christineyu commited on
Commit
edde79e
1 Parent(s): c2983e1

Model save

Browse files
Files changed (4) hide show
  1. README.md +57 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +925 -0
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: mistralai/Mistral-7B-Instruct-v0.2
3
+ library_name: transformers
4
+ model_name: prometheus-7b-v1.5-beta-1-overfit
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for prometheus-7b-v1.5-beta-1-overfit
13
+
14
+ This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="christineyu/prometheus-7b-v1.5-beta-1-overfit", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/myexp/huggingface/runs/6ne9m82s)
31
+
32
+ This model was trained with SFT.
33
+
34
+ ### Framework versions
35
+
36
+ - TRL: 0.12.2
37
+ - Transformers: 4.46.3
38
+ - Pytorch: 2.3.0
39
+ - Datasets: 3.2.0
40
+ - Tokenizers: 0.20.3
41
+
42
+ ## Citations
43
+
44
+
45
+
46
+ Cite TRL as:
47
+
48
+ ```bibtex
49
+ @misc{vonwerra2022trl,
50
+ title = {{TRL: Transformer Reinforcement Learning}},
51
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
52
+ year = 2020,
53
+ journal = {GitHub repository},
54
+ publisher = {GitHub},
55
+ howpublished = {\url{https://github.com/huggingface/trl}}
56
+ }
57
+ ```
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 5.590502947474637e+16,
4
+ "train_loss": 0.06471517855121244,
5
+ "train_runtime": 1309.8316,
6
+ "train_samples": 990,
7
+ "train_samples_per_second": 0.473,
8
+ "train_steps_per_second": 0.473
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 5.590502947474637e+16,
4
+ "train_loss": 0.06471517855121244,
5
+ "train_runtime": 1309.8316,
6
+ "train_samples": 990,
7
+ "train_samples_per_second": 0.473,
8
+ "train_steps_per_second": 0.473
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,925 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 620,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0016129032258064516,
13
+ "grad_norm": 8.676656723022461,
14
+ "learning_rate": 1.6129032258064516e-06,
15
+ "loss": 1.2418,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.008064516129032258,
20
+ "grad_norm": 4.749965190887451,
21
+ "learning_rate": 8.064516129032258e-06,
22
+ "loss": 1.1522,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.016129032258064516,
27
+ "grad_norm": 4.172841548919678,
28
+ "learning_rate": 1.6129032258064517e-05,
29
+ "loss": 0.9234,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.024193548387096774,
34
+ "grad_norm": 3.307924747467041,
35
+ "learning_rate": 2.4193548387096777e-05,
36
+ "loss": 0.5665,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.03225806451612903,
41
+ "grad_norm": 3.2205514907836914,
42
+ "learning_rate": 3.2258064516129034e-05,
43
+ "loss": 0.205,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.04032258064516129,
48
+ "grad_norm": 1.3202248811721802,
49
+ "learning_rate": 4.032258064516129e-05,
50
+ "loss": 0.0567,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.04838709677419355,
55
+ "grad_norm": 2.270284652709961,
56
+ "learning_rate": 4.8387096774193554e-05,
57
+ "loss": 0.0348,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.056451612903225805,
62
+ "grad_norm": 1.9001686573028564,
63
+ "learning_rate": 5.645161290322582e-05,
64
+ "loss": 0.0256,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.06451612903225806,
69
+ "grad_norm": 0.8090413212776184,
70
+ "learning_rate": 6.451612903225807e-05,
71
+ "loss": 0.0234,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.07258064516129033,
76
+ "grad_norm": 0.7908226251602173,
77
+ "learning_rate": 7.258064516129033e-05,
78
+ "loss": 0.0177,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.08064516129032258,
83
+ "grad_norm": 3.9440441131591797,
84
+ "learning_rate": 8.064516129032258e-05,
85
+ "loss": 0.0148,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.08870967741935484,
90
+ "grad_norm": 0.7870569825172424,
91
+ "learning_rate": 8.870967741935484e-05,
92
+ "loss": 0.0186,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.0967741935483871,
97
+ "grad_norm": 1.7015715837478638,
98
+ "learning_rate": 9.677419354838711e-05,
99
+ "loss": 0.0197,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.10483870967741936,
104
+ "grad_norm": 0.7983545660972595,
105
+ "learning_rate": 9.99928681279855e-05,
106
+ "loss": 0.0148,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.11290322580645161,
111
+ "grad_norm": 1.1775078773498535,
112
+ "learning_rate": 9.994929183335238e-05,
113
+ "loss": 0.0139,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.12096774193548387,
118
+ "grad_norm": 4.544706344604492,
119
+ "learning_rate": 9.986613588305434e-05,
120
+ "loss": 0.0201,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.12903225806451613,
125
+ "grad_norm": 1.9559928178787231,
126
+ "learning_rate": 9.974346616959476e-05,
127
+ "loss": 0.0157,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.13709677419354838,
132
+ "grad_norm": 1.6153545379638672,
133
+ "learning_rate": 9.95813798960538e-05,
134
+ "loss": 0.0188,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.14516129032258066,
139
+ "grad_norm": 300.6769104003906,
140
+ "learning_rate": 9.938000549906508e-05,
141
+ "loss": 0.0761,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.1532258064516129,
146
+ "grad_norm": 0.5071149468421936,
147
+ "learning_rate": 9.913950254704291e-05,
148
+ "loss": 0.0236,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.16129032258064516,
153
+ "grad_norm": 1.064196228981018,
154
+ "learning_rate": 9.88600616137407e-05,
155
+ "loss": 1.1514,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.1693548387096774,
160
+ "grad_norm": 415.9707946777344,
161
+ "learning_rate": 9.854190412724114e-05,
162
+ "loss": 0.0835,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.1774193548387097,
167
+ "grad_norm": 1.221725583076477,
168
+ "learning_rate": 9.818528219449705e-05,
169
+ "loss": 0.0189,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.18548387096774194,
174
+ "grad_norm": 0.40028271079063416,
175
+ "learning_rate": 9.779047840156288e-05,
176
+ "loss": 0.0197,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.1935483870967742,
181
+ "grad_norm": 3.2106378078460693,
182
+ "learning_rate": 9.735780558967434e-05,
183
+ "loss": 1.0777,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.20161290322580644,
188
+ "grad_norm": 544.4739379882812,
189
+ "learning_rate": 9.688760660735402e-05,
190
+ "loss": 0.0779,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.20967741935483872,
195
+ "grad_norm": 0.48118099570274353,
196
+ "learning_rate": 9.63802540387394e-05,
197
+ "loss": 0.7549,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.21774193548387097,
202
+ "grad_norm": 0.43920445442199707,
203
+ "learning_rate": 9.583614990834828e-05,
204
+ "loss": 0.0193,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.22580645161290322,
209
+ "grad_norm": 1.4750373363494873,
210
+ "learning_rate": 9.525572536251607e-05,
211
+ "loss": 0.02,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.23387096774193547,
216
+ "grad_norm": 2.262516498565674,
217
+ "learning_rate": 9.463944032775658e-05,
218
+ "loss": 0.0216,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.24193548387096775,
223
+ "grad_norm": 1.0215312242507935,
224
+ "learning_rate": 9.398778314631801e-05,
225
+ "loss": 0.018,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.25,
230
+ "grad_norm": 0.8741445541381836,
231
+ "learning_rate": 9.330127018922194e-05,
232
+ "loss": 0.0178,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.25806451612903225,
237
+ "grad_norm": 0.5452761650085449,
238
+ "learning_rate": 9.258044544709274e-05,
239
+ "loss": 0.0323,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.2661290322580645,
244
+ "grad_norm": 0.43162572383880615,
245
+ "learning_rate": 9.182588009910118e-05,
246
+ "loss": 0.0201,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.27419354838709675,
251
+ "grad_norm": 0.25898757576942444,
252
+ "learning_rate": 9.103817206036382e-05,
253
+ "loss": 0.0198,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.28225806451612906,
258
+ "grad_norm": 2.178893566131592,
259
+ "learning_rate": 9.021794550815713e-05,
260
+ "loss": 0.0159,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.2903225806451613,
265
+ "grad_norm": 145.104736328125,
266
+ "learning_rate": 8.936585038732141e-05,
267
+ "loss": 0.5195,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.29838709677419356,
272
+ "grad_norm": 6.8297834396362305,
273
+ "learning_rate": 8.848256189524662e-05,
274
+ "loss": 0.021,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.3064516129032258,
279
+ "grad_norm": 1.4729770421981812,
280
+ "learning_rate": 8.756877994684818e-05,
281
+ "loss": 0.0159,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.31451612903225806,
286
+ "grad_norm": 2.1729257106781006,
287
+ "learning_rate": 8.66252286199567e-05,
288
+ "loss": 0.0152,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.3225806451612903,
293
+ "grad_norm": 0.6754555106163025,
294
+ "learning_rate": 8.565265558156101e-05,
295
+ "loss": 0.0117,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.33064516129032256,
300
+ "grad_norm": 0.3731600046157837,
301
+ "learning_rate": 8.46518314953594e-05,
302
+ "loss": 0.0101,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.3387096774193548,
307
+ "grad_norm": 0.3129900395870209,
308
+ "learning_rate": 8.362354941108802e-05,
309
+ "loss": 0.0161,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.3467741935483871,
314
+ "grad_norm": 0.3701813519001007,
315
+ "learning_rate": 8.256862413611113e-05,
316
+ "loss": 0.0156,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.3548387096774194,
321
+ "grad_norm": 10.514623641967773,
322
+ "learning_rate": 8.148789158977013e-05,
323
+ "loss": 0.0072,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.3629032258064516,
328
+ "grad_norm": 0.25649479031562805,
329
+ "learning_rate": 8.038220814100403e-05,
330
+ "loss": 0.0089,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.3709677419354839,
335
+ "grad_norm": 0.36153924465179443,
336
+ "learning_rate": 7.925244992976538e-05,
337
+ "loss": 0.0091,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.3790322580645161,
342
+ "grad_norm": 0.15555064380168915,
343
+ "learning_rate": 7.809951217276985e-05,
344
+ "loss": 0.1414,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.3870967741935484,
349
+ "grad_norm": 0.5916360020637512,
350
+ "learning_rate": 7.692430845412945e-05,
351
+ "loss": 0.0153,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.3951612903225806,
356
+ "grad_norm": 0.47686517238616943,
357
+ "learning_rate": 7.572777000143145e-05,
358
+ "loss": 0.0154,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.4032258064516129,
363
+ "grad_norm": 0.5245814919471741,
364
+ "learning_rate": 7.451084494783668e-05,
365
+ "loss": 0.02,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.4112903225806452,
370
+ "grad_norm": 3.145771026611328,
371
+ "learning_rate": 7.327449758078193e-05,
372
+ "loss": 0.0121,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.41935483870967744,
377
+ "grad_norm": 0.47086361050605774,
378
+ "learning_rate": 7.201970757788172e-05,
379
+ "loss": 0.01,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.4274193548387097,
384
+ "grad_norm": 0.8778440356254578,
385
+ "learning_rate": 7.074746923063497e-05,
386
+ "loss": 0.0083,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.43548387096774194,
391
+ "grad_norm": 0.1912572830915451,
392
+ "learning_rate": 6.945879065655165e-05,
393
+ "loss": 0.018,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.4435483870967742,
398
+ "grad_norm": 0.3541414737701416,
399
+ "learning_rate": 6.815469300032374e-05,
400
+ "loss": 0.0103,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.45161290322580644,
405
+ "grad_norm": 0.39805546402931213,
406
+ "learning_rate": 6.683620962467357e-05,
407
+ "loss": 0.0123,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.4596774193548387,
412
+ "grad_norm": 0.3862231373786926,
413
+ "learning_rate": 6.550438529152056e-05,
414
+ "loss": 0.0085,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.46774193548387094,
419
+ "grad_norm": 0.7873578071594238,
420
+ "learning_rate": 6.416027533411519e-05,
421
+ "loss": 0.0078,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.47580645161290325,
426
+ "grad_norm": 0.9767886996269226,
427
+ "learning_rate": 6.28049448207966e-05,
428
+ "loss": 0.0089,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.4838709677419355,
433
+ "grad_norm": 0.659842312335968,
434
+ "learning_rate": 6.14394677110356e-05,
435
+ "loss": 0.0085,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.49193548387096775,
440
+ "grad_norm": 0.15667948126792908,
441
+ "learning_rate": 6.006492600443301e-05,
442
+ "loss": 0.0106,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 0.5,
447
+ "grad_norm": 0.37710684537887573,
448
+ "learning_rate": 5.868240888334653e-05,
449
+ "loss": 0.0125,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 0.5080645161290323,
454
+ "grad_norm": 0.22124210000038147,
455
+ "learning_rate": 5.7293011849826217e-05,
456
+ "loss": 0.0121,
457
+ "step": 315
458
+ },
459
+ {
460
+ "epoch": 0.5161290322580645,
461
+ "grad_norm": 0.894884467124939,
462
+ "learning_rate": 5.5897835857542317e-05,
463
+ "loss": 0.0059,
464
+ "step": 320
465
+ },
466
+ {
467
+ "epoch": 0.5241935483870968,
468
+ "grad_norm": 0.2709536850452423,
469
+ "learning_rate": 5.449798643939304e-05,
470
+ "loss": 0.0074,
471
+ "step": 325
472
+ },
473
+ {
474
+ "epoch": 0.532258064516129,
475
+ "grad_norm": 0.15670421719551086,
476
+ "learning_rate": 5.30945728314841e-05,
477
+ "loss": 0.0131,
478
+ "step": 330
479
+ },
480
+ {
481
+ "epoch": 0.5403225806451613,
482
+ "grad_norm": 0.31018203496932983,
483
+ "learning_rate": 5.168870709417342e-05,
484
+ "loss": 0.0101,
485
+ "step": 335
486
+ },
487
+ {
488
+ "epoch": 0.5483870967741935,
489
+ "grad_norm": 0.20672403275966644,
490
+ "learning_rate": 5.02815032308783e-05,
491
+ "loss": 0.0057,
492
+ "step": 340
493
+ },
494
+ {
495
+ "epoch": 0.5564516129032258,
496
+ "grad_norm": 0.18754148483276367,
497
+ "learning_rate": 4.887407630534271e-05,
498
+ "loss": 0.0063,
499
+ "step": 345
500
+ },
501
+ {
502
+ "epoch": 0.5645161290322581,
503
+ "grad_norm": 0.29782626032829285,
504
+ "learning_rate": 4.746754155806437e-05,
505
+ "loss": 0.0071,
506
+ "step": 350
507
+ },
508
+ {
509
+ "epoch": 0.5725806451612904,
510
+ "grad_norm": 0.1842086762189865,
511
+ "learning_rate": 4.6063013522581915e-05,
512
+ "loss": 0.012,
513
+ "step": 355
514
+ },
515
+ {
516
+ "epoch": 0.5806451612903226,
517
+ "grad_norm": 3.5033631324768066,
518
+ "learning_rate": 4.466160514232206e-05,
519
+ "loss": 0.0071,
520
+ "step": 360
521
+ },
522
+ {
523
+ "epoch": 0.5887096774193549,
524
+ "grad_norm": 0.11994602531194687,
525
+ "learning_rate": 4.3264426888706965e-05,
526
+ "loss": 0.0056,
527
+ "step": 365
528
+ },
529
+ {
530
+ "epoch": 0.5967741935483871,
531
+ "grad_norm": 0.21798093616962433,
532
+ "learning_rate": 4.187258588122019e-05,
533
+ "loss": 0.0096,
534
+ "step": 370
535
+ },
536
+ {
537
+ "epoch": 0.6048387096774194,
538
+ "grad_norm": 1.366426944732666,
539
+ "learning_rate": 4.048718501012895e-05,
540
+ "loss": 0.0113,
541
+ "step": 375
542
+ },
543
+ {
544
+ "epoch": 0.6129032258064516,
545
+ "grad_norm": 0.2101077437400818,
546
+ "learning_rate": 3.9109322062557424e-05,
547
+ "loss": 0.0105,
548
+ "step": 380
549
+ },
550
+ {
551
+ "epoch": 0.6209677419354839,
552
+ "grad_norm": 0.2682858407497406,
553
+ "learning_rate": 3.77400888526038e-05,
554
+ "loss": 0.0079,
555
+ "step": 385
556
+ },
557
+ {
558
+ "epoch": 0.6290322580645161,
559
+ "grad_norm": 0.3401815593242645,
560
+ "learning_rate": 3.6380570356190344e-05,
561
+ "loss": 0.0087,
562
+ "step": 390
563
+ },
564
+ {
565
+ "epoch": 0.6370967741935484,
566
+ "grad_norm": 1.0749493837356567,
567
+ "learning_rate": 3.5031843851332104e-05,
568
+ "loss": 0.0077,
569
+ "step": 395
570
+ },
571
+ {
572
+ "epoch": 0.6451612903225806,
573
+ "grad_norm": 0.600229024887085,
574
+ "learning_rate": 3.3694978064505254e-05,
575
+ "loss": 0.0088,
576
+ "step": 400
577
+ },
578
+ {
579
+ "epoch": 0.6532258064516129,
580
+ "grad_norm": 11.734594345092773,
581
+ "learning_rate": 3.2371032323791754e-05,
582
+ "loss": 0.0107,
583
+ "step": 405
584
+ },
585
+ {
586
+ "epoch": 0.6612903225806451,
587
+ "grad_norm": 0.9927507638931274,
588
+ "learning_rate": 3.10610557194712e-05,
589
+ "loss": 0.0146,
590
+ "step": 410
591
+ },
592
+ {
593
+ "epoch": 0.6693548387096774,
594
+ "grad_norm": 0.46601602435112,
595
+ "learning_rate": 2.9766086272725196e-05,
596
+ "loss": 0.0123,
597
+ "step": 415
598
+ },
599
+ {
600
+ "epoch": 0.6774193548387096,
601
+ "grad_norm": 11.156332969665527,
602
+ "learning_rate": 2.848715011311271e-05,
603
+ "loss": 0.0134,
604
+ "step": 420
605
+ },
606
+ {
607
+ "epoch": 0.6854838709677419,
608
+ "grad_norm": 0.34076985716819763,
609
+ "learning_rate": 2.72252606654683e-05,
610
+ "loss": 0.011,
611
+ "step": 425
612
+ },
613
+ {
614
+ "epoch": 0.6935483870967742,
615
+ "grad_norm": 0.14691153168678284,
616
+ "learning_rate": 2.598141784686775e-05,
617
+ "loss": 0.0113,
618
+ "step": 430
619
+ },
620
+ {
621
+ "epoch": 0.7016129032258065,
622
+ "grad_norm": 0.19039344787597656,
623
+ "learning_rate": 2.4756607274296846e-05,
624
+ "loss": 0.0066,
625
+ "step": 435
626
+ },
627
+ {
628
+ "epoch": 0.7096774193548387,
629
+ "grad_norm": 0.4377165734767914,
630
+ "learning_rate": 2.3551799483651894e-05,
631
+ "loss": 0.0056,
632
+ "step": 440
633
+ },
634
+ {
635
+ "epoch": 0.717741935483871,
636
+ "grad_norm": 0.16886833310127258,
637
+ "learning_rate": 2.2367949160690072e-05,
638
+ "loss": 0.0075,
639
+ "step": 445
640
+ },
641
+ {
642
+ "epoch": 0.7258064516129032,
643
+ "grad_norm": 5.967025279998779,
644
+ "learning_rate": 2.120599438453968e-05,
645
+ "loss": 0.0083,
646
+ "step": 450
647
+ },
648
+ {
649
+ "epoch": 0.7338709677419355,
650
+ "grad_norm": 0.15985387563705444,
651
+ "learning_rate": 2.0066855884369244e-05,
652
+ "loss": 0.004,
653
+ "step": 455
654
+ },
655
+ {
656
+ "epoch": 0.7419354838709677,
657
+ "grad_norm": 0.18566212058067322,
658
+ "learning_rate": 1.8951436309804765e-05,
659
+ "loss": 0.0049,
660
+ "step": 460
661
+ },
662
+ {
663
+ "epoch": 0.75,
664
+ "grad_norm": 0.2376997023820877,
665
+ "learning_rate": 1.7860619515673033e-05,
666
+ "loss": 0.0071,
667
+ "step": 465
668
+ },
669
+ {
670
+ "epoch": 0.7580645161290323,
671
+ "grad_norm": 0.24823206663131714,
672
+ "learning_rate": 1.6795269861638042e-05,
673
+ "loss": 0.0068,
674
+ "step": 470
675
+ },
676
+ {
677
+ "epoch": 0.7661290322580645,
678
+ "grad_norm": 0.15533004701137543,
679
+ "learning_rate": 1.575623152728518e-05,
680
+ "loss": 0.0054,
681
+ "step": 475
682
+ },
683
+ {
684
+ "epoch": 0.7741935483870968,
685
+ "grad_norm": 0.1936763972043991,
686
+ "learning_rate": 1.4744327843196043e-05,
687
+ "loss": 0.0053,
688
+ "step": 480
689
+ },
690
+ {
691
+ "epoch": 0.782258064516129,
692
+ "grad_norm": 0.24305252730846405,
693
+ "learning_rate": 1.3760360638544012e-05,
694
+ "loss": 0.0042,
695
+ "step": 485
696
+ },
697
+ {
698
+ "epoch": 0.7903225806451613,
699
+ "grad_norm": 0.23395785689353943,
700
+ "learning_rate": 1.280510960572745e-05,
701
+ "loss": 0.006,
702
+ "step": 490
703
+ },
704
+ {
705
+ "epoch": 0.7983870967741935,
706
+ "grad_norm": 0.160566508769989,
707
+ "learning_rate": 1.187933168254397e-05,
708
+ "loss": 0.0036,
709
+ "step": 495
710
+ },
711
+ {
712
+ "epoch": 0.8064516129032258,
713
+ "grad_norm": 0.173016756772995,
714
+ "learning_rate": 1.0983760452395413e-05,
715
+ "loss": 0.004,
716
+ "step": 500
717
+ },
718
+ {
719
+ "epoch": 0.8145161290322581,
720
+ "grad_norm": 0.4186556935310364,
721
+ "learning_rate": 1.0119105562998698e-05,
722
+ "loss": 0.0062,
723
+ "step": 505
724
+ },
725
+ {
726
+ "epoch": 0.8225806451612904,
727
+ "grad_norm": 0.29756516218185425,
728
+ "learning_rate": 9.28605216406337e-06,
729
+ "loss": 0.0046,
730
+ "step": 510
731
+ },
732
+ {
733
+ "epoch": 0.8306451612903226,
734
+ "grad_norm": 0.1282072514295578,
735
+ "learning_rate": 8.485260364381187e-06,
736
+ "loss": 0.0033,
737
+ "step": 515
738
+ },
739
+ {
740
+ "epoch": 0.8387096774193549,
741
+ "grad_norm": 0.1337880641222,
742
+ "learning_rate": 7.717364708758024e-06,
743
+ "loss": 0.0049,
744
+ "step": 520
745
+ },
746
+ {
747
+ "epoch": 0.8467741935483871,
748
+ "grad_norm": 0.2575572431087494,
749
+ "learning_rate": 6.982973675202676e-06,
750
+ "loss": 0.0058,
751
+ "step": 525
752
+ },
753
+ {
754
+ "epoch": 0.8548387096774194,
755
+ "grad_norm": 0.1366008073091507,
756
+ "learning_rate": 6.282669192770896e-06,
757
+ "loss": 0.005,
758
+ "step": 530
759
+ },
760
+ {
761
+ "epoch": 0.8629032258064516,
762
+ "grad_norm": 2.0787034034729004,
763
+ "learning_rate": 5.617006180446688e-06,
764
+ "loss": 0.0075,
765
+ "step": 535
766
+ },
767
+ {
768
+ "epoch": 0.8709677419354839,
769
+ "grad_norm": 0.17773574590682983,
770
+ "learning_rate": 4.986512107426283e-06,
771
+ "loss": 0.0054,
772
+ "step": 540
773
+ },
774
+ {
775
+ "epoch": 0.8790322580645161,
776
+ "grad_norm": 0.0983528196811676,
777
+ "learning_rate": 4.391686575153331e-06,
778
+ "loss": 0.0043,
779
+ "step": 545
780
+ },
781
+ {
782
+ "epoch": 0.8870967741935484,
783
+ "grad_norm": 0.13671350479125977,
784
+ "learning_rate": 3.833000921436319e-06,
785
+ "loss": 0.0031,
786
+ "step": 550
787
+ },
788
+ {
789
+ "epoch": 0.8951612903225806,
790
+ "grad_norm": 0.6432064175605774,
791
+ "learning_rate": 3.310897846962041e-06,
792
+ "loss": 0.0048,
793
+ "step": 555
794
+ },
795
+ {
796
+ "epoch": 0.9032258064516129,
797
+ "grad_norm": 0.5728097558021545,
798
+ "learning_rate": 2.825791064500993e-06,
799
+ "loss": 0.0053,
800
+ "step": 560
801
+ },
802
+ {
803
+ "epoch": 0.9112903225806451,
804
+ "grad_norm": 0.6017826199531555,
805
+ "learning_rate": 2.3780649710827554e-06,
806
+ "loss": 0.0058,
807
+ "step": 565
808
+ },
809
+ {
810
+ "epoch": 0.9193548387096774,
811
+ "grad_norm": 0.3495374321937561,
812
+ "learning_rate": 1.9680743434010384e-06,
813
+ "loss": 0.007,
814
+ "step": 570
815
+ },
816
+ {
817
+ "epoch": 0.9274193548387096,
818
+ "grad_norm": 0.34849557280540466,
819
+ "learning_rate": 1.5961440566897912e-06,
820
+ "loss": 0.003,
821
+ "step": 575
822
+ },
823
+ {
824
+ "epoch": 0.9354838709677419,
825
+ "grad_norm": 0.24747046828269958,
826
+ "learning_rate": 1.2625688272930926e-06,
827
+ "loss": 0.0042,
828
+ "step": 580
829
+ },
830
+ {
831
+ "epoch": 0.9435483870967742,
832
+ "grad_norm": 0.5230942964553833,
833
+ "learning_rate": 9.67612979132948e-07,
834
+ "loss": 0.0045,
835
+ "step": 585
836
+ },
837
+ {
838
+ "epoch": 0.9516129032258065,
839
+ "grad_norm": 0.14619432389736176,
840
+ "learning_rate": 7.115102342598101e-07,
841
+ "loss": 0.0034,
842
+ "step": 590
843
+ },
844
+ {
845
+ "epoch": 0.9596774193548387,
846
+ "grad_norm": 0.47683650255203247,
847
+ "learning_rate": 4.944635276520393e-07,
848
+ "loss": 0.0036,
849
+ "step": 595
850
+ },
851
+ {
852
+ "epoch": 0.967741935483871,
853
+ "grad_norm": 0.20121371746063232,
854
+ "learning_rate": 3.1664484641086287e-07,
855
+ "loss": 0.0054,
856
+ "step": 600
857
+ },
858
+ {
859
+ "epoch": 0.9758064516129032,
860
+ "grad_norm": 0.12017033249139786,
861
+ "learning_rate": 1.781950934783505e-07,
862
+ "loss": 0.0059,
863
+ "step": 605
864
+ },
865
+ {
866
+ "epoch": 0.9838709677419355,
867
+ "grad_norm": 0.39845508337020874,
868
+ "learning_rate": 7.92239759864255e-08,
869
+ "loss": 0.0043,
870
+ "step": 610
871
+ },
872
+ {
873
+ "epoch": 0.9919354838709677,
874
+ "grad_norm": 0.21712489426136017,
875
+ "learning_rate": 1.980991832524759e-08,
876
+ "loss": 0.0056,
877
+ "step": 615
878
+ },
879
+ {
880
+ "epoch": 1.0,
881
+ "grad_norm": 0.13640886545181274,
882
+ "learning_rate": 0.0,
883
+ "loss": 0.0046,
884
+ "step": 620
885
+ },
886
+ {
887
+ "epoch": 1.0,
888
+ "eval_loss": 0.0026028414722532034,
889
+ "eval_runtime": 3.8766,
890
+ "eval_samples_per_second": 1.548,
891
+ "eval_steps_per_second": 1.548,
892
+ "step": 620
893
+ },
894
+ {
895
+ "epoch": 1.0,
896
+ "step": 620,
897
+ "total_flos": 5.590502947474637e+16,
898
+ "train_loss": 0.06471517855121244,
899
+ "train_runtime": 1309.8316,
900
+ "train_samples_per_second": 0.473,
901
+ "train_steps_per_second": 0.473
902
+ }
903
+ ],
904
+ "logging_steps": 5,
905
+ "max_steps": 620,
906
+ "num_input_tokens_seen": 0,
907
+ "num_train_epochs": 1,
908
+ "save_steps": 100,
909
+ "stateful_callbacks": {
910
+ "TrainerControl": {
911
+ "args": {
912
+ "should_epoch_stop": false,
913
+ "should_evaluate": false,
914
+ "should_log": false,
915
+ "should_save": true,
916
+ "should_training_stop": true
917
+ },
918
+ "attributes": {}
919
+ }
920
+ },
921
+ "total_flos": 5.590502947474637e+16,
922
+ "train_batch_size": 1,
923
+ "trial_name": null,
924
+ "trial_params": null
925
+ }