206897,
188
+ "grad_norm": 0.4447080012153922,
189
+ "learning_rate": 7.736995761178399e-05,
190
+ "loss": 0.4415,
191
+ "step": 26
192
+ },
193
+ {
194
+ "epoch": 1.0738916256157636,
195
+ "grad_norm": 0.302803670590809,
196
+ "learning_rate": 7.695518130045147e-05,
197
+ "loss": 0.4339,
198
+ "step": 27
199
+ },
200
+ {
201
+ "epoch": 1.1133004926108374,
202
+ "grad_norm": 0.35078704995091015,
203
+ "learning_rate": 7.651133060245276e-05,
204
+ "loss": 0.4334,
205
+ "step": 28
206
+ },
207
+ {
208
+ "epoch": 1.1527093596059113,
209
+ "grad_norm": 0.347929638232275,
210
+ "learning_rate": 7.603875471609677e-05,
211
+ "loss": 0.4306,
212
+ "step": 29
213
+ },
214
+ {
215
+ "epoch": 1.1921182266009853,
216
+ "grad_norm": 0.34132416804989185,
217
+ "learning_rate": 7.55378254391549e-05,
218
+ "loss": 0.4303,
219
+ "step": 30
220
+ },
221
+ {
222
+ "epoch": 1.2315270935960592,
223
+ "grad_norm": 0.2780341991318012,
224
+ "learning_rate": 7.500893687635015e-05,
225
+ "loss": 0.4187,
226
+ "step": 31
227
+ },
228
+ {
229
+ "epoch": 1.270935960591133,
230
+ "grad_norm": 0.236156587933769,
231
+ "learning_rate": 7.445250512929637e-05,
232
+ "loss": 0.4163,
233
+ "step": 32
234
+ },
235
+ {
236
+ "epoch": 1.3103448275862069,
237
+ "grad_norm": 0.28381264582341237,
238
+ "learning_rate": 7.386896796913137e-05,
239
+ "loss": 0.4112,
240
+ "step": 33
241
+ },
242
+ {
243
+ "epoch": 1.3497536945812807,
244
+ "grad_norm": 0.222292178815377,
245
+ "learning_rate": 7.325878449210182e-05,
246
+ "loss": 0.4167,
247
+ "step": 34
248
+ },
249
+ {
250
+ "epoch": 1.3891625615763548,
251
+ "grad_norm": 0.35512090170091437,
252
+ "learning_rate": 7.262243475837041e-05,
253
+ "loss": 0.4109,
254
+ "step": 35
255
+ },
256
+ {
257
+ "epoch": 1.4285714285714286,
258
+ "grad_norm": 0.2511586648870301,
259
+ "learning_rate": 7.196041941432998e-05,
260
+ "loss": 0.4124,
261
+ "step": 36
262
+ },
263
+ {
264
+ "epoch": 1.4679802955665024,
265
+ "grad_norm": 0.22170782679952175,
266
+ "learning_rate": 7.12732592987212e-05,
267
+ "loss": 0.4056,
268
+ "step": 37
269
+ },
270
+ {
271
+ "epoch": 1.5073891625615765,
272
+ "grad_norm": 0.18785361140671847,
273
+ "learning_rate": 7.05614950328643e-05,
274
+ "loss": 0.4087,
275
+ "step": 38
276
+ },
277
+ {
278
+ "epoch": 1.5467980295566504,
279
+ "grad_norm": 0.20698082992129718,
280
+ "learning_rate": 6.982568659532663e-05,
281
+ "loss": 0.412,
282
+ "step": 39
283
+ },
284
+ {
285
+ "epoch": 1.5862068965517242,
286
+ "grad_norm": 0.2923360568345124,
287
+ "learning_rate": 6.906641288136109e-05,
288
+ "loss": 0.4077,
289
+ "step": 40
290
+ },
291
+ {
292
+ "epoch": 1.625615763546798,
293
+ "grad_norm": 0.36347305309861494,
294
+ "learning_rate": 6.828427124746191e-05,
295
+ "loss": 0.4135,
296
+ "step": 41
297
+ },
298
+ {
299
+ "epoch": 1.6650246305418719,
300
+ "grad_norm": 0.42985470055529523,
301
+ "learning_rate": 6.747987704139607e-05,
302
+ "loss": 0.408,
303
+ "step": 42
304
+ },
305
+ {
306
+ "epoch": 1.7044334975369457,
307
+ "grad_norm": 0.505450596225966,
308
+ "learning_rate": 6.665386311808017e-05,
309
+ "loss": 0.4125,
310
+ "step": 43
311
+ },
312
+ {
313
+ "epoch": 1.7438423645320196,
314
+ "grad_norm": 0.49870209403284876,
315
+ "learning_rate": 6.580687934168352e-05,
316
+ "loss": 0.4028,
317
+ "step": 44
318
+ },
319
+ {
320
+ "epoch": 1.7832512315270936,
321
+ "grad_norm": 0.3113343805759419,
322
+ "learning_rate": 6.493959207434934e-05,
323
+ "loss": 0.403,
324
+ "step": 45
325
+ },
326
+ {
327
+ "epoch": 1.8226600985221675,
328
+ "grad_norm": 0.2080418139358356,
329
+ "learning_rate": 6.405268365193624e-05,
330
+ "loss": 0.4143,
331
+ "step": 46
332
+ },
333
+ {
334
+ "epoch": 1.8620689655172413,
335
+ "grad_norm": 0.35947872339994125,
336
+ "learning_rate": 6.314685184719224e-05,
337
+ "loss": 0.3986,
338
+ "step": 47
339
+ },
340
+ {
341
+ "epoch": 1.9014778325123154,
342
+ "grad_norm": 0.2889476841026003,
343
+ "learning_rate": 6.22228093207841e-05,
344
+ "loss": 0.4034,
345
+ "step": 48
346
+ },
347
+ {
348
+ "epoch": 1.9408866995073892,
349
+ "grad_norm": 0.16926057758761603,
350
+ "learning_rate": 6.128128306061347e-05,
351
+ "loss": 0.398,
352
+ "step": 49
353
+ },
354
+ {
355
+ "epoch": 1.980295566502463,
356
+ "grad_norm": 0.3015239387441465,
357
+ "learning_rate": 6.0323013809861185e-05,
358
+ "loss": 0.3946,
359
+ "step": 50
360
+ },
361
+ {
362
+ "epoch": 2.0295566502463056,
363
+ "grad_norm": 0.27369640017832153,
364
+ "learning_rate": 5.9348755484209597e-05,
365
+ "loss": 0.3927,
366
+ "step": 51
367
+ },
368
+ {
369
+ "epoch": 2.0689655172413794,
370
+ "grad_norm": 0.1434834184101561,
371
+ "learning_rate": 5.835927457870151e-05,
372
+ "loss": 0.3811,
373
+ "step": 52
374
+ },
375
+ {
376
+ "epoch": 2.1083743842364533,
377
+ "grad_norm": 0.2854039627965572,
378
+ "learning_rate": 5.735534956470233e-05,
379
+ "loss": 0.3757,
380
+ "step": 53
381
+ },
382
+ {
383
+ "epoch": 2.147783251231527,
384
+ "grad_norm": 0.3009695490584673,
385
+ "learning_rate": 5.6337770277439854e-05,
386
+ "loss": 0.3818,
387
+ "step": 54
388
+ },
389
+ {
390
+ "epoch": 2.187192118226601,
391
+ "grad_norm": 0.154882062870217,
392
+ "learning_rate": 5.5307337294603595e-05,
393
+ "loss": 0.3758,
394
+ "step": 55
395
+ },
396
+ {
397
+ "epoch": 2.226600985221675,
398
+ "grad_norm": 0.23852846311077655,
399
+ "learning_rate": 5.4264861306492525e-05,
400
+ "loss": 0.3667,
401
+ "step": 56
402
+ },
403
+ {
404
+ "epoch": 2.2660098522167487,
405
+ "grad_norm": 0.22608925683357714,
406
+ "learning_rate": 5.321116247820669e-05,
407
+ "loss": 0.3711,
408
+ "step": 57
409
+ },
410
+ {
411
+ "epoch": 2.3054187192118225,
412
+ "grad_norm": 0.1583983272233887,
413
+ "learning_rate": 5.214706980438459e-05,
414
+ "loss": 0.367,
415
+ "step": 58
416
+ },
417
+ {
418
+ "epoch": 2.344827586206897,
419
+ "grad_norm": 0.205572865908627,
420
+ "learning_rate": 5.107342045699397e-05,
421
+ "loss": 0.3651,
422
+ "step": 59
423
+ },
424
+ {
425
+ "epoch": 2.3842364532019706,
426
+ "grad_norm": 0.2250613392996445,
427
+ "learning_rate": 4.999105912668908e-05,
428
+ "loss": 0.3723,
429
+ "step": 60
430
+ },
431
+ {
432
+ "epoch": 2.4236453201970445,
433
+ "grad_norm": 0.12838953670182573,
434
+ "learning_rate": 4.890083735825258e-05,
435
+ "loss": 0.3696,
436
+ "step": 61
437
+ },
438
+ {
439
+ "epoch": 2.4630541871921183,
440
+ "grad_norm": 0.2251988369582669,
441
+ "learning_rate": 4.780361288064514e-05,
442
+ "loss": 0.3676,
443
+ "step": 62
444
+ },
445
+ {
446
+ "epoch": 2.502463054187192,
447
+ "grad_norm": 0.1725969347073438,
448
+ "learning_rate": 4.670024893218946e-05,
449
+ "loss": 0.3697,
450
+ "step": 63
451
+ },
452
+ {
453
+ "epoch": 2.541871921182266,
454
+ "grad_norm": 0.12188885415056715,
455
+ "learning_rate": 4.5591613581419984e-05,
456
+ "loss": 0.3576,
457
+ "step": 64
458
+ },
459
+ {
460
+ "epoch": 2.58128078817734,
461
+ "grad_norm": 0.18927822293311972,
462
+ "learning_rate": 4.4478579044132314e-05,
463
+ "loss": 0.3667,
464
+ "step": 65
465
+ },
466
+ {
467
+ "epoch": 2.6206896551724137,
468
+ "grad_norm": 0.11561614146910962,
469
+ "learning_rate": 4.336202099716991e-05,
470
+ "loss": 0.3658,
471
+ "step": 66
472
+ },
473
+ {
474
+ "epoch": 2.6600985221674875,
475
+ "grad_norm": 0.1516544571012481,
476
+ "learning_rate": 4.2242817889487676e-05,
477
+ "loss": 0.3648,
478
+ "step": 67
479
+ },
480
+ {
481
+ "epoch": 2.6995073891625614,
482
+ "grad_norm": 0.1317279280788641,
483
+ "learning_rate": 4.112185025103476e-05,
484
+ "loss": 0.3615,
485
+ "step": 68
486
+ },
487
+ {
488
+ "epoch": 2.7389162561576352,
489
+ "grad_norm": 0.11760148271456246,
490
+ "learning_rate": 4e-05,
491
+ "loss": 0.3663,
492
+ "step": 69
493
+ },
494
+ {
495
+ "epoch": 2.7783251231527095,
496
+ "grad_norm": 0.12906163898612472,
497
+ "learning_rate": 3.8878149748965245e-05,
498
+ "loss": 0.3606,
499
+ "step": 70
500
+ },
501
+ {
502
+ "epoch": 2.8177339901477834,
503
+ "grad_norm": 0.09949229840426103,
504
+ "learning_rate": 3.775718211051233e-05,
505
+ "loss": 0.3667,
506
+ "step": 71
507
+ },
508
+ {
509
+ "epoch": 2.857142857142857,
510
+ "grad_norm": 0.12817295262296305,
511
+ "learning_rate": 3.6637979002830106e-05,
512
+ "loss": 0.3691,
513
+ "step": 72
514
+ },
515
+ {
516
+ "epoch": 2.896551724137931,
517
+ "grad_norm": 0.09926866475598084,
518
+ "learning_rate": 3.552142095586769e-05,
519
+ "loss": 0.3664,
520
+ "step": 73
521
+ },
522
+ {
523
+ "epoch": 2.935960591133005,
524
+ "grad_norm": 0.1188738983781825,
525
+ "learning_rate": 3.4408386418580036e-05,
526
+ "loss": 0.3717,
527
+ "step": 74
528
+ },
529
+ {
530
+ "epoch": 2.9753694581280787,
531
+ "grad_norm": 0.09857847772154998,
532
+ "learning_rate": 3.329975106781055e-05,
533
+ "loss": 0.358,
534
+ "step": 75
535
+ },
536
+ {
537
+ "epoch": 3.0246305418719213,
538
+ "grad_norm": 0.12437210969829017,
539
+ "learning_rate": 3.219638711935488e-05,
540
+ "loss": 0.3487,
541
+ "step": 76
542
+ },
543
+ {
544
+ "epoch": 3.064039408866995,
545
+ "grad_norm": 0.13092335355733034,
546
+ "learning_rate": 3.109916264174743e-05,
547
+ "loss": 0.3378,
548
+ "step": 77
549
+ },
550
+ {
551
+ "epoch": 3.103448275862069,
552
+ "grad_norm": 0.1380029588610029,
553
+ "learning_rate": 3.000894087331092e-05,
554
+ "loss": 0.3471,
555
+ "step": 78
556
+ },
557
+ {
558
+ "epoch": 3.142857142857143,
559
+ "grad_norm": 0.16659609357249366,
560
+ "learning_rate": 2.892657954300603e-05,
561
+ "loss": 0.345,
562
+ "step": 79
563
+ },
564
+ {
565
+ "epoch": 3.1822660098522166,
566
+ "grad_norm": 0.1491210789680348,
567
+ "learning_rate": 2.7852930195615413e-05,
568
+ "loss": 0.3423,
569
+ "step": 80
570
+ },
571
+ {
572
+ "epoch": 3.2216748768472905,
573
+ "grad_norm": 0.13323974852056628,
574
+ "learning_rate": 2.678883752179333e-05,
575
+ "loss": 0.3375,
576
+ "step": 81
577
+ },
578
+ {
579
+ "epoch": 3.2610837438423648,
580
+ "grad_norm": 0.16071703950070282,
581
+ "learning_rate": 2.573513869350748e-05,
582
+ "loss": 0.3449,
583
+ "step": 82
584
+ },
585
+ {
586
+ "epoch": 3.3004926108374386,
587
+ "grad_norm": 0.11154281193190216,
588
+ "learning_rate": 2.4692662705396412e-05,
589
+ "loss": 0.3477,
590
+ "step": 83
591
+ },
592
+ {
593
+ "epoch": 3.3399014778325125,
594
+ "grad_norm": 0.13730160221975535,
595
+ "learning_rate": 2.366222972256016e-05,
596
+ "loss": 0.3393,
597
+ "step": 84
598
+ },
599
+ {
600
+ "epoch": 3.3793103448275863,
601
+ "grad_norm": 0.1001261256156619,
602
+ "learning_rate": 2.264465043529768e-05,
603
+ "loss": 0.3386,
604
+ "step": 85
605
+ },
606
+ {
607
+ "epoch": 3.41871921182266,
608
+ "grad_norm": 0.11871409719140899,
609
+ "learning_rate": 2.1640725421298487e-05,
610
+ "loss": 0.3413,
611
+ "step": 86
612
+ },
613
+ {
614
+ "epoch": 3.458128078817734,
615
+ "grad_norm": 0.10822197872193218,
616
+ "learning_rate": 2.065124451579041e-05,
617
+ "loss": 0.3415,
618
+ "step": 87
619
+ },
620
+ {
621
+ "epoch": 3.497536945812808,
622
+ "grad_norm": 0.09848707693593217,
623
+ "learning_rate": 1.9676986190138835e-05,
624
+ "loss": 0.3424,
625
+ "step": 88
626
+ },
627
+ {
628
+ "epoch": 3.5369458128078817,
629
+ "grad_norm": 0.10650565727211944,
630
+ "learning_rate": 1.8718716939386543e-05,
631
+ "loss": 0.3438,
632
+ "step": 89
633
+ },
634
+ {
635
+ "epoch": 3.5763546798029555,
636
+ "grad_norm": 0.09005212054070572,
637
+ "learning_rate": 1.7777190679215923e-05,
638
+ "loss": 0.3413,
639
+ "step": 90
640
+ },
641
+ {
642
+ "epoch": 3.6157635467980294,
643
+ "grad_norm": 0.09471297169488595,
644
+ "learning_rate": 1.6853148152807774e-05,
645
+ "loss": 0.3394,
646
+ "step": 91
647
+ },
648
+ {
649
+ "epoch": 3.655172413793103,
650
+ "grad_norm": 0.08803273896652859,
651
+ "learning_rate": 1.5947316348063764e-05,
652
+ "loss": 0.3452,
653
+ "step": 92
654
+ },
655
+ {
656
+ "epoch": 3.6945812807881775,
657
+ "grad_norm": 0.08555929815288703,
658
+ "learning_rate": 1.5060407925650662e-05,
659
+ "loss": 0.3386,
660
+ "step": 93
661
+ },
662
+ {
663
+ "epoch": 3.7339901477832513,
664
+ "grad_norm": 0.07676232199734123,
665
+ "learning_rate": 1.4193120658316506e-05,
666
+ "loss": 0.3384,
667
+ "step": 94
668
+ },
669
+ {
670
+ "epoch": 3.773399014778325,
671
+ "grad_norm": 0.07008988332755012,
672
+ "learning_rate": 1.3346136881919845e-05,
673
+ "loss": 0.3423,
674
+ "step": 95
675
+ },
676
+ {
677
+ "epoch": 3.812807881773399,
678
+ "grad_norm": 0.07699181822522967,
679
+ "learning_rate": 1.2520122958603933e-05,
680
+ "loss": 0.3394,
681
+ "step": 96
682
+ },
683
+ {
684
+ "epoch": 3.852216748768473,
685
+ "grad_norm": 0.07037612515608051,
686
+ "learning_rate": 1.1715728752538103e-05,
687
+ "loss": 0.3377,
688
+ "step": 97
689
+ },
690
+ {
691
+ "epoch": 3.8916256157635467,
692
+ "grad_norm": 0.06692075874239917,
693
+ "learning_rate": 1.0933587118638927e-05,
694
+ "loss": 0.3389,
695
+ "step": 98
696
+ },
697
+ {
698
+ "epoch": 3.9310344827586206,
699
+ "grad_norm": 0.07239315796297541,
700
+ "learning_rate": 1.0174313404673378e-05,
701
+ "loss": 0.3386,
702
+ "step": 99
703
+ },
704
+ {
705
+ "epoch": 3.970443349753695,
706
+ "grad_norm": 0.06509354323682291,
707
+ "learning_rate": 9.438504967135703e-06,
708
+ "loss": 0.3435,
709
+ "step": 100
710
+ },
711
+ {
712
+ "epoch": 4.019704433497537,
713
+ "grad_norm": 0.07415226571205577,
714
+ "learning_rate": 8.72674070127881e-06,
715
+ "loss": 0.3299,
716
+ "step": 101
717
+ },
718
+ {
719
+ "epoch": 4.059113300492611,
720
+ "grad_norm": 0.09427673195861289,
721
+ "learning_rate": 8.039580585670047e-06,
722
+ "loss": 0.3271,
723
+ "step": 102
724
+ },
725
+ {
726
+ "epoch": 4.098522167487685,
727
+ "grad_norm": 0.07768702640271646,
728
+ "learning_rate": 7.3775652416295936e-06,
729
+ "loss": 0.3298,
730
+ "step": 103
731
+ },
732
+ {
733
+ "epoch": 4.137931034482759,
734
+ "grad_norm": 0.07261233473354416,
735
+ "learning_rate": 6.7412155078981865e-06,
736
+ "loss": 0.3295,
737
+ "step": 104
738
+ },
739
+ {
740
+ "epoch": 4.177339901477833,
741
+ "grad_norm": 0.07340342692044237,
742
+ "learning_rate": 6.1310320308686354e-06,
743
+ "loss": 0.3274,
744
+ "step": 105
745
+ },
746
+ {
747
+ "epoch": 4.216748768472907,
748
+ "grad_norm": 0.07153254469501634,
749
+ "learning_rate": 5.547494870703642e-06,
750
+ "loss": 0.3307,
751
+ "step": 106
752
+ },
753
+ {
754
+ "epoch": 4.25615763546798,
755
+ "grad_norm": 0.07858955565979121,
756
+ "learning_rate": 4.991063123649853e-06,
757
+ "loss": 0.326,
758
+ "step": 107
759
+ },
760
+ {
761
+ "epoch": 4.295566502463054,
762
+ "grad_norm": 0.0747794841861464,
763
+ "learning_rate": 4.462174560845114e-06,
764
+ "loss": 0.3273,
765
+ "step": 108
766
+ },
767
+ {
768
+ "epoch": 4.334975369458128,
769
+ "grad_norm": 0.06814386298200781,
770
+ "learning_rate": 3.961245283903239e-06,
771
+ "loss": 0.3304,
772
+ "step": 109
773
+ },
774
+ {
775
+ "epoch": 4.374384236453202,
776
+ "grad_norm": 0.06709307703349002,
777
+ "learning_rate": 3.4886693975472443e-06,
778
+ "loss": 0.3285,
779
+ "step": 110
780
+ },
781
+ {
782
+ "epoch": 4.413793103448276,
783
+ "grad_norm": 0.06195820663002009,
784
+ "learning_rate": 3.0448186995485307e-06,
785
+ "loss": 0.328,
786
+ "step": 111
787
+ },
788
+ {
789
+ "epoch": 4.45320197044335,
790
+ "grad_norm": 0.06590977322439409,
791
+ "learning_rate": 2.630042388216012e-06,
792
+ "loss": 0.3314,
793
+ "step": 112
794
+ },
795
+ {
796
+ "epoch": 4.4926108374384235,
797
+ "grad_norm": 0.06216983243982217,
798
+ "learning_rate": 2.244666787665297e-06,
799
+ "loss": 0.3285,
800
+ "step": 113
801
+ },
802
+ {
803
+ "epoch": 4.532019704433497,
804
+ "grad_norm": 0.0568288191054022,
805
+ "learning_rate": 1.888995091084147e-06,
806
+ "loss": 0.3267,
807
+ "step": 114
808
+ },
809
+ {
810
+ "epoch": 4.571428571428571,
811
+ "grad_norm": 0.05502800210878154,
812
+ "learning_rate": 1.5633071221960205e-06,
813
+ "loss": 0.3199,
814
+ "step": 115
815
+ },
816
+ {
817
+ "epoch": 4.610837438423645,
818
+ "grad_norm": 0.05085194354160944,
819
+ "learning_rate": 1.2678591151095466e-06,
820
+ "loss": 0.3291,
821
+ "step": 116
822
+ },
823
+ {
824
+ "epoch": 4.650246305418719,
825
+ "grad_norm": 0.0542030183423217,
826
+ "learning_rate": 1.0028835127270553e-06,
827
+ "loss": 0.328,
828
+ "step": 117
829
+ },
830
+ {
831
+ "epoch": 4.689655172413794,
832
+ "grad_norm": 0.04943720054109848,
833
+ "learning_rate": 7.685887838707828e-07,
834
+ "loss": 0.3282,
835
+ "step": 118
836
+ },
837
+ {
838
+ "epoch": 4.7290640394088665,
839
+ "grad_norm": 0.05350280931369455,
840
+ "learning_rate": 5.651592592705646e-07,
841
+ "loss": 0.3358,
842
+ "step": 119
843
+ },
844
+ {
845
+ "epoch": 4.768472906403941,
846
+ "grad_norm": 0.051675651653332454,
847
+ "learning_rate": 3.9275498654217425e-07,
848
+ "loss": 0.3231,
849
+ "step": 120
850
+ },
851
+ {
852
+ "epoch": 4.807881773399015,
853
+ "grad_norm": 0.049565582110987425,
854
+ "learning_rate": 2.5151160427029584e-07,
855
+ "loss": 0.3273,
856
+ "step": 121
857
+ },
858
+ {
859
+ "epoch": 4.847290640394089,
860
+ "grad_norm": 0.05086813508645347,
861
+ "learning_rate": 1.4154023529523663e-07,
862
+ "loss": 0.3252,
863
+ "step": 122
864
+ },
865
+ {
866
+ "epoch": 4.886699507389163,
867
+ "grad_norm": 0.04981788886449752,
868
+ "learning_rate": 6.292739928733582e-08,
869
+ "loss": 0.3251,
870
+ "step": 123
871
+ },
872
+ {
873
+ "epoch": 4.926108374384237,
874
+ "grad_norm": 0.051983785964860994,
875
+ "learning_rate": 1.5734944677885388e-08,
876
+ "loss": 0.3261,
877
+ "step": 124
878
+ },
879
+ {
880
+ "epoch": 4.9655172413793105,
881
+ "grad_norm": 0.05048075450561713,
882
+ "learning_rate": 0.0,
883
+ "loss": 0.3278,
884
+ "step": 125
885
+ },
886
+ {
887
+ "epoch": 4.9655172413793105,
888
+ "step": 125,
889
+ "total_flos": 4.29306832130723e+18,
890
+ "train_loss": 0.40043015813827515,
891
+ "train_runtime": 25003.3066,
892
+ "train_samples_per_second": 2.597,
893
+ "train_steps_per_second": 0.005
894
+ }
895
+ ],
896
+ "logging_steps": 1,
897
+ "max_steps": 125,
898
+ "num_input_tokens_seen": 0,
899
+ "num_train_epochs": 5,
900
+ "save_steps": 500,
901
+ "stateful_callbacks": {
902
+ "TrainerControl": {
903
+ "args": {
904
+ "should_epoch_stop": false,
905
+ "should_evaluate": false,
906
+ "should_log": false,
907
+ "should_save": true,
908
+ "should_training_stop": true
909
+ },
910
+ "attributes": {}
911
+ }
912
+ },
913
+ "total_flos": 4.29306832130723e+18,
914
+ "train_batch_size": 1,
915
+ "trial_name": null,
916
+ "trial_params": null
917
+ }
training_loss.png ADDED