weizhepei commited on
Commit
ae0c54b
·
verified ·
1 Parent(s): 390ba9d

Model save

Browse files
Files changed (5) hide show
  1. README.md +58 -0
  2. all_results.json +8 -0
  3. generation_config.json +14 -0
  4. train_results.json +8 -0
  5. trainer_state.json +1078 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-3B-Instruct
3
+ library_name: transformers
4
+ model_name: Qwen2.5-3B-WebArena-Lite-SFT-epoch-10
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for Qwen2.5-3B-WebArena-Lite-SFT-epoch-10
13
+
14
+ This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct).
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="weizhepei/Qwen2.5-3B-WebArena-Lite-SFT-epoch-10", 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/uva-llm/huggingface/runs/11rxohqk)
31
+
32
+
33
+ This model was trained with SFT.
34
+
35
+ ### Framework versions
36
+
37
+ - TRL: 0.16.0.dev0
38
+ - Transformers: 4.49.0
39
+ - Pytorch: 2.5.1
40
+ - Datasets: 3.4.1
41
+ - Tokenizers: 0.21.1
42
+
43
+ ## Citations
44
+
45
+
46
+
47
+ Cite TRL as:
48
+
49
+ ```bibtex
50
+ @misc{vonwerra2022trl,
51
+ title = {{TRL: Transformer Reinforcement Learning}},
52
+ 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},
53
+ year = 2020,
54
+ journal = {GitHub repository},
55
+ publisher = {GitHub},
56
+ howpublished = {\url{https://github.com/huggingface/trl}}
57
+ }
58
+ ```
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 1253741243269120.0,
3
+ "train_loss": 0.05100687126286101,
4
+ "train_runtime": 15432.4671,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 6.13,
7
+ "train_steps_per_second": 0.048
8
+ }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.49.0"
14
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 1253741243269120.0,
3
+ "train_loss": 0.05100687126286101,
4
+ "train_runtime": 15432.4671,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 6.13,
7
+ "train_steps_per_second": 0.048
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1078 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 10.0,
5
+ "eval_steps": 500,
6
+ "global_step": 740,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.06756756756756757,
13
+ "grad_norm": 6.5182561675653,
14
+ "learning_rate": 6.7567567567567575e-06,
15
+ "loss": 0.6032,
16
+ "step": 5
17
+ },
18
+ {
19
+ "epoch": 0.13513513513513514,
20
+ "grad_norm": 2.467090492652078,
21
+ "learning_rate": 1.3513513513513515e-05,
22
+ "loss": 0.4268,
23
+ "step": 10
24
+ },
25
+ {
26
+ "epoch": 0.20270270270270271,
27
+ "grad_norm": 2.0035666433223756,
28
+ "learning_rate": 2.0270270270270273e-05,
29
+ "loss": 0.327,
30
+ "step": 15
31
+ },
32
+ {
33
+ "epoch": 0.2702702702702703,
34
+ "grad_norm": 1.5202456605148946,
35
+ "learning_rate": 2.702702702702703e-05,
36
+ "loss": 0.2619,
37
+ "step": 20
38
+ },
39
+ {
40
+ "epoch": 0.33783783783783783,
41
+ "grad_norm": 1.7543743381326309,
42
+ "learning_rate": 3.3783783783783784e-05,
43
+ "loss": 0.21,
44
+ "step": 25
45
+ },
46
+ {
47
+ "epoch": 0.40540540540540543,
48
+ "grad_norm": 3.413879980820011,
49
+ "learning_rate": 4.0540540540540545e-05,
50
+ "loss": 0.2167,
51
+ "step": 30
52
+ },
53
+ {
54
+ "epoch": 0.47297297297297297,
55
+ "grad_norm": 1.1607159273353023,
56
+ "learning_rate": 4.72972972972973e-05,
57
+ "loss": 0.1978,
58
+ "step": 35
59
+ },
60
+ {
61
+ "epoch": 0.5405405405405406,
62
+ "grad_norm": 1.0076230382756923,
63
+ "learning_rate": 4.9997978016429605e-05,
64
+ "loss": 0.1987,
65
+ "step": 40
66
+ },
67
+ {
68
+ "epoch": 0.6081081081081081,
69
+ "grad_norm": 1.162970602606406,
70
+ "learning_rate": 4.9985622766211935e-05,
71
+ "loss": 0.1928,
72
+ "step": 45
73
+ },
74
+ {
75
+ "epoch": 0.6756756756756757,
76
+ "grad_norm": 0.9302783027187795,
77
+ "learning_rate": 4.996204175076325e-05,
78
+ "loss": 0.179,
79
+ "step": 50
80
+ },
81
+ {
82
+ "epoch": 0.7432432432432432,
83
+ "grad_norm": 0.9702952191590688,
84
+ "learning_rate": 4.99272467427147e-05,
85
+ "loss": 0.1888,
86
+ "step": 55
87
+ },
88
+ {
89
+ "epoch": 0.8108108108108109,
90
+ "grad_norm": 0.9759423611984822,
91
+ "learning_rate": 4.9881255113192526e-05,
92
+ "loss": 0.1929,
93
+ "step": 60
94
+ },
95
+ {
96
+ "epoch": 0.8783783783783784,
97
+ "grad_norm": 0.880390195886552,
98
+ "learning_rate": 4.982408982314565e-05,
99
+ "loss": 0.1896,
100
+ "step": 65
101
+ },
102
+ {
103
+ "epoch": 0.9459459459459459,
104
+ "grad_norm": 0.8204121532507659,
105
+ "learning_rate": 4.975577941188258e-05,
106
+ "loss": 0.1713,
107
+ "step": 70
108
+ },
109
+ {
110
+ "epoch": 1.0135135135135136,
111
+ "grad_norm": 0.5777465635036437,
112
+ "learning_rate": 4.967635798282344e-05,
113
+ "loss": 0.162,
114
+ "step": 75
115
+ },
116
+ {
117
+ "epoch": 1.0810810810810811,
118
+ "grad_norm": 0.6949612712064188,
119
+ "learning_rate": 4.958586518647407e-05,
120
+ "loss": 0.1094,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 1.1486486486486487,
125
+ "grad_norm": 0.7652099300477704,
126
+ "learning_rate": 4.9484346200630855e-05,
127
+ "loss": 0.1215,
128
+ "step": 85
129
+ },
130
+ {
131
+ "epoch": 1.2162162162162162,
132
+ "grad_norm": 0.723315962513987,
133
+ "learning_rate": 4.937185170782607e-05,
134
+ "loss": 0.1019,
135
+ "step": 90
136
+ },
137
+ {
138
+ "epoch": 1.2837837837837838,
139
+ "grad_norm": 0.8479152677405634,
140
+ "learning_rate": 4.9248437870025035e-05,
141
+ "loss": 0.1069,
142
+ "step": 95
143
+ },
144
+ {
145
+ "epoch": 1.3513513513513513,
146
+ "grad_norm": 0.6337675422347736,
147
+ "learning_rate": 4.911416630058772e-05,
148
+ "loss": 0.1077,
149
+ "step": 100
150
+ },
151
+ {
152
+ "epoch": 1.4189189189189189,
153
+ "grad_norm": 0.7402025515878772,
154
+ "learning_rate": 4.896910403350873e-05,
155
+ "loss": 0.1046,
156
+ "step": 105
157
+ },
158
+ {
159
+ "epoch": 1.4864864864864864,
160
+ "grad_norm": 1.1176372786369246,
161
+ "learning_rate": 4.88133234899512e-05,
162
+ "loss": 0.0993,
163
+ "step": 110
164
+ },
165
+ {
166
+ "epoch": 1.554054054054054,
167
+ "grad_norm": 7.67563476349744,
168
+ "learning_rate": 4.864690244209105e-05,
169
+ "loss": 0.1096,
170
+ "step": 115
171
+ },
172
+ {
173
+ "epoch": 1.6216216216216215,
174
+ "grad_norm": 2.0857374248215415,
175
+ "learning_rate": 4.8469923974289874e-05,
176
+ "loss": 0.1025,
177
+ "step": 120
178
+ },
179
+ {
180
+ "epoch": 1.689189189189189,
181
+ "grad_norm": 0.5388629684725229,
182
+ "learning_rate": 4.828247644161577e-05,
183
+ "loss": 0.1021,
184
+ "step": 125
185
+ },
186
+ {
187
+ "epoch": 1.7567567567567568,
188
+ "grad_norm": 0.6274012190681492,
189
+ "learning_rate": 4.808465342573274e-05,
190
+ "loss": 0.1089,
191
+ "step": 130
192
+ },
193
+ {
194
+ "epoch": 1.8243243243243243,
195
+ "grad_norm": 0.5671573351859114,
196
+ "learning_rate": 4.787655368818087e-05,
197
+ "loss": 0.094,
198
+ "step": 135
199
+ },
200
+ {
201
+ "epoch": 1.8918918918918919,
202
+ "grad_norm": 0.5368848884663832,
203
+ "learning_rate": 4.765828112107034e-05,
204
+ "loss": 0.0897,
205
+ "step": 140
206
+ },
207
+ {
208
+ "epoch": 1.9594594594594594,
209
+ "grad_norm": 0.6463078769066465,
210
+ "learning_rate": 4.742994469521421e-05,
211
+ "loss": 0.0976,
212
+ "step": 145
213
+ },
214
+ {
215
+ "epoch": 2.027027027027027,
216
+ "grad_norm": 0.5546106089284485,
217
+ "learning_rate": 4.719165840572557e-05,
218
+ "loss": 0.0799,
219
+ "step": 150
220
+ },
221
+ {
222
+ "epoch": 2.0945945945945947,
223
+ "grad_norm": 0.5556979492761135,
224
+ "learning_rate": 4.694354121510644e-05,
225
+ "loss": 0.0692,
226
+ "step": 155
227
+ },
228
+ {
229
+ "epoch": 2.1621621621621623,
230
+ "grad_norm": 0.5153408393477845,
231
+ "learning_rate": 4.668571699385668e-05,
232
+ "loss": 0.0643,
233
+ "step": 160
234
+ },
235
+ {
236
+ "epoch": 2.22972972972973,
237
+ "grad_norm": 0.44312390146493885,
238
+ "learning_rate": 4.641831445863265e-05,
239
+ "loss": 0.056,
240
+ "step": 165
241
+ },
242
+ {
243
+ "epoch": 2.2972972972972974,
244
+ "grad_norm": 0.853737761210155,
245
+ "learning_rate": 4.614146710798645e-05,
246
+ "loss": 0.0647,
247
+ "step": 170
248
+ },
249
+ {
250
+ "epoch": 2.364864864864865,
251
+ "grad_norm": 0.41643642657814545,
252
+ "learning_rate": 4.585531315571788e-05,
253
+ "loss": 0.0609,
254
+ "step": 175
255
+ },
256
+ {
257
+ "epoch": 2.4324324324324325,
258
+ "grad_norm": 0.5004584107448025,
259
+ "learning_rate": 4.555999546187229e-05,
260
+ "loss": 0.0623,
261
+ "step": 180
262
+ },
263
+ {
264
+ "epoch": 2.5,
265
+ "grad_norm": 0.5484821364819039,
266
+ "learning_rate": 4.5255661461418854e-05,
267
+ "loss": 0.0607,
268
+ "step": 185
269
+ },
270
+ {
271
+ "epoch": 2.5675675675675675,
272
+ "grad_norm": 0.740378611251539,
273
+ "learning_rate": 4.4942463090644896e-05,
274
+ "loss": 0.0631,
275
+ "step": 190
276
+ },
277
+ {
278
+ "epoch": 2.635135135135135,
279
+ "grad_norm": 0.6794703674337778,
280
+ "learning_rate": 4.462055671130289e-05,
281
+ "loss": 0.0634,
282
+ "step": 195
283
+ },
284
+ {
285
+ "epoch": 2.7027027027027026,
286
+ "grad_norm": 0.4944963404173982,
287
+ "learning_rate": 4.4290103032548094e-05,
288
+ "loss": 0.0575,
289
+ "step": 200
290
+ },
291
+ {
292
+ "epoch": 2.77027027027027,
293
+ "grad_norm": 0.5398626146014253,
294
+ "learning_rate": 4.395126703070589e-05,
295
+ "loss": 0.0659,
296
+ "step": 205
297
+ },
298
+ {
299
+ "epoch": 2.8378378378378377,
300
+ "grad_norm": 0.5017193338659323,
301
+ "learning_rate": 4.360421786690862e-05,
302
+ "loss": 0.0615,
303
+ "step": 210
304
+ },
305
+ {
306
+ "epoch": 2.9054054054054053,
307
+ "grad_norm": 0.5177174186838486,
308
+ "learning_rate": 4.324912880264326e-05,
309
+ "loss": 0.0666,
310
+ "step": 215
311
+ },
312
+ {
313
+ "epoch": 2.972972972972973,
314
+ "grad_norm": 0.39586341268090364,
315
+ "learning_rate": 4.288617711325207e-05,
316
+ "loss": 0.0571,
317
+ "step": 220
318
+ },
319
+ {
320
+ "epoch": 3.0405405405405403,
321
+ "grad_norm": 0.3625971418475026,
322
+ "learning_rate": 4.251554399942928e-05,
323
+ "loss": 0.0476,
324
+ "step": 225
325
+ },
326
+ {
327
+ "epoch": 3.108108108108108,
328
+ "grad_norm": 0.45196738729892844,
329
+ "learning_rate": 4.21374144967581e-05,
330
+ "loss": 0.0365,
331
+ "step": 230
332
+ },
333
+ {
334
+ "epoch": 3.175675675675676,
335
+ "grad_norm": 0.4119519496515403,
336
+ "learning_rate": 4.1751977383333224e-05,
337
+ "loss": 0.0418,
338
+ "step": 235
339
+ },
340
+ {
341
+ "epoch": 3.2432432432432434,
342
+ "grad_norm": 0.39819242944877964,
343
+ "learning_rate": 4.1359425085514906e-05,
344
+ "loss": 0.0366,
345
+ "step": 240
346
+ },
347
+ {
348
+ "epoch": 3.310810810810811,
349
+ "grad_norm": 0.424750779792785,
350
+ "learning_rate": 4.095995358186162e-05,
351
+ "loss": 0.0416,
352
+ "step": 245
353
+ },
354
+ {
355
+ "epoch": 3.3783783783783785,
356
+ "grad_norm": 0.35953093850347395,
357
+ "learning_rate": 4.055376230528936e-05,
358
+ "loss": 0.0382,
359
+ "step": 250
360
+ },
361
+ {
362
+ "epoch": 3.445945945945946,
363
+ "grad_norm": 0.43321981154907985,
364
+ "learning_rate": 4.0141054043506406e-05,
365
+ "loss": 0.0437,
366
+ "step": 255
367
+ },
368
+ {
369
+ "epoch": 3.5135135135135136,
370
+ "grad_norm": 0.4421355103321677,
371
+ "learning_rate": 3.972203483777315e-05,
372
+ "loss": 0.0388,
373
+ "step": 260
374
+ },
375
+ {
376
+ "epoch": 3.581081081081081,
377
+ "grad_norm": 0.4158553312353602,
378
+ "learning_rate": 3.929691388003772e-05,
379
+ "loss": 0.0444,
380
+ "step": 265
381
+ },
382
+ {
383
+ "epoch": 3.6486486486486487,
384
+ "grad_norm": 0.4057889642816714,
385
+ "learning_rate": 3.886590340849852e-05,
386
+ "loss": 0.0451,
387
+ "step": 270
388
+ },
389
+ {
390
+ "epoch": 3.7162162162162162,
391
+ "grad_norm": 0.43411808520930756,
392
+ "learning_rate": 3.842921860164607e-05,
393
+ "loss": 0.0446,
394
+ "step": 275
395
+ },
396
+ {
397
+ "epoch": 3.7837837837837838,
398
+ "grad_norm": 0.47135409811439094,
399
+ "learning_rate": 3.798707747083694e-05,
400
+ "loss": 0.0437,
401
+ "step": 280
402
+ },
403
+ {
404
+ "epoch": 3.8513513513513513,
405
+ "grad_norm": 0.3634190854519526,
406
+ "learning_rate": 3.753970075145322e-05,
407
+ "loss": 0.0383,
408
+ "step": 285
409
+ },
410
+ {
411
+ "epoch": 3.918918918918919,
412
+ "grad_norm": 0.4113954814959505,
413
+ "learning_rate": 3.7087311792702265e-05,
414
+ "loss": 0.0437,
415
+ "step": 290
416
+ },
417
+ {
418
+ "epoch": 3.9864864864864864,
419
+ "grad_norm": 0.5620689578203624,
420
+ "learning_rate": 3.663013644611139e-05,
421
+ "loss": 0.044,
422
+ "step": 295
423
+ },
424
+ {
425
+ "epoch": 4.054054054054054,
426
+ "grad_norm": 0.43501925674622166,
427
+ "learning_rate": 3.616840295277328e-05,
428
+ "loss": 0.029,
429
+ "step": 300
430
+ },
431
+ {
432
+ "epoch": 4.121621621621622,
433
+ "grad_norm": 0.40992682304298633,
434
+ "learning_rate": 3.5702341829398525e-05,
435
+ "loss": 0.0241,
436
+ "step": 305
437
+ },
438
+ {
439
+ "epoch": 4.1891891891891895,
440
+ "grad_norm": 0.36148624086580156,
441
+ "learning_rate": 3.523218575323198e-05,
442
+ "loss": 0.026,
443
+ "step": 310
444
+ },
445
+ {
446
+ "epoch": 4.256756756756757,
447
+ "grad_norm": 0.2792981727538657,
448
+ "learning_rate": 3.475816944589058e-05,
449
+ "loss": 0.0277,
450
+ "step": 315
451
+ },
452
+ {
453
+ "epoch": 4.324324324324325,
454
+ "grad_norm": 0.6042932142708992,
455
+ "learning_rate": 3.4280529556180404e-05,
456
+ "loss": 0.0269,
457
+ "step": 320
458
+ },
459
+ {
460
+ "epoch": 4.391891891891892,
461
+ "grad_norm": 0.45641619996504024,
462
+ "learning_rate": 3.379950454195172e-05,
463
+ "loss": 0.0244,
464
+ "step": 325
465
+ },
466
+ {
467
+ "epoch": 4.45945945945946,
468
+ "grad_norm": 0.3319573209641707,
469
+ "learning_rate": 3.331533455105084e-05,
470
+ "loss": 0.0271,
471
+ "step": 330
472
+ },
473
+ {
474
+ "epoch": 4.527027027027027,
475
+ "grad_norm": 0.303898417134011,
476
+ "learning_rate": 3.2828261301428206e-05,
477
+ "loss": 0.0252,
478
+ "step": 335
479
+ },
480
+ {
481
+ "epoch": 4.594594594594595,
482
+ "grad_norm": 0.32553696908478624,
483
+ "learning_rate": 3.23385279604627e-05,
484
+ "loss": 0.0251,
485
+ "step": 340
486
+ },
487
+ {
488
+ "epoch": 4.662162162162162,
489
+ "grad_norm": 0.2550335003989632,
490
+ "learning_rate": 3.18463790235623e-05,
491
+ "loss": 0.0255,
492
+ "step": 345
493
+ },
494
+ {
495
+ "epoch": 4.72972972972973,
496
+ "grad_norm": 0.40503372366040374,
497
+ "learning_rate": 3.135206019210167e-05,
498
+ "loss": 0.0268,
499
+ "step": 350
500
+ },
501
+ {
502
+ "epoch": 4.797297297297297,
503
+ "grad_norm": 0.5104789962705262,
504
+ "learning_rate": 3.085581825075782e-05,
505
+ "loss": 0.0287,
506
+ "step": 355
507
+ },
508
+ {
509
+ "epoch": 4.864864864864865,
510
+ "grad_norm": 0.347089797482406,
511
+ "learning_rate": 3.0357900944304774e-05,
512
+ "loss": 0.0274,
513
+ "step": 360
514
+ },
515
+ {
516
+ "epoch": 4.9324324324324325,
517
+ "grad_norm": 0.46095293238337387,
518
+ "learning_rate": 2.9858556853929048e-05,
519
+ "loss": 0.0276,
520
+ "step": 365
521
+ },
522
+ {
523
+ "epoch": 5.0,
524
+ "grad_norm": 0.30461903604327145,
525
+ "learning_rate": 2.9358035273127483e-05,
526
+ "loss": 0.0258,
527
+ "step": 370
528
+ },
529
+ {
530
+ "epoch": 5.0675675675675675,
531
+ "grad_norm": 0.2538348451100868,
532
+ "learning_rate": 2.8856586083249487e-05,
533
+ "loss": 0.0151,
534
+ "step": 375
535
+ },
536
+ {
537
+ "epoch": 5.135135135135135,
538
+ "grad_norm": 0.23696050364844123,
539
+ "learning_rate": 2.83544596287458e-05,
540
+ "loss": 0.0145,
541
+ "step": 380
542
+ },
543
+ {
544
+ "epoch": 5.202702702702703,
545
+ "grad_norm": 0.317781039944712,
546
+ "learning_rate": 2.785190659218604e-05,
547
+ "loss": 0.0141,
548
+ "step": 385
549
+ },
550
+ {
551
+ "epoch": 5.27027027027027,
552
+ "grad_norm": 0.285237643213249,
553
+ "learning_rate": 2.7349177869107462e-05,
554
+ "loss": 0.0151,
555
+ "step": 390
556
+ },
557
+ {
558
+ "epoch": 5.337837837837838,
559
+ "grad_norm": 0.29901748039669407,
560
+ "learning_rate": 2.684652444275741e-05,
561
+ "loss": 0.016,
562
+ "step": 395
563
+ },
564
+ {
565
+ "epoch": 5.405405405405405,
566
+ "grad_norm": 0.4048747946319297,
567
+ "learning_rate": 2.634419725879193e-05,
568
+ "loss": 0.0141,
569
+ "step": 400
570
+ },
571
+ {
572
+ "epoch": 5.472972972972973,
573
+ "grad_norm": 0.362450553880386,
574
+ "learning_rate": 2.58424470999932e-05,
575
+ "loss": 0.0162,
576
+ "step": 405
577
+ },
578
+ {
579
+ "epoch": 5.54054054054054,
580
+ "grad_norm": 0.25875308146518555,
581
+ "learning_rate": 2.534152446106825e-05,
582
+ "loss": 0.017,
583
+ "step": 410
584
+ },
585
+ {
586
+ "epoch": 5.608108108108108,
587
+ "grad_norm": 0.547918758419167,
588
+ "learning_rate": 2.4841679423591523e-05,
589
+ "loss": 0.0147,
590
+ "step": 415
591
+ },
592
+ {
593
+ "epoch": 5.675675675675675,
594
+ "grad_norm": 0.38452896016200555,
595
+ "learning_rate": 2.4343161531153647e-05,
596
+ "loss": 0.0156,
597
+ "step": 420
598
+ },
599
+ {
600
+ "epoch": 5.743243243243243,
601
+ "grad_norm": 0.29252334164098,
602
+ "learning_rate": 2.3846219664778824e-05,
603
+ "loss": 0.015,
604
+ "step": 425
605
+ },
606
+ {
607
+ "epoch": 5.8108108108108105,
608
+ "grad_norm": 0.2524439030025814,
609
+ "learning_rate": 2.3351101918672985e-05,
610
+ "loss": 0.0167,
611
+ "step": 430
612
+ },
613
+ {
614
+ "epoch": 5.878378378378378,
615
+ "grad_norm": 0.34963079065242764,
616
+ "learning_rate": 2.2858055476364822e-05,
617
+ "loss": 0.0165,
618
+ "step": 435
619
+ },
620
+ {
621
+ "epoch": 5.945945945945946,
622
+ "grad_norm": 0.25040336257702267,
623
+ "learning_rate": 2.2367326487301317e-05,
624
+ "loss": 0.014,
625
+ "step": 440
626
+ },
627
+ {
628
+ "epoch": 6.013513513513513,
629
+ "grad_norm": 0.3479947117611859,
630
+ "learning_rate": 2.1879159943959686e-05,
631
+ "loss": 0.0141,
632
+ "step": 445
633
+ },
634
+ {
635
+ "epoch": 6.081081081081081,
636
+ "grad_norm": 0.20659430889580188,
637
+ "learning_rate": 2.139379955953686e-05,
638
+ "loss": 0.0071,
639
+ "step": 450
640
+ },
641
+ {
642
+ "epoch": 6.148648648648648,
643
+ "grad_norm": 0.220203705014732,
644
+ "learning_rate": 2.0911487646277623e-05,
645
+ "loss": 0.0076,
646
+ "step": 455
647
+ },
648
+ {
649
+ "epoch": 6.216216216216216,
650
+ "grad_norm": 0.17947459431869756,
651
+ "learning_rate": 2.0432464994502203e-05,
652
+ "loss": 0.0069,
653
+ "step": 460
654
+ },
655
+ {
656
+ "epoch": 6.283783783783784,
657
+ "grad_norm": 0.21636057374547216,
658
+ "learning_rate": 1.995697075239365e-05,
659
+ "loss": 0.0076,
660
+ "step": 465
661
+ },
662
+ {
663
+ "epoch": 6.351351351351352,
664
+ "grad_norm": 0.29024707610053163,
665
+ "learning_rate": 1.9485242306605028e-05,
666
+ "loss": 0.0071,
667
+ "step": 470
668
+ },
669
+ {
670
+ "epoch": 6.418918918918919,
671
+ "grad_norm": 0.25019577078633803,
672
+ "learning_rate": 1.9017515163746058e-05,
673
+ "loss": 0.006,
674
+ "step": 475
675
+ },
676
+ {
677
+ "epoch": 6.486486486486487,
678
+ "grad_norm": 0.1844009452055872,
679
+ "learning_rate": 1.855402283280836e-05,
680
+ "loss": 0.0063,
681
+ "step": 480
682
+ },
683
+ {
684
+ "epoch": 6.554054054054054,
685
+ "grad_norm": 0.30319539004527984,
686
+ "learning_rate": 1.8094996708587958e-05,
687
+ "loss": 0.0081,
688
+ "step": 485
689
+ },
690
+ {
691
+ "epoch": 6.621621621621622,
692
+ "grad_norm": 0.2036348164101663,
693
+ "learning_rate": 1.7640665956163306e-05,
694
+ "loss": 0.0084,
695
+ "step": 490
696
+ },
697
+ {
698
+ "epoch": 6.6891891891891895,
699
+ "grad_norm": 0.3106159021115643,
700
+ "learning_rate": 1.719125739648648e-05,
701
+ "loss": 0.008,
702
+ "step": 495
703
+ },
704
+ {
705
+ "epoch": 6.756756756756757,
706
+ "grad_norm": 0.21194797013825406,
707
+ "learning_rate": 1.6746995393144668e-05,
708
+ "loss": 0.0071,
709
+ "step": 500
710
+ },
711
+ {
712
+ "epoch": 6.824324324324325,
713
+ "grad_norm": 0.22026182202638217,
714
+ "learning_rate": 1.6308101740348433e-05,
715
+ "loss": 0.0066,
716
+ "step": 505
717
+ },
718
+ {
719
+ "epoch": 6.891891891891892,
720
+ "grad_norm": 0.2816260648539415,
721
+ "learning_rate": 1.5874795552202773e-05,
722
+ "loss": 0.0066,
723
+ "step": 510
724
+ },
725
+ {
726
+ "epoch": 6.95945945945946,
727
+ "grad_norm": 0.22182276522514063,
728
+ "learning_rate": 1.5447293153316163e-05,
729
+ "loss": 0.0067,
730
+ "step": 515
731
+ },
732
+ {
733
+ "epoch": 7.027027027027027,
734
+ "grad_norm": 0.11067954194940728,
735
+ "learning_rate": 1.5025807970802252e-05,
736
+ "loss": 0.0046,
737
+ "step": 520
738
+ },
739
+ {
740
+ "epoch": 7.094594594594595,
741
+ "grad_norm": 0.10657051658609974,
742
+ "learning_rate": 1.4610550427728103e-05,
743
+ "loss": 0.0031,
744
+ "step": 525
745
+ },
746
+ {
747
+ "epoch": 7.162162162162162,
748
+ "grad_norm": 0.19612699613612214,
749
+ "learning_rate": 1.4201727838062181e-05,
750
+ "loss": 0.0028,
751
+ "step": 530
752
+ },
753
+ {
754
+ "epoch": 7.22972972972973,
755
+ "grad_norm": 0.14244125673198543,
756
+ "learning_rate": 1.3799544303174514e-05,
757
+ "loss": 0.0031,
758
+ "step": 535
759
+ },
760
+ {
761
+ "epoch": 7.297297297297297,
762
+ "grad_norm": 0.39421131084217764,
763
+ "learning_rate": 1.3404200609940754e-05,
764
+ "loss": 0.0024,
765
+ "step": 540
766
+ },
767
+ {
768
+ "epoch": 7.364864864864865,
769
+ "grad_norm": 0.3350757057485832,
770
+ "learning_rate": 1.3015894130500977e-05,
771
+ "loss": 0.0027,
772
+ "step": 545
773
+ },
774
+ {
775
+ "epoch": 7.4324324324324325,
776
+ "grad_norm": 0.22416604127845918,
777
+ "learning_rate": 1.2634818723723174e-05,
778
+ "loss": 0.0038,
779
+ "step": 550
780
+ },
781
+ {
782
+ "epoch": 7.5,
783
+ "grad_norm": 0.09935074986475295,
784
+ "learning_rate": 1.2261164638420832e-05,
785
+ "loss": 0.0021,
786
+ "step": 555
787
+ },
788
+ {
789
+ "epoch": 7.5675675675675675,
790
+ "grad_norm": 0.2604369794575155,
791
+ "learning_rate": 1.1895118418372734e-05,
792
+ "loss": 0.0035,
793
+ "step": 560
794
+ },
795
+ {
796
+ "epoch": 7.635135135135135,
797
+ "grad_norm": 0.1806357452528696,
798
+ "learning_rate": 1.1536862809192518e-05,
799
+ "loss": 0.0036,
800
+ "step": 565
801
+ },
802
+ {
803
+ "epoch": 7.702702702702703,
804
+ "grad_norm": 0.17567109094612535,
805
+ "learning_rate": 1.1186576667094342e-05,
806
+ "loss": 0.0027,
807
+ "step": 570
808
+ },
809
+ {
810
+ "epoch": 7.77027027027027,
811
+ "grad_norm": 0.14038431867767823,
812
+ "learning_rate": 1.0844434869600428e-05,
813
+ "loss": 0.002,
814
+ "step": 575
815
+ },
816
+ {
817
+ "epoch": 7.837837837837838,
818
+ "grad_norm": 0.23961488241206388,
819
+ "learning_rate": 1.0510608228234848e-05,
820
+ "loss": 0.0037,
821
+ "step": 580
822
+ },
823
+ {
824
+ "epoch": 7.905405405405405,
825
+ "grad_norm": 0.15447418897239665,
826
+ "learning_rate": 1.0185263403247256e-05,
827
+ "loss": 0.0025,
828
+ "step": 585
829
+ },
830
+ {
831
+ "epoch": 7.972972972972973,
832
+ "grad_norm": 0.16071816453151258,
833
+ "learning_rate": 9.868562820409103e-06,
834
+ "loss": 0.0023,
835
+ "step": 590
836
+ },
837
+ {
838
+ "epoch": 8.04054054054054,
839
+ "grad_norm": 0.039839669116423025,
840
+ "learning_rate": 9.560664589923895e-06,
841
+ "loss": 0.001,
842
+ "step": 595
843
+ },
844
+ {
845
+ "epoch": 8.108108108108109,
846
+ "grad_norm": 0.10387037183648117,
847
+ "learning_rate": 9.261722427491953e-06,
848
+ "loss": 0.0009,
849
+ "step": 600
850
+ },
851
+ {
852
+ "epoch": 8.175675675675675,
853
+ "grad_norm": 0.07137878766657882,
854
+ "learning_rate": 8.971885577569058e-06,
855
+ "loss": 0.0012,
856
+ "step": 605
857
+ },
858
+ {
859
+ "epoch": 8.243243243243244,
860
+ "grad_norm": 0.053255760667021,
861
+ "learning_rate": 8.691298738857432e-06,
862
+ "loss": 0.0008,
863
+ "step": 610
864
+ },
865
+ {
866
+ "epoch": 8.31081081081081,
867
+ "grad_norm": 0.04651063458672636,
868
+ "learning_rate": 8.420101992066028e-06,
869
+ "loss": 0.0008,
870
+ "step": 615
871
+ },
872
+ {
873
+ "epoch": 8.378378378378379,
874
+ "grad_norm": 0.05177808162440591,
875
+ "learning_rate": 8.158430729976372e-06,
876
+ "loss": 0.0018,
877
+ "step": 620
878
+ },
879
+ {
880
+ "epoch": 8.445945945945946,
881
+ "grad_norm": 0.04613561942598636,
882
+ "learning_rate": 7.906415589848834e-06,
883
+ "loss": 0.0013,
884
+ "step": 625
885
+ },
886
+ {
887
+ "epoch": 8.513513513513514,
888
+ "grad_norm": 0.04399594723846995,
889
+ "learning_rate": 7.664182388203037e-06,
890
+ "loss": 0.0011,
891
+ "step": 630
892
+ },
893
+ {
894
+ "epoch": 8.58108108108108,
895
+ "grad_norm": 0.14384542504848677,
896
+ "learning_rate": 7.4318520580049444e-06,
897
+ "loss": 0.0011,
898
+ "step": 635
899
+ },
900
+ {
901
+ "epoch": 8.64864864864865,
902
+ "grad_norm": 0.04312970027184161,
903
+ "learning_rate": 7.209540588292083e-06,
904
+ "loss": 0.0006,
905
+ "step": 640
906
+ },
907
+ {
908
+ "epoch": 8.716216216216216,
909
+ "grad_norm": 0.05426799779526829,
910
+ "learning_rate": 6.9973589662669455e-06,
911
+ "loss": 0.0011,
912
+ "step": 645
913
+ },
914
+ {
915
+ "epoch": 8.783783783783784,
916
+ "grad_norm": 0.08582632371830831,
917
+ "learning_rate": 6.7954131218875404e-06,
918
+ "loss": 0.0008,
919
+ "step": 650
920
+ },
921
+ {
922
+ "epoch": 8.85135135135135,
923
+ "grad_norm": 0.061172982105000405,
924
+ "learning_rate": 6.603803874982687e-06,
925
+ "loss": 0.0009,
926
+ "step": 655
927
+ },
928
+ {
929
+ "epoch": 8.91891891891892,
930
+ "grad_norm": 0.051037483562492506,
931
+ "learning_rate": 6.422626884918559e-06,
932
+ "loss": 0.0008,
933
+ "step": 660
934
+ },
935
+ {
936
+ "epoch": 8.986486486486486,
937
+ "grad_norm": 0.20515275466613603,
938
+ "learning_rate": 6.2519726028415145e-06,
939
+ "loss": 0.0011,
940
+ "step": 665
941
+ },
942
+ {
943
+ "epoch": 9.054054054054054,
944
+ "grad_norm": 0.011066746080717657,
945
+ "learning_rate": 6.091926226521089e-06,
946
+ "loss": 0.0006,
947
+ "step": 670
948
+ },
949
+ {
950
+ "epoch": 9.121621621621621,
951
+ "grad_norm": 0.029851844251962547,
952
+ "learning_rate": 5.942567657815696e-06,
953
+ "loss": 0.0006,
954
+ "step": 675
955
+ },
956
+ {
957
+ "epoch": 9.18918918918919,
958
+ "grad_norm": 0.02610342765221721,
959
+ "learning_rate": 5.8039714627822754e-06,
960
+ "loss": 0.0003,
961
+ "step": 680
962
+ },
963
+ {
964
+ "epoch": 9.256756756756756,
965
+ "grad_norm": 0.0735024063013663,
966
+ "learning_rate": 5.676206834449797e-06,
967
+ "loss": 0.0005,
968
+ "step": 685
969
+ },
970
+ {
971
+ "epoch": 9.324324324324325,
972
+ "grad_norm": 0.011455265457415495,
973
+ "learning_rate": 5.55933755827518e-06,
974
+ "loss": 0.0008,
975
+ "step": 690
976
+ },
977
+ {
978
+ "epoch": 9.391891891891891,
979
+ "grad_norm": 0.02585028852152204,
980
+ "learning_rate": 5.453421980298957e-06,
981
+ "loss": 0.0004,
982
+ "step": 695
983
+ },
984
+ {
985
+ "epoch": 9.45945945945946,
986
+ "grad_norm": 0.04445082056397156,
987
+ "learning_rate": 5.358512978016445e-06,
988
+ "loss": 0.0006,
989
+ "step": 700
990
+ },
991
+ {
992
+ "epoch": 9.527027027027026,
993
+ "grad_norm": 0.15020322618596355,
994
+ "learning_rate": 5.27465793397911e-06,
995
+ "loss": 0.0005,
996
+ "step": 705
997
+ },
998
+ {
999
+ "epoch": 9.594594594594595,
1000
+ "grad_norm": 0.03409990489832852,
1001
+ "learning_rate": 5.201898712139201e-06,
1002
+ "loss": 0.0003,
1003
+ "step": 710
1004
+ },
1005
+ {
1006
+ "epoch": 9.662162162162161,
1007
+ "grad_norm": 0.01537043466452054,
1008
+ "learning_rate": 5.1402716369495194e-06,
1009
+ "loss": 0.0003,
1010
+ "step": 715
1011
+ },
1012
+ {
1013
+ "epoch": 9.72972972972973,
1014
+ "grad_norm": 0.09960225768992337,
1015
+ "learning_rate": 5.089807475228711e-06,
1016
+ "loss": 0.0004,
1017
+ "step": 720
1018
+ },
1019
+ {
1020
+ "epoch": 9.797297297297296,
1021
+ "grad_norm": 0.012056853972605454,
1022
+ "learning_rate": 5.050531420801205e-06,
1023
+ "loss": 0.0005,
1024
+ "step": 725
1025
+ },
1026
+ {
1027
+ "epoch": 9.864864864864865,
1028
+ "grad_norm": 0.032615314716716644,
1029
+ "learning_rate": 5.022463081919386e-06,
1030
+ "loss": 0.0003,
1031
+ "step": 730
1032
+ },
1033
+ {
1034
+ "epoch": 9.932432432432432,
1035
+ "grad_norm": 0.042062745236192,
1036
+ "learning_rate": 5.005616471474332e-06,
1037
+ "loss": 0.0003,
1038
+ "step": 735
1039
+ },
1040
+ {
1041
+ "epoch": 10.0,
1042
+ "grad_norm": 0.015918832195194826,
1043
+ "learning_rate": 5e-06,
1044
+ "loss": 0.0004,
1045
+ "step": 740
1046
+ },
1047
+ {
1048
+ "epoch": 10.0,
1049
+ "step": 740,
1050
+ "total_flos": 1253741243269120.0,
1051
+ "train_loss": 0.05100687126286101,
1052
+ "train_runtime": 15432.4671,
1053
+ "train_samples_per_second": 6.13,
1054
+ "train_steps_per_second": 0.048
1055
+ }
1056
+ ],
1057
+ "logging_steps": 5,
1058
+ "max_steps": 740,
1059
+ "num_input_tokens_seen": 0,
1060
+ "num_train_epochs": 10,
1061
+ "save_steps": 500,
1062
+ "stateful_callbacks": {
1063
+ "TrainerControl": {
1064
+ "args": {
1065
+ "should_epoch_stop": false,
1066
+ "should_evaluate": false,
1067
+ "should_log": false,
1068
+ "should_save": true,
1069
+ "should_training_stop": true
1070
+ },
1071
+ "attributes": {}
1072
+ }
1073
+ },
1074
+ "total_flos": 1253741243269120.0,
1075
+ "train_batch_size": 16,
1076
+ "trial_name": null,
1077
+ "trial_params": null
1078
+ }