weizhepei commited on
Commit
e6fd19b
·
verified ·
1 Parent(s): 604e113

Model save

Browse files
Files changed (5) hide show
  1. README.md +58 -0
  2. all_results.json +8 -0
  3. generation_config.json +12 -0
  4. train_results.json +8 -0
  5. trainer_state.json +1078 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.1-8B-Instruct
3
+ library_name: transformers
4
+ model_name: Llama3.1-8B-WebArena-Lite-SFT-epoch-5
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for Llama3.1-8B-WebArena-Lite-SFT-epoch-5
13
+
14
+ This model is a fine-tuned version of [meta-llama/Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-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/Llama3.1-8B-WebArena-Lite-SFT-epoch-5", 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/ht64pcn4)
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": 519543052959744.0,
3
+ "train_loss": 0.11201934994676628,
4
+ "train_runtime": 13636.3107,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 3.469,
7
+ "train_steps_per_second": 0.054
8
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 128000,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 128001,
6
+ 128008,
7
+ 128009
8
+ ],
9
+ "temperature": 0.6,
10
+ "top_p": 0.9,
11
+ "transformers_version": "4.49.0"
12
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 519543052959744.0,
3
+ "train_loss": 0.11201934994676628,
4
+ "train_runtime": 13636.3107,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 3.469,
7
+ "train_steps_per_second": 0.054
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1078 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
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": 0.033783783783783786,
13
+ "grad_norm": 8.29013495953546,
14
+ "learning_rate": 6.7567567567567575e-06,
15
+ "loss": 0.5181,
16
+ "step": 5
17
+ },
18
+ {
19
+ "epoch": 0.06756756756756757,
20
+ "grad_norm": 4.196123392105515,
21
+ "learning_rate": 1.3513513513513515e-05,
22
+ "loss": 0.3567,
23
+ "step": 10
24
+ },
25
+ {
26
+ "epoch": 0.10135135135135136,
27
+ "grad_norm": 3.2047846933733215,
28
+ "learning_rate": 2.0270270270270273e-05,
29
+ "loss": 0.3399,
30
+ "step": 15
31
+ },
32
+ {
33
+ "epoch": 0.13513513513513514,
34
+ "grad_norm": 3.152224162446417,
35
+ "learning_rate": 2.702702702702703e-05,
36
+ "loss": 0.2552,
37
+ "step": 20
38
+ },
39
+ {
40
+ "epoch": 0.16891891891891891,
41
+ "grad_norm": 3.649791800007489,
42
+ "learning_rate": 3.3783783783783784e-05,
43
+ "loss": 0.2705,
44
+ "step": 25
45
+ },
46
+ {
47
+ "epoch": 0.20270270270270271,
48
+ "grad_norm": 3.4730019079608607,
49
+ "learning_rate": 4.0540540540540545e-05,
50
+ "loss": 0.2647,
51
+ "step": 30
52
+ },
53
+ {
54
+ "epoch": 0.23648648648648649,
55
+ "grad_norm": 5.8436815551949195,
56
+ "learning_rate": 4.72972972972973e-05,
57
+ "loss": 0.3513,
58
+ "step": 35
59
+ },
60
+ {
61
+ "epoch": 0.2702702702702703,
62
+ "grad_norm": 2.9019987565526995,
63
+ "learning_rate": 4.9997978016429605e-05,
64
+ "loss": 0.4888,
65
+ "step": 40
66
+ },
67
+ {
68
+ "epoch": 0.30405405405405406,
69
+ "grad_norm": 2.9397780048282027,
70
+ "learning_rate": 4.9985622766211935e-05,
71
+ "loss": 0.2713,
72
+ "step": 45
73
+ },
74
+ {
75
+ "epoch": 0.33783783783783783,
76
+ "grad_norm": 2.8421186668651917,
77
+ "learning_rate": 4.996204175076325e-05,
78
+ "loss": 0.2629,
79
+ "step": 50
80
+ },
81
+ {
82
+ "epoch": 0.3716216216216216,
83
+ "grad_norm": 2.798031701878566,
84
+ "learning_rate": 4.99272467427147e-05,
85
+ "loss": 0.3088,
86
+ "step": 55
87
+ },
88
+ {
89
+ "epoch": 0.40540540540540543,
90
+ "grad_norm": 2.70572378151596,
91
+ "learning_rate": 4.9881255113192526e-05,
92
+ "loss": 0.2351,
93
+ "step": 60
94
+ },
95
+ {
96
+ "epoch": 0.4391891891891892,
97
+ "grad_norm": 2.6573100082008914,
98
+ "learning_rate": 4.982408982314565e-05,
99
+ "loss": 0.2722,
100
+ "step": 65
101
+ },
102
+ {
103
+ "epoch": 0.47297297297297297,
104
+ "grad_norm": 2.613947983515025,
105
+ "learning_rate": 4.975577941188258e-05,
106
+ "loss": 0.2561,
107
+ "step": 70
108
+ },
109
+ {
110
+ "epoch": 0.5067567567567568,
111
+ "grad_norm": 2.4731517146590503,
112
+ "learning_rate": 4.967635798282344e-05,
113
+ "loss": 0.2895,
114
+ "step": 75
115
+ },
116
+ {
117
+ "epoch": 0.5405405405405406,
118
+ "grad_norm": 2.024939244762624,
119
+ "learning_rate": 4.958586518647407e-05,
120
+ "loss": 0.2432,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.5743243243243243,
125
+ "grad_norm": 2.075950925366443,
126
+ "learning_rate": 4.9484346200630855e-05,
127
+ "loss": 0.272,
128
+ "step": 85
129
+ },
130
+ {
131
+ "epoch": 0.6081081081081081,
132
+ "grad_norm": 2.0824590269407257,
133
+ "learning_rate": 4.937185170782607e-05,
134
+ "loss": 0.2465,
135
+ "step": 90
136
+ },
137
+ {
138
+ "epoch": 0.6418918918918919,
139
+ "grad_norm": 2.1834298841545765,
140
+ "learning_rate": 4.9248437870025035e-05,
141
+ "loss": 0.2561,
142
+ "step": 95
143
+ },
144
+ {
145
+ "epoch": 0.6756756756756757,
146
+ "grad_norm": 3.0028526613842756,
147
+ "learning_rate": 4.911416630058772e-05,
148
+ "loss": 0.2112,
149
+ "step": 100
150
+ },
151
+ {
152
+ "epoch": 0.7094594594594594,
153
+ "grad_norm": 6.923011878473891,
154
+ "learning_rate": 4.896910403350873e-05,
155
+ "loss": 0.2703,
156
+ "step": 105
157
+ },
158
+ {
159
+ "epoch": 0.7432432432432432,
160
+ "grad_norm": 1.5341551012824088,
161
+ "learning_rate": 4.88133234899512e-05,
162
+ "loss": 0.2445,
163
+ "step": 110
164
+ },
165
+ {
166
+ "epoch": 0.777027027027027,
167
+ "grad_norm": 1.9243858701641752,
168
+ "learning_rate": 4.864690244209105e-05,
169
+ "loss": 0.208,
170
+ "step": 115
171
+ },
172
+ {
173
+ "epoch": 0.8108108108108109,
174
+ "grad_norm": 1.6400221556474566,
175
+ "learning_rate": 4.8469923974289874e-05,
176
+ "loss": 0.2879,
177
+ "step": 120
178
+ },
179
+ {
180
+ "epoch": 0.8445945945945946,
181
+ "grad_norm": 2.7494322371602946,
182
+ "learning_rate": 4.828247644161577e-05,
183
+ "loss": 0.2482,
184
+ "step": 125
185
+ },
186
+ {
187
+ "epoch": 0.8783783783783784,
188
+ "grad_norm": 1.7599344699077832,
189
+ "learning_rate": 4.808465342573274e-05,
190
+ "loss": 0.2587,
191
+ "step": 130
192
+ },
193
+ {
194
+ "epoch": 0.9121621621621622,
195
+ "grad_norm": 1.6037855553499427,
196
+ "learning_rate": 4.787655368818087e-05,
197
+ "loss": 0.2024,
198
+ "step": 135
199
+ },
200
+ {
201
+ "epoch": 0.9459459459459459,
202
+ "grad_norm": 1.3955654859145117,
203
+ "learning_rate": 4.765828112107034e-05,
204
+ "loss": 0.2557,
205
+ "step": 140
206
+ },
207
+ {
208
+ "epoch": 0.9797297297297297,
209
+ "grad_norm": 1.288730609348733,
210
+ "learning_rate": 4.742994469521421e-05,
211
+ "loss": 0.2232,
212
+ "step": 145
213
+ },
214
+ {
215
+ "epoch": 1.0135135135135136,
216
+ "grad_norm": 1.7061202812270857,
217
+ "learning_rate": 4.719165840572557e-05,
218
+ "loss": 0.2171,
219
+ "step": 150
220
+ },
221
+ {
222
+ "epoch": 1.0472972972972974,
223
+ "grad_norm": 2.1625112154042254,
224
+ "learning_rate": 4.694354121510644e-05,
225
+ "loss": 0.1566,
226
+ "step": 155
227
+ },
228
+ {
229
+ "epoch": 1.0810810810810811,
230
+ "grad_norm": 2.211819043240105,
231
+ "learning_rate": 4.668571699385668e-05,
232
+ "loss": 0.1689,
233
+ "step": 160
234
+ },
235
+ {
236
+ "epoch": 1.114864864864865,
237
+ "grad_norm": 1.4126203201911418,
238
+ "learning_rate": 4.641831445863265e-05,
239
+ "loss": 0.1716,
240
+ "step": 165
241
+ },
242
+ {
243
+ "epoch": 1.1486486486486487,
244
+ "grad_norm": 2.0610795868124194,
245
+ "learning_rate": 4.614146710798645e-05,
246
+ "loss": 0.1605,
247
+ "step": 170
248
+ },
249
+ {
250
+ "epoch": 1.1824324324324325,
251
+ "grad_norm": 1.7608279284054584,
252
+ "learning_rate": 4.585531315571788e-05,
253
+ "loss": 0.1594,
254
+ "step": 175
255
+ },
256
+ {
257
+ "epoch": 1.2162162162162162,
258
+ "grad_norm": 1.5806659537041632,
259
+ "learning_rate": 4.555999546187229e-05,
260
+ "loss": 0.1484,
261
+ "step": 180
262
+ },
263
+ {
264
+ "epoch": 1.25,
265
+ "grad_norm": 1.5935232233887906,
266
+ "learning_rate": 4.5255661461418854e-05,
267
+ "loss": 0.1383,
268
+ "step": 185
269
+ },
270
+ {
271
+ "epoch": 1.2837837837837838,
272
+ "grad_norm": 1.609634736806882,
273
+ "learning_rate": 4.4942463090644896e-05,
274
+ "loss": 0.1547,
275
+ "step": 190
276
+ },
277
+ {
278
+ "epoch": 1.3175675675675675,
279
+ "grad_norm": 1.588101249847805,
280
+ "learning_rate": 4.462055671130289e-05,
281
+ "loss": 0.1482,
282
+ "step": 195
283
+ },
284
+ {
285
+ "epoch": 1.3513513513513513,
286
+ "grad_norm": 1.5585026200892942,
287
+ "learning_rate": 4.4290103032548094e-05,
288
+ "loss": 0.146,
289
+ "step": 200
290
+ },
291
+ {
292
+ "epoch": 1.385135135135135,
293
+ "grad_norm": 1.1467440704851377,
294
+ "learning_rate": 4.395126703070589e-05,
295
+ "loss": 0.1446,
296
+ "step": 205
297
+ },
298
+ {
299
+ "epoch": 1.4189189189189189,
300
+ "grad_norm": 0.9207690008726266,
301
+ "learning_rate": 4.360421786690862e-05,
302
+ "loss": 0.151,
303
+ "step": 210
304
+ },
305
+ {
306
+ "epoch": 1.4527027027027026,
307
+ "grad_norm": 1.1893240387302906,
308
+ "learning_rate": 4.324912880264326e-05,
309
+ "loss": 0.1451,
310
+ "step": 215
311
+ },
312
+ {
313
+ "epoch": 1.4864864864864864,
314
+ "grad_norm": 1.1348088575018935,
315
+ "learning_rate": 4.288617711325207e-05,
316
+ "loss": 0.133,
317
+ "step": 220
318
+ },
319
+ {
320
+ "epoch": 1.5202702702702702,
321
+ "grad_norm": 1.2969069765004013,
322
+ "learning_rate": 4.251554399942928e-05,
323
+ "loss": 0.1482,
324
+ "step": 225
325
+ },
326
+ {
327
+ "epoch": 1.554054054054054,
328
+ "grad_norm": 1.1549621334885833,
329
+ "learning_rate": 4.21374144967581e-05,
330
+ "loss": 0.1589,
331
+ "step": 230
332
+ },
333
+ {
334
+ "epoch": 1.5878378378378377,
335
+ "grad_norm": 1.644137743103509,
336
+ "learning_rate": 4.1751977383333224e-05,
337
+ "loss": 0.1407,
338
+ "step": 235
339
+ },
340
+ {
341
+ "epoch": 1.6216216216216215,
342
+ "grad_norm": 1.3525975163496182,
343
+ "learning_rate": 4.1359425085514906e-05,
344
+ "loss": 0.1363,
345
+ "step": 240
346
+ },
347
+ {
348
+ "epoch": 1.6554054054054053,
349
+ "grad_norm": 2.0901857582974217,
350
+ "learning_rate": 4.095995358186162e-05,
351
+ "loss": 0.1772,
352
+ "step": 245
353
+ },
354
+ {
355
+ "epoch": 1.689189189189189,
356
+ "grad_norm": 1.6270221012827155,
357
+ "learning_rate": 4.055376230528936e-05,
358
+ "loss": 0.1387,
359
+ "step": 250
360
+ },
361
+ {
362
+ "epoch": 1.722972972972973,
363
+ "grad_norm": 1.430683434287028,
364
+ "learning_rate": 4.0141054043506406e-05,
365
+ "loss": 0.1655,
366
+ "step": 255
367
+ },
368
+ {
369
+ "epoch": 1.7567567567567568,
370
+ "grad_norm": 0.8114050618763757,
371
+ "learning_rate": 3.972203483777315e-05,
372
+ "loss": 0.1366,
373
+ "step": 260
374
+ },
375
+ {
376
+ "epoch": 1.7905405405405406,
377
+ "grad_norm": 1.269223528535312,
378
+ "learning_rate": 3.929691388003772e-05,
379
+ "loss": 0.1485,
380
+ "step": 265
381
+ },
382
+ {
383
+ "epoch": 1.8243243243243243,
384
+ "grad_norm": 1.13503713932199,
385
+ "learning_rate": 3.886590340849852e-05,
386
+ "loss": 0.1197,
387
+ "step": 270
388
+ },
389
+ {
390
+ "epoch": 1.8581081081081081,
391
+ "grad_norm": 0.9877113608753568,
392
+ "learning_rate": 3.842921860164607e-05,
393
+ "loss": 0.1334,
394
+ "step": 275
395
+ },
396
+ {
397
+ "epoch": 1.8918918918918919,
398
+ "grad_norm": 0.8713265904231167,
399
+ "learning_rate": 3.798707747083694e-05,
400
+ "loss": 0.1334,
401
+ "step": 280
402
+ },
403
+ {
404
+ "epoch": 1.9256756756756757,
405
+ "grad_norm": 0.8661557486157838,
406
+ "learning_rate": 3.753970075145322e-05,
407
+ "loss": 0.1429,
408
+ "step": 285
409
+ },
410
+ {
411
+ "epoch": 1.9594594594594594,
412
+ "grad_norm": 1.219122887579692,
413
+ "learning_rate": 3.7087311792702265e-05,
414
+ "loss": 0.1356,
415
+ "step": 290
416
+ },
417
+ {
418
+ "epoch": 1.9932432432432432,
419
+ "grad_norm": 1.1038755273323901,
420
+ "learning_rate": 3.663013644611139e-05,
421
+ "loss": 0.1349,
422
+ "step": 295
423
+ },
424
+ {
425
+ "epoch": 2.027027027027027,
426
+ "grad_norm": 0.7179183269742409,
427
+ "learning_rate": 3.616840295277328e-05,
428
+ "loss": 0.0975,
429
+ "step": 300
430
+ },
431
+ {
432
+ "epoch": 2.060810810810811,
433
+ "grad_norm": 0.7521542943209245,
434
+ "learning_rate": 3.5702341829398525e-05,
435
+ "loss": 0.096,
436
+ "step": 305
437
+ },
438
+ {
439
+ "epoch": 2.0945945945945947,
440
+ "grad_norm": 0.7622848192686278,
441
+ "learning_rate": 3.523218575323198e-05,
442
+ "loss": 0.0947,
443
+ "step": 310
444
+ },
445
+ {
446
+ "epoch": 2.1283783783783785,
447
+ "grad_norm": 1.25703408706443,
448
+ "learning_rate": 3.475816944589058e-05,
449
+ "loss": 0.0937,
450
+ "step": 315
451
+ },
452
+ {
453
+ "epoch": 2.1621621621621623,
454
+ "grad_norm": 0.910003923674188,
455
+ "learning_rate": 3.4280529556180404e-05,
456
+ "loss": 0.0998,
457
+ "step": 320
458
+ },
459
+ {
460
+ "epoch": 2.195945945945946,
461
+ "grad_norm": 0.6770052864732585,
462
+ "learning_rate": 3.379950454195172e-05,
463
+ "loss": 0.0714,
464
+ "step": 325
465
+ },
466
+ {
467
+ "epoch": 2.22972972972973,
468
+ "grad_norm": 0.9501804841017806,
469
+ "learning_rate": 3.331533455105084e-05,
470
+ "loss": 0.0856,
471
+ "step": 330
472
+ },
473
+ {
474
+ "epoch": 2.2635135135135136,
475
+ "grad_norm": 0.9502759678893843,
476
+ "learning_rate": 3.2828261301428206e-05,
477
+ "loss": 0.0996,
478
+ "step": 335
479
+ },
480
+ {
481
+ "epoch": 2.2972972972972974,
482
+ "grad_norm": 1.40596962540061,
483
+ "learning_rate": 3.23385279604627e-05,
484
+ "loss": 0.0691,
485
+ "step": 340
486
+ },
487
+ {
488
+ "epoch": 2.331081081081081,
489
+ "grad_norm": 0.850391778909355,
490
+ "learning_rate": 3.18463790235623e-05,
491
+ "loss": 0.0828,
492
+ "step": 345
493
+ },
494
+ {
495
+ "epoch": 2.364864864864865,
496
+ "grad_norm": 0.9046270921885601,
497
+ "learning_rate": 3.135206019210167e-05,
498
+ "loss": 0.0829,
499
+ "step": 350
500
+ },
501
+ {
502
+ "epoch": 2.3986486486486487,
503
+ "grad_norm": 1.344251528726258,
504
+ "learning_rate": 3.085581825075782e-05,
505
+ "loss": 0.0761,
506
+ "step": 355
507
+ },
508
+ {
509
+ "epoch": 2.4324324324324325,
510
+ "grad_norm": 1.3818381192158777,
511
+ "learning_rate": 3.0357900944304774e-05,
512
+ "loss": 0.0912,
513
+ "step": 360
514
+ },
515
+ {
516
+ "epoch": 2.4662162162162162,
517
+ "grad_norm": 0.8834881506524517,
518
+ "learning_rate": 2.9858556853929048e-05,
519
+ "loss": 0.0843,
520
+ "step": 365
521
+ },
522
+ {
523
+ "epoch": 2.5,
524
+ "grad_norm": 0.9107496260787339,
525
+ "learning_rate": 2.9358035273127483e-05,
526
+ "loss": 0.0825,
527
+ "step": 370
528
+ },
529
+ {
530
+ "epoch": 2.5337837837837838,
531
+ "grad_norm": 0.5784107872034662,
532
+ "learning_rate": 2.8856586083249487e-05,
533
+ "loss": 0.0721,
534
+ "step": 375
535
+ },
536
+ {
537
+ "epoch": 2.5675675675675675,
538
+ "grad_norm": 0.7773606211954556,
539
+ "learning_rate": 2.83544596287458e-05,
540
+ "loss": 0.0874,
541
+ "step": 380
542
+ },
543
+ {
544
+ "epoch": 2.6013513513513513,
545
+ "grad_norm": 0.6678593340764952,
546
+ "learning_rate": 2.785190659218604e-05,
547
+ "loss": 0.0859,
548
+ "step": 385
549
+ },
550
+ {
551
+ "epoch": 2.635135135135135,
552
+ "grad_norm": 0.5939162753157207,
553
+ "learning_rate": 2.7349177869107462e-05,
554
+ "loss": 0.0809,
555
+ "step": 390
556
+ },
557
+ {
558
+ "epoch": 2.668918918918919,
559
+ "grad_norm": 0.7283077606960382,
560
+ "learning_rate": 2.684652444275741e-05,
561
+ "loss": 0.0733,
562
+ "step": 395
563
+ },
564
+ {
565
+ "epoch": 2.7027027027027026,
566
+ "grad_norm": 0.6502593242253658,
567
+ "learning_rate": 2.634419725879193e-05,
568
+ "loss": 0.0726,
569
+ "step": 400
570
+ },
571
+ {
572
+ "epoch": 2.7364864864864864,
573
+ "grad_norm": 0.6122029804466594,
574
+ "learning_rate": 2.58424470999932e-05,
575
+ "loss": 0.088,
576
+ "step": 405
577
+ },
578
+ {
579
+ "epoch": 2.77027027027027,
580
+ "grad_norm": 0.799053842544273,
581
+ "learning_rate": 2.534152446106825e-05,
582
+ "loss": 0.0729,
583
+ "step": 410
584
+ },
585
+ {
586
+ "epoch": 2.804054054054054,
587
+ "grad_norm": 0.7095023200640361,
588
+ "learning_rate": 2.4841679423591523e-05,
589
+ "loss": 0.0668,
590
+ "step": 415
591
+ },
592
+ {
593
+ "epoch": 2.8378378378378377,
594
+ "grad_norm": 0.5870222541144963,
595
+ "learning_rate": 2.4343161531153647e-05,
596
+ "loss": 0.0811,
597
+ "step": 420
598
+ },
599
+ {
600
+ "epoch": 2.8716216216216215,
601
+ "grad_norm": 0.7486446531081316,
602
+ "learning_rate": 2.3846219664778824e-05,
603
+ "loss": 0.0743,
604
+ "step": 425
605
+ },
606
+ {
607
+ "epoch": 2.9054054054054053,
608
+ "grad_norm": 0.626979823529596,
609
+ "learning_rate": 2.3351101918672985e-05,
610
+ "loss": 0.0827,
611
+ "step": 430
612
+ },
613
+ {
614
+ "epoch": 2.939189189189189,
615
+ "grad_norm": 0.7519449237853378,
616
+ "learning_rate": 2.2858055476364822e-05,
617
+ "loss": 0.0626,
618
+ "step": 435
619
+ },
620
+ {
621
+ "epoch": 2.972972972972973,
622
+ "grad_norm": 0.6067887360251734,
623
+ "learning_rate": 2.2367326487301317e-05,
624
+ "loss": 0.065,
625
+ "step": 440
626
+ },
627
+ {
628
+ "epoch": 3.0067567567567566,
629
+ "grad_norm": 0.43015841504665747,
630
+ "learning_rate": 2.1879159943959686e-05,
631
+ "loss": 0.0588,
632
+ "step": 445
633
+ },
634
+ {
635
+ "epoch": 3.0405405405405403,
636
+ "grad_norm": 0.5166996425697965,
637
+ "learning_rate": 2.139379955953686e-05,
638
+ "loss": 0.0403,
639
+ "step": 450
640
+ },
641
+ {
642
+ "epoch": 3.074324324324324,
643
+ "grad_norm": 1.533221487720139,
644
+ "learning_rate": 2.0911487646277623e-05,
645
+ "loss": 0.045,
646
+ "step": 455
647
+ },
648
+ {
649
+ "epoch": 3.108108108108108,
650
+ "grad_norm": 1.3591414304325815,
651
+ "learning_rate": 2.0432464994502203e-05,
652
+ "loss": 0.0428,
653
+ "step": 460
654
+ },
655
+ {
656
+ "epoch": 3.141891891891892,
657
+ "grad_norm": 0.4695191476769462,
658
+ "learning_rate": 1.995697075239365e-05,
659
+ "loss": 0.0397,
660
+ "step": 465
661
+ },
662
+ {
663
+ "epoch": 3.175675675675676,
664
+ "grad_norm": 0.7689436070838513,
665
+ "learning_rate": 1.9485242306605028e-05,
666
+ "loss": 0.0388,
667
+ "step": 470
668
+ },
669
+ {
670
+ "epoch": 3.2094594594594597,
671
+ "grad_norm": 0.5707450348910151,
672
+ "learning_rate": 1.9017515163746058e-05,
673
+ "loss": 0.0398,
674
+ "step": 475
675
+ },
676
+ {
677
+ "epoch": 3.2432432432432434,
678
+ "grad_norm": 0.6311825777201817,
679
+ "learning_rate": 1.855402283280836e-05,
680
+ "loss": 0.0397,
681
+ "step": 480
682
+ },
683
+ {
684
+ "epoch": 3.277027027027027,
685
+ "grad_norm": 0.6341709550171472,
686
+ "learning_rate": 1.8094996708587958e-05,
687
+ "loss": 0.0366,
688
+ "step": 485
689
+ },
690
+ {
691
+ "epoch": 3.310810810810811,
692
+ "grad_norm": 0.37977836752911515,
693
+ "learning_rate": 1.7640665956163306e-05,
694
+ "loss": 0.0364,
695
+ "step": 490
696
+ },
697
+ {
698
+ "epoch": 3.3445945945945947,
699
+ "grad_norm": 0.3786127389464406,
700
+ "learning_rate": 1.719125739648648e-05,
701
+ "loss": 0.0316,
702
+ "step": 495
703
+ },
704
+ {
705
+ "epoch": 3.3783783783783785,
706
+ "grad_norm": 0.4003900559035512,
707
+ "learning_rate": 1.6746995393144668e-05,
708
+ "loss": 0.0348,
709
+ "step": 500
710
+ },
711
+ {
712
+ "epoch": 3.4121621621621623,
713
+ "grad_norm": 0.7612366325017373,
714
+ "learning_rate": 1.6308101740348433e-05,
715
+ "loss": 0.0373,
716
+ "step": 505
717
+ },
718
+ {
719
+ "epoch": 3.445945945945946,
720
+ "grad_norm": 0.5347572825256244,
721
+ "learning_rate": 1.5874795552202773e-05,
722
+ "loss": 0.0409,
723
+ "step": 510
724
+ },
725
+ {
726
+ "epoch": 3.47972972972973,
727
+ "grad_norm": 0.3804098138882223,
728
+ "learning_rate": 1.5447293153316163e-05,
729
+ "loss": 0.0288,
730
+ "step": 515
731
+ },
732
+ {
733
+ "epoch": 3.5135135135135136,
734
+ "grad_norm": 0.7846805185343385,
735
+ "learning_rate": 1.5025807970802252e-05,
736
+ "loss": 0.0367,
737
+ "step": 520
738
+ },
739
+ {
740
+ "epoch": 3.5472972972972974,
741
+ "grad_norm": 0.6955590006040003,
742
+ "learning_rate": 1.4610550427728103e-05,
743
+ "loss": 0.0355,
744
+ "step": 525
745
+ },
746
+ {
747
+ "epoch": 3.581081081081081,
748
+ "grad_norm": 0.6081379754765148,
749
+ "learning_rate": 1.4201727838062181e-05,
750
+ "loss": 0.0384,
751
+ "step": 530
752
+ },
753
+ {
754
+ "epoch": 3.614864864864865,
755
+ "grad_norm": 0.6589981768125784,
756
+ "learning_rate": 1.3799544303174514e-05,
757
+ "loss": 0.0454,
758
+ "step": 535
759
+ },
760
+ {
761
+ "epoch": 3.6486486486486487,
762
+ "grad_norm": 0.6007971088374728,
763
+ "learning_rate": 1.3404200609940754e-05,
764
+ "loss": 0.0397,
765
+ "step": 540
766
+ },
767
+ {
768
+ "epoch": 3.6824324324324325,
769
+ "grad_norm": 0.7325992049485736,
770
+ "learning_rate": 1.3015894130500977e-05,
771
+ "loss": 0.0374,
772
+ "step": 545
773
+ },
774
+ {
775
+ "epoch": 3.7162162162162162,
776
+ "grad_norm": 0.37094326467204003,
777
+ "learning_rate": 1.2634818723723174e-05,
778
+ "loss": 0.0318,
779
+ "step": 550
780
+ },
781
+ {
782
+ "epoch": 3.75,
783
+ "grad_norm": 0.9146146847365886,
784
+ "learning_rate": 1.2261164638420832e-05,
785
+ "loss": 0.0328,
786
+ "step": 555
787
+ },
788
+ {
789
+ "epoch": 3.7837837837837838,
790
+ "grad_norm": 0.5648038565491271,
791
+ "learning_rate": 1.1895118418372734e-05,
792
+ "loss": 0.0356,
793
+ "step": 560
794
+ },
795
+ {
796
+ "epoch": 3.8175675675675675,
797
+ "grad_norm": 0.37538544226428106,
798
+ "learning_rate": 1.1536862809192518e-05,
799
+ "loss": 0.0398,
800
+ "step": 565
801
+ },
802
+ {
803
+ "epoch": 3.8513513513513513,
804
+ "grad_norm": 0.4038884653715223,
805
+ "learning_rate": 1.1186576667094342e-05,
806
+ "loss": 0.0271,
807
+ "step": 570
808
+ },
809
+ {
810
+ "epoch": 3.885135135135135,
811
+ "grad_norm": 0.6627132587861069,
812
+ "learning_rate": 1.0844434869600428e-05,
813
+ "loss": 0.0338,
814
+ "step": 575
815
+ },
816
+ {
817
+ "epoch": 3.918918918918919,
818
+ "grad_norm": 0.7041617251510983,
819
+ "learning_rate": 1.0510608228234848e-05,
820
+ "loss": 0.0391,
821
+ "step": 580
822
+ },
823
+ {
824
+ "epoch": 3.9527027027027026,
825
+ "grad_norm": 0.8131142697010945,
826
+ "learning_rate": 1.0185263403247256e-05,
827
+ "loss": 0.0342,
828
+ "step": 585
829
+ },
830
+ {
831
+ "epoch": 3.9864864864864864,
832
+ "grad_norm": 0.5920567955107612,
833
+ "learning_rate": 9.868562820409103e-06,
834
+ "loss": 0.0337,
835
+ "step": 590
836
+ },
837
+ {
838
+ "epoch": 4.02027027027027,
839
+ "grad_norm": 0.3120828177149261,
840
+ "learning_rate": 9.560664589923895e-06,
841
+ "loss": 0.0226,
842
+ "step": 595
843
+ },
844
+ {
845
+ "epoch": 4.054054054054054,
846
+ "grad_norm": 0.2900775148851475,
847
+ "learning_rate": 9.261722427491953e-06,
848
+ "loss": 0.0147,
849
+ "step": 600
850
+ },
851
+ {
852
+ "epoch": 4.087837837837838,
853
+ "grad_norm": 0.22784089723926296,
854
+ "learning_rate": 8.971885577569058e-06,
855
+ "loss": 0.0122,
856
+ "step": 605
857
+ },
858
+ {
859
+ "epoch": 4.121621621621622,
860
+ "grad_norm": 0.2693260557173335,
861
+ "learning_rate": 8.691298738857432e-06,
862
+ "loss": 0.0119,
863
+ "step": 610
864
+ },
865
+ {
866
+ "epoch": 4.155405405405405,
867
+ "grad_norm": 0.21618402703820352,
868
+ "learning_rate": 8.420101992066028e-06,
869
+ "loss": 0.0156,
870
+ "step": 615
871
+ },
872
+ {
873
+ "epoch": 4.1891891891891895,
874
+ "grad_norm": 0.24421224725586618,
875
+ "learning_rate": 8.158430729976372e-06,
876
+ "loss": 0.0119,
877
+ "step": 620
878
+ },
879
+ {
880
+ "epoch": 4.222972972972973,
881
+ "grad_norm": 0.33320132947770426,
882
+ "learning_rate": 7.906415589848834e-06,
883
+ "loss": 0.0146,
884
+ "step": 625
885
+ },
886
+ {
887
+ "epoch": 4.256756756756757,
888
+ "grad_norm": 0.33015864806759715,
889
+ "learning_rate": 7.664182388203037e-06,
890
+ "loss": 0.0145,
891
+ "step": 630
892
+ },
893
+ {
894
+ "epoch": 4.29054054054054,
895
+ "grad_norm": 0.36028113378267884,
896
+ "learning_rate": 7.4318520580049444e-06,
897
+ "loss": 0.015,
898
+ "step": 635
899
+ },
900
+ {
901
+ "epoch": 4.324324324324325,
902
+ "grad_norm": 0.26126678042031565,
903
+ "learning_rate": 7.209540588292083e-06,
904
+ "loss": 0.0134,
905
+ "step": 640
906
+ },
907
+ {
908
+ "epoch": 4.358108108108108,
909
+ "grad_norm": 0.549701638585684,
910
+ "learning_rate": 6.9973589662669455e-06,
911
+ "loss": 0.0136,
912
+ "step": 645
913
+ },
914
+ {
915
+ "epoch": 4.391891891891892,
916
+ "grad_norm": 0.30168954386975444,
917
+ "learning_rate": 6.7954131218875404e-06,
918
+ "loss": 0.0151,
919
+ "step": 650
920
+ },
921
+ {
922
+ "epoch": 4.425675675675675,
923
+ "grad_norm": 0.43179940752419366,
924
+ "learning_rate": 6.603803874982687e-06,
925
+ "loss": 0.017,
926
+ "step": 655
927
+ },
928
+ {
929
+ "epoch": 4.45945945945946,
930
+ "grad_norm": 0.26869880474008145,
931
+ "learning_rate": 6.422626884918559e-06,
932
+ "loss": 0.0131,
933
+ "step": 660
934
+ },
935
+ {
936
+ "epoch": 4.493243243243243,
937
+ "grad_norm": 0.4055684732783103,
938
+ "learning_rate": 6.2519726028415145e-06,
939
+ "loss": 0.017,
940
+ "step": 665
941
+ },
942
+ {
943
+ "epoch": 4.527027027027027,
944
+ "grad_norm": 0.334734546863203,
945
+ "learning_rate": 6.091926226521089e-06,
946
+ "loss": 0.0135,
947
+ "step": 670
948
+ },
949
+ {
950
+ "epoch": 4.5608108108108105,
951
+ "grad_norm": 0.5546791873819881,
952
+ "learning_rate": 5.942567657815696e-06,
953
+ "loss": 0.0138,
954
+ "step": 675
955
+ },
956
+ {
957
+ "epoch": 4.594594594594595,
958
+ "grad_norm": 0.4128562858606657,
959
+ "learning_rate": 5.8039714627822754e-06,
960
+ "loss": 0.0123,
961
+ "step": 680
962
+ },
963
+ {
964
+ "epoch": 4.628378378378378,
965
+ "grad_norm": 0.33985067404728464,
966
+ "learning_rate": 5.676206834449797e-06,
967
+ "loss": 0.014,
968
+ "step": 685
969
+ },
970
+ {
971
+ "epoch": 4.662162162162162,
972
+ "grad_norm": 0.3646278429623203,
973
+ "learning_rate": 5.55933755827518e-06,
974
+ "loss": 0.0105,
975
+ "step": 690
976
+ },
977
+ {
978
+ "epoch": 4.695945945945946,
979
+ "grad_norm": 0.2663115053288746,
980
+ "learning_rate": 5.453421980298957e-06,
981
+ "loss": 0.012,
982
+ "step": 695
983
+ },
984
+ {
985
+ "epoch": 4.72972972972973,
986
+ "grad_norm": 0.30374384905412244,
987
+ "learning_rate": 5.358512978016445e-06,
988
+ "loss": 0.0108,
989
+ "step": 700
990
+ },
991
+ {
992
+ "epoch": 4.763513513513513,
993
+ "grad_norm": 0.47429438189169476,
994
+ "learning_rate": 5.27465793397911e-06,
995
+ "loss": 0.0143,
996
+ "step": 705
997
+ },
998
+ {
999
+ "epoch": 4.797297297297297,
1000
+ "grad_norm": 0.2683488363965782,
1001
+ "learning_rate": 5.201898712139201e-06,
1002
+ "loss": 0.0121,
1003
+ "step": 710
1004
+ },
1005
+ {
1006
+ "epoch": 4.831081081081081,
1007
+ "grad_norm": 0.4245927469089233,
1008
+ "learning_rate": 5.1402716369495194e-06,
1009
+ "loss": 0.0125,
1010
+ "step": 715
1011
+ },
1012
+ {
1013
+ "epoch": 4.864864864864865,
1014
+ "grad_norm": 0.40748023019876795,
1015
+ "learning_rate": 5.089807475228711e-06,
1016
+ "loss": 0.0151,
1017
+ "step": 720
1018
+ },
1019
+ {
1020
+ "epoch": 4.898648648648649,
1021
+ "grad_norm": 0.2454774836937338,
1022
+ "learning_rate": 5.050531420801205e-06,
1023
+ "loss": 0.0112,
1024
+ "step": 725
1025
+ },
1026
+ {
1027
+ "epoch": 4.9324324324324325,
1028
+ "grad_norm": 0.343369387002044,
1029
+ "learning_rate": 5.022463081919386e-06,
1030
+ "loss": 0.0126,
1031
+ "step": 730
1032
+ },
1033
+ {
1034
+ "epoch": 4.966216216216216,
1035
+ "grad_norm": 0.28684110351791725,
1036
+ "learning_rate": 5.005616471474332e-06,
1037
+ "loss": 0.0116,
1038
+ "step": 735
1039
+ },
1040
+ {
1041
+ "epoch": 5.0,
1042
+ "grad_norm": 0.3365561300076367,
1043
+ "learning_rate": 5e-06,
1044
+ "loss": 0.013,
1045
+ "step": 740
1046
+ },
1047
+ {
1048
+ "epoch": 5.0,
1049
+ "step": 740,
1050
+ "total_flos": 519543052959744.0,
1051
+ "train_loss": 0.11201934994676628,
1052
+ "train_runtime": 13636.3107,
1053
+ "train_samples_per_second": 3.469,
1054
+ "train_steps_per_second": 0.054
1055
+ }
1056
+ ],
1057
+ "logging_steps": 5,
1058
+ "max_steps": 740,
1059
+ "num_input_tokens_seen": 0,
1060
+ "num_train_epochs": 5,
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": 519543052959744.0,
1075
+ "train_batch_size": 8,
1076
+ "trial_name": null,
1077
+ "trial_params": null
1078
+ }