weizhepei commited on
Commit
bcc8f1a
·
verified ·
1 Parent(s): 6e4781e

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-no-sys
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-no-sys
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-no-sys", 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/xqvns6es)
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": 435868890169344.0,
3
+ "train_loss": 0.11167457006066232,
4
+ "train_runtime": 12339.9122,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 3.833,
7
+ "train_steps_per_second": 0.06
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": 435868890169344.0,
3
+ "train_loss": 0.11167457006066232,
4
+ "train_runtime": 12339.9122,
5
+ "train_samples": 9460,
6
+ "train_samples_per_second": 3.833,
7
+ "train_steps_per_second": 0.06
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": 11.91859090228093,
14
+ "learning_rate": 6.7567567567567575e-06,
15
+ "loss": 0.9462,
16
+ "step": 5
17
+ },
18
+ {
19
+ "epoch": 0.06756756756756757,
20
+ "grad_norm": 4.53680244078793,
21
+ "learning_rate": 1.3513513513513515e-05,
22
+ "loss": 0.4666,
23
+ "step": 10
24
+ },
25
+ {
26
+ "epoch": 0.10135135135135136,
27
+ "grad_norm": 3.573903925081475,
28
+ "learning_rate": 2.0270270270270273e-05,
29
+ "loss": 0.3568,
30
+ "step": 15
31
+ },
32
+ {
33
+ "epoch": 0.13513513513513514,
34
+ "grad_norm": 3.832108029582817,
35
+ "learning_rate": 2.702702702702703e-05,
36
+ "loss": 0.264,
37
+ "step": 20
38
+ },
39
+ {
40
+ "epoch": 0.16891891891891891,
41
+ "grad_norm": 2.7780428812660225,
42
+ "learning_rate": 3.3783783783783784e-05,
43
+ "loss": 0.2717,
44
+ "step": 25
45
+ },
46
+ {
47
+ "epoch": 0.20270270270270271,
48
+ "grad_norm": 3.2015639656457204,
49
+ "learning_rate": 4.0540540540540545e-05,
50
+ "loss": 0.2704,
51
+ "step": 30
52
+ },
53
+ {
54
+ "epoch": 0.23648648648648649,
55
+ "grad_norm": 3.8305742015126354,
56
+ "learning_rate": 4.72972972972973e-05,
57
+ "loss": 0.2625,
58
+ "step": 35
59
+ },
60
+ {
61
+ "epoch": 0.2702702702702703,
62
+ "grad_norm": 2.4187710692208926,
63
+ "learning_rate": 4.9997978016429605e-05,
64
+ "loss": 0.2675,
65
+ "step": 40
66
+ },
67
+ {
68
+ "epoch": 0.30405405405405406,
69
+ "grad_norm": 2.635487937809038,
70
+ "learning_rate": 4.9985622766211935e-05,
71
+ "loss": 0.2499,
72
+ "step": 45
73
+ },
74
+ {
75
+ "epoch": 0.33783783783783783,
76
+ "grad_norm": 2.316225114816826,
77
+ "learning_rate": 4.996204175076325e-05,
78
+ "loss": 0.2621,
79
+ "step": 50
80
+ },
81
+ {
82
+ "epoch": 0.3716216216216216,
83
+ "grad_norm": 2.146594431614266,
84
+ "learning_rate": 4.99272467427147e-05,
85
+ "loss": 0.2884,
86
+ "step": 55
87
+ },
88
+ {
89
+ "epoch": 0.40540540540540543,
90
+ "grad_norm": 2.398859815430172,
91
+ "learning_rate": 4.9881255113192526e-05,
92
+ "loss": 0.2398,
93
+ "step": 60
94
+ },
95
+ {
96
+ "epoch": 0.4391891891891892,
97
+ "grad_norm": 2.954151145902332,
98
+ "learning_rate": 4.982408982314565e-05,
99
+ "loss": 0.2543,
100
+ "step": 65
101
+ },
102
+ {
103
+ "epoch": 0.47297297297297297,
104
+ "grad_norm": 1.7929864319507005,
105
+ "learning_rate": 4.975577941188258e-05,
106
+ "loss": 0.2502,
107
+ "step": 70
108
+ },
109
+ {
110
+ "epoch": 0.5067567567567568,
111
+ "grad_norm": 1.840568109556661,
112
+ "learning_rate": 4.967635798282344e-05,
113
+ "loss": 0.277,
114
+ "step": 75
115
+ },
116
+ {
117
+ "epoch": 0.5405405405405406,
118
+ "grad_norm": 1.771348286216397,
119
+ "learning_rate": 4.958586518647407e-05,
120
+ "loss": 0.2518,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.5743243243243243,
125
+ "grad_norm": 2.623290643440281,
126
+ "learning_rate": 4.9484346200630855e-05,
127
+ "loss": 0.269,
128
+ "step": 85
129
+ },
130
+ {
131
+ "epoch": 0.6081081081081081,
132
+ "grad_norm": 2.6715602356902712,
133
+ "learning_rate": 4.937185170782607e-05,
134
+ "loss": 0.2469,
135
+ "step": 90
136
+ },
137
+ {
138
+ "epoch": 0.6418918918918919,
139
+ "grad_norm": 1.5618566633614335,
140
+ "learning_rate": 4.9248437870025035e-05,
141
+ "loss": 0.2389,
142
+ "step": 95
143
+ },
144
+ {
145
+ "epoch": 0.6756756756756757,
146
+ "grad_norm": 1.6043766275036224,
147
+ "learning_rate": 4.911416630058772e-05,
148
+ "loss": 0.2034,
149
+ "step": 100
150
+ },
151
+ {
152
+ "epoch": 0.7094594594594594,
153
+ "grad_norm": 1.745799222586184,
154
+ "learning_rate": 4.896910403350873e-05,
155
+ "loss": 0.2718,
156
+ "step": 105
157
+ },
158
+ {
159
+ "epoch": 0.7432432432432432,
160
+ "grad_norm": 1.6346813492093817,
161
+ "learning_rate": 4.88133234899512e-05,
162
+ "loss": 0.2441,
163
+ "step": 110
164
+ },
165
+ {
166
+ "epoch": 0.777027027027027,
167
+ "grad_norm": 2.0436191339698486,
168
+ "learning_rate": 4.864690244209105e-05,
169
+ "loss": 0.2135,
170
+ "step": 115
171
+ },
172
+ {
173
+ "epoch": 0.8108108108108109,
174
+ "grad_norm": 1.4009046058211432,
175
+ "learning_rate": 4.8469923974289874e-05,
176
+ "loss": 0.2855,
177
+ "step": 120
178
+ },
179
+ {
180
+ "epoch": 0.8445945945945946,
181
+ "grad_norm": 2.1420433385829907,
182
+ "learning_rate": 4.828247644161577e-05,
183
+ "loss": 0.2443,
184
+ "step": 125
185
+ },
186
+ {
187
+ "epoch": 0.8783783783783784,
188
+ "grad_norm": 2.029711117623661,
189
+ "learning_rate": 4.808465342573274e-05,
190
+ "loss": 0.2413,
191
+ "step": 130
192
+ },
193
+ {
194
+ "epoch": 0.9121621621621622,
195
+ "grad_norm": 2.216487452899117,
196
+ "learning_rate": 4.787655368818087e-05,
197
+ "loss": 0.2136,
198
+ "step": 135
199
+ },
200
+ {
201
+ "epoch": 0.9459459459459459,
202
+ "grad_norm": 1.9397316210901858,
203
+ "learning_rate": 4.765828112107034e-05,
204
+ "loss": 0.2541,
205
+ "step": 140
206
+ },
207
+ {
208
+ "epoch": 0.9797297297297297,
209
+ "grad_norm": 1.2829673381620292,
210
+ "learning_rate": 4.742994469521421e-05,
211
+ "loss": 0.224,
212
+ "step": 145
213
+ },
214
+ {
215
+ "epoch": 1.0135135135135136,
216
+ "grad_norm": 1.8944274655064726,
217
+ "learning_rate": 4.719165840572557e-05,
218
+ "loss": 0.2169,
219
+ "step": 150
220
+ },
221
+ {
222
+ "epoch": 1.0472972972972974,
223
+ "grad_norm": 1.7790710619051624,
224
+ "learning_rate": 4.694354121510644e-05,
225
+ "loss": 0.1452,
226
+ "step": 155
227
+ },
228
+ {
229
+ "epoch": 1.0810810810810811,
230
+ "grad_norm": 1.5418464625119979,
231
+ "learning_rate": 4.668571699385668e-05,
232
+ "loss": 0.1531,
233
+ "step": 160
234
+ },
235
+ {
236
+ "epoch": 1.114864864864865,
237
+ "grad_norm": 1.5780966926016784,
238
+ "learning_rate": 4.641831445863265e-05,
239
+ "loss": 0.1694,
240
+ "step": 165
241
+ },
242
+ {
243
+ "epoch": 1.1486486486486487,
244
+ "grad_norm": 4.6428360428861515,
245
+ "learning_rate": 4.614146710798645e-05,
246
+ "loss": 0.1619,
247
+ "step": 170
248
+ },
249
+ {
250
+ "epoch": 1.1824324324324325,
251
+ "grad_norm": 1.4418340555320202,
252
+ "learning_rate": 4.585531315571788e-05,
253
+ "loss": 0.1611,
254
+ "step": 175
255
+ },
256
+ {
257
+ "epoch": 1.2162162162162162,
258
+ "grad_norm": 2.0529270411153333,
259
+ "learning_rate": 4.555999546187229e-05,
260
+ "loss": 0.1383,
261
+ "step": 180
262
+ },
263
+ {
264
+ "epoch": 1.25,
265
+ "grad_norm": 1.1879187127096535,
266
+ "learning_rate": 4.5255661461418854e-05,
267
+ "loss": 0.1448,
268
+ "step": 185
269
+ },
270
+ {
271
+ "epoch": 1.2837837837837838,
272
+ "grad_norm": 1.0143016739135113,
273
+ "learning_rate": 4.4942463090644896e-05,
274
+ "loss": 0.1465,
275
+ "step": 190
276
+ },
277
+ {
278
+ "epoch": 1.3175675675675675,
279
+ "grad_norm": 2.090466049666836,
280
+ "learning_rate": 4.462055671130289e-05,
281
+ "loss": 0.1615,
282
+ "step": 195
283
+ },
284
+ {
285
+ "epoch": 1.3513513513513513,
286
+ "grad_norm": 1.5608930671166168,
287
+ "learning_rate": 4.4290103032548094e-05,
288
+ "loss": 0.1412,
289
+ "step": 200
290
+ },
291
+ {
292
+ "epoch": 1.385135135135135,
293
+ "grad_norm": 1.2368408671304418,
294
+ "learning_rate": 4.395126703070589e-05,
295
+ "loss": 0.1571,
296
+ "step": 205
297
+ },
298
+ {
299
+ "epoch": 1.4189189189189189,
300
+ "grad_norm": 1.0805073107430125,
301
+ "learning_rate": 4.360421786690862e-05,
302
+ "loss": 0.1518,
303
+ "step": 210
304
+ },
305
+ {
306
+ "epoch": 1.4527027027027026,
307
+ "grad_norm": 1.9178479550142697,
308
+ "learning_rate": 4.324912880264326e-05,
309
+ "loss": 0.1476,
310
+ "step": 215
311
+ },
312
+ {
313
+ "epoch": 1.4864864864864864,
314
+ "grad_norm": 1.2390954008536137,
315
+ "learning_rate": 4.288617711325207e-05,
316
+ "loss": 0.1436,
317
+ "step": 220
318
+ },
319
+ {
320
+ "epoch": 1.5202702702702702,
321
+ "grad_norm": 1.2743622522980955,
322
+ "learning_rate": 4.251554399942928e-05,
323
+ "loss": 0.1494,
324
+ "step": 225
325
+ },
326
+ {
327
+ "epoch": 1.554054054054054,
328
+ "grad_norm": 1.1175039450703055,
329
+ "learning_rate": 4.21374144967581e-05,
330
+ "loss": 0.1482,
331
+ "step": 230
332
+ },
333
+ {
334
+ "epoch": 1.5878378378378377,
335
+ "grad_norm": 0.8990095650984934,
336
+ "learning_rate": 4.1751977383333224e-05,
337
+ "loss": 0.1412,
338
+ "step": 235
339
+ },
340
+ {
341
+ "epoch": 1.6216216216216215,
342
+ "grad_norm": 1.242740863487836,
343
+ "learning_rate": 4.1359425085514906e-05,
344
+ "loss": 0.1366,
345
+ "step": 240
346
+ },
347
+ {
348
+ "epoch": 1.6554054054054053,
349
+ "grad_norm": 1.1355786140547892,
350
+ "learning_rate": 4.095995358186162e-05,
351
+ "loss": 0.1667,
352
+ "step": 245
353
+ },
354
+ {
355
+ "epoch": 1.689189189189189,
356
+ "grad_norm": 1.1926029161081027,
357
+ "learning_rate": 4.055376230528936e-05,
358
+ "loss": 0.1308,
359
+ "step": 250
360
+ },
361
+ {
362
+ "epoch": 1.722972972972973,
363
+ "grad_norm": 1.0477428608330859,
364
+ "learning_rate": 4.0141054043506406e-05,
365
+ "loss": 0.1588,
366
+ "step": 255
367
+ },
368
+ {
369
+ "epoch": 1.7567567567567568,
370
+ "grad_norm": 2.434364212868205,
371
+ "learning_rate": 3.972203483777315e-05,
372
+ "loss": 0.1454,
373
+ "step": 260
374
+ },
375
+ {
376
+ "epoch": 1.7905405405405406,
377
+ "grad_norm": 1.025288482977718,
378
+ "learning_rate": 3.929691388003772e-05,
379
+ "loss": 0.1464,
380
+ "step": 265
381
+ },
382
+ {
383
+ "epoch": 1.8243243243243243,
384
+ "grad_norm": 1.0185166028149573,
385
+ "learning_rate": 3.886590340849852e-05,
386
+ "loss": 0.1228,
387
+ "step": 270
388
+ },
389
+ {
390
+ "epoch": 1.8581081081081081,
391
+ "grad_norm": 0.8097111792595093,
392
+ "learning_rate": 3.842921860164607e-05,
393
+ "loss": 0.13,
394
+ "step": 275
395
+ },
396
+ {
397
+ "epoch": 1.8918918918918919,
398
+ "grad_norm": 1.0056347069881086,
399
+ "learning_rate": 3.798707747083694e-05,
400
+ "loss": 0.1333,
401
+ "step": 280
402
+ },
403
+ {
404
+ "epoch": 1.9256756756756757,
405
+ "grad_norm": 1.0801697150500822,
406
+ "learning_rate": 3.753970075145322e-05,
407
+ "loss": 0.1323,
408
+ "step": 285
409
+ },
410
+ {
411
+ "epoch": 1.9594594594594594,
412
+ "grad_norm": 1.1000500869217074,
413
+ "learning_rate": 3.7087311792702265e-05,
414
+ "loss": 0.1337,
415
+ "step": 290
416
+ },
417
+ {
418
+ "epoch": 1.9932432432432432,
419
+ "grad_norm": 1.5122187810870562,
420
+ "learning_rate": 3.663013644611139e-05,
421
+ "loss": 0.1353,
422
+ "step": 295
423
+ },
424
+ {
425
+ "epoch": 2.027027027027027,
426
+ "grad_norm": 0.9578127795346132,
427
+ "learning_rate": 3.616840295277328e-05,
428
+ "loss": 0.0908,
429
+ "step": 300
430
+ },
431
+ {
432
+ "epoch": 2.060810810810811,
433
+ "grad_norm": 0.5934461784804044,
434
+ "learning_rate": 3.5702341829398525e-05,
435
+ "loss": 0.0851,
436
+ "step": 305
437
+ },
438
+ {
439
+ "epoch": 2.0945945945945947,
440
+ "grad_norm": 1.0725119181506966,
441
+ "learning_rate": 3.523218575323198e-05,
442
+ "loss": 0.0929,
443
+ "step": 310
444
+ },
445
+ {
446
+ "epoch": 2.1283783783783785,
447
+ "grad_norm": 0.7850220371401365,
448
+ "learning_rate": 3.475816944589058e-05,
449
+ "loss": 0.0911,
450
+ "step": 315
451
+ },
452
+ {
453
+ "epoch": 2.1621621621621623,
454
+ "grad_norm": 0.8527224202602178,
455
+ "learning_rate": 3.4280529556180404e-05,
456
+ "loss": 0.0929,
457
+ "step": 320
458
+ },
459
+ {
460
+ "epoch": 2.195945945945946,
461
+ "grad_norm": 1.04326596490438,
462
+ "learning_rate": 3.379950454195172e-05,
463
+ "loss": 0.0758,
464
+ "step": 325
465
+ },
466
+ {
467
+ "epoch": 2.22972972972973,
468
+ "grad_norm": 1.0501264893771993,
469
+ "learning_rate": 3.331533455105084e-05,
470
+ "loss": 0.0818,
471
+ "step": 330
472
+ },
473
+ {
474
+ "epoch": 2.2635135135135136,
475
+ "grad_norm": 0.8205870434886383,
476
+ "learning_rate": 3.2828261301428206e-05,
477
+ "loss": 0.0977,
478
+ "step": 335
479
+ },
480
+ {
481
+ "epoch": 2.2972972972972974,
482
+ "grad_norm": 0.8849982112270377,
483
+ "learning_rate": 3.23385279604627e-05,
484
+ "loss": 0.0719,
485
+ "step": 340
486
+ },
487
+ {
488
+ "epoch": 2.331081081081081,
489
+ "grad_norm": 0.7183927930700003,
490
+ "learning_rate": 3.18463790235623e-05,
491
+ "loss": 0.0873,
492
+ "step": 345
493
+ },
494
+ {
495
+ "epoch": 2.364864864864865,
496
+ "grad_norm": 0.7716121512809517,
497
+ "learning_rate": 3.135206019210167e-05,
498
+ "loss": 0.0771,
499
+ "step": 350
500
+ },
501
+ {
502
+ "epoch": 2.3986486486486487,
503
+ "grad_norm": 1.139603479425331,
504
+ "learning_rate": 3.085581825075782e-05,
505
+ "loss": 0.0709,
506
+ "step": 355
507
+ },
508
+ {
509
+ "epoch": 2.4324324324324325,
510
+ "grad_norm": 1.4786085416010715,
511
+ "learning_rate": 3.0357900944304774e-05,
512
+ "loss": 0.0902,
513
+ "step": 360
514
+ },
515
+ {
516
+ "epoch": 2.4662162162162162,
517
+ "grad_norm": 0.9071145610807975,
518
+ "learning_rate": 2.9858556853929048e-05,
519
+ "loss": 0.0833,
520
+ "step": 365
521
+ },
522
+ {
523
+ "epoch": 2.5,
524
+ "grad_norm": 0.7033676015416719,
525
+ "learning_rate": 2.9358035273127483e-05,
526
+ "loss": 0.0713,
527
+ "step": 370
528
+ },
529
+ {
530
+ "epoch": 2.5337837837837838,
531
+ "grad_norm": 0.673070630957069,
532
+ "learning_rate": 2.8856586083249487e-05,
533
+ "loss": 0.0672,
534
+ "step": 375
535
+ },
536
+ {
537
+ "epoch": 2.5675675675675675,
538
+ "grad_norm": 0.8537399411767074,
539
+ "learning_rate": 2.83544596287458e-05,
540
+ "loss": 0.0859,
541
+ "step": 380
542
+ },
543
+ {
544
+ "epoch": 2.6013513513513513,
545
+ "grad_norm": 0.5796181912861205,
546
+ "learning_rate": 2.785190659218604e-05,
547
+ "loss": 0.0794,
548
+ "step": 385
549
+ },
550
+ {
551
+ "epoch": 2.635135135135135,
552
+ "grad_norm": 0.6511959977506814,
553
+ "learning_rate": 2.7349177869107462e-05,
554
+ "loss": 0.078,
555
+ "step": 390
556
+ },
557
+ {
558
+ "epoch": 2.668918918918919,
559
+ "grad_norm": 0.975876535380539,
560
+ "learning_rate": 2.684652444275741e-05,
561
+ "loss": 0.0646,
562
+ "step": 395
563
+ },
564
+ {
565
+ "epoch": 2.7027027027027026,
566
+ "grad_norm": 0.6915954278323423,
567
+ "learning_rate": 2.634419725879193e-05,
568
+ "loss": 0.0753,
569
+ "step": 400
570
+ },
571
+ {
572
+ "epoch": 2.7364864864864864,
573
+ "grad_norm": 0.6789310321110773,
574
+ "learning_rate": 2.58424470999932e-05,
575
+ "loss": 0.0825,
576
+ "step": 405
577
+ },
578
+ {
579
+ "epoch": 2.77027027027027,
580
+ "grad_norm": 0.8180023876519836,
581
+ "learning_rate": 2.534152446106825e-05,
582
+ "loss": 0.0744,
583
+ "step": 410
584
+ },
585
+ {
586
+ "epoch": 2.804054054054054,
587
+ "grad_norm": 0.7821847539221951,
588
+ "learning_rate": 2.4841679423591523e-05,
589
+ "loss": 0.0684,
590
+ "step": 415
591
+ },
592
+ {
593
+ "epoch": 2.8378378378378377,
594
+ "grad_norm": 0.6251388492073695,
595
+ "learning_rate": 2.4343161531153647e-05,
596
+ "loss": 0.0764,
597
+ "step": 420
598
+ },
599
+ {
600
+ "epoch": 2.8716216216216215,
601
+ "grad_norm": 0.5954138849180285,
602
+ "learning_rate": 2.3846219664778824e-05,
603
+ "loss": 0.0739,
604
+ "step": 425
605
+ },
606
+ {
607
+ "epoch": 2.9054054054054053,
608
+ "grad_norm": 0.7069008525263534,
609
+ "learning_rate": 2.3351101918672985e-05,
610
+ "loss": 0.0772,
611
+ "step": 430
612
+ },
613
+ {
614
+ "epoch": 2.939189189189189,
615
+ "grad_norm": 0.6296050119251395,
616
+ "learning_rate": 2.2858055476364822e-05,
617
+ "loss": 0.0675,
618
+ "step": 435
619
+ },
620
+ {
621
+ "epoch": 2.972972972972973,
622
+ "grad_norm": 0.6716148252050718,
623
+ "learning_rate": 2.2367326487301317e-05,
624
+ "loss": 0.0612,
625
+ "step": 440
626
+ },
627
+ {
628
+ "epoch": 3.0067567567567566,
629
+ "grad_norm": 0.41293433850786937,
630
+ "learning_rate": 2.1879159943959686e-05,
631
+ "loss": 0.0583,
632
+ "step": 445
633
+ },
634
+ {
635
+ "epoch": 3.0405405405405403,
636
+ "grad_norm": 0.48242208510460555,
637
+ "learning_rate": 2.139379955953686e-05,
638
+ "loss": 0.04,
639
+ "step": 450
640
+ },
641
+ {
642
+ "epoch": 3.074324324324324,
643
+ "grad_norm": 0.5002995507342688,
644
+ "learning_rate": 2.0911487646277623e-05,
645
+ "loss": 0.0379,
646
+ "step": 455
647
+ },
648
+ {
649
+ "epoch": 3.108108108108108,
650
+ "grad_norm": 0.7307976680525754,
651
+ "learning_rate": 2.0432464994502203e-05,
652
+ "loss": 0.0402,
653
+ "step": 460
654
+ },
655
+ {
656
+ "epoch": 3.141891891891892,
657
+ "grad_norm": 0.4456729161007252,
658
+ "learning_rate": 1.995697075239365e-05,
659
+ "loss": 0.0417,
660
+ "step": 465
661
+ },
662
+ {
663
+ "epoch": 3.175675675675676,
664
+ "grad_norm": 0.5552051849040736,
665
+ "learning_rate": 1.9485242306605028e-05,
666
+ "loss": 0.0368,
667
+ "step": 470
668
+ },
669
+ {
670
+ "epoch": 3.2094594594594597,
671
+ "grad_norm": 0.48465471615450373,
672
+ "learning_rate": 1.9017515163746058e-05,
673
+ "loss": 0.0398,
674
+ "step": 475
675
+ },
676
+ {
677
+ "epoch": 3.2432432432432434,
678
+ "grad_norm": 0.40973628373713084,
679
+ "learning_rate": 1.855402283280836e-05,
680
+ "loss": 0.0395,
681
+ "step": 480
682
+ },
683
+ {
684
+ "epoch": 3.277027027027027,
685
+ "grad_norm": 0.5006185178881579,
686
+ "learning_rate": 1.8094996708587958e-05,
687
+ "loss": 0.0334,
688
+ "step": 485
689
+ },
690
+ {
691
+ "epoch": 3.310810810810811,
692
+ "grad_norm": 0.5022161933000103,
693
+ "learning_rate": 1.7640665956163306e-05,
694
+ "loss": 0.0331,
695
+ "step": 490
696
+ },
697
+ {
698
+ "epoch": 3.3445945945945947,
699
+ "grad_norm": 0.45753568314378656,
700
+ "learning_rate": 1.719125739648648e-05,
701
+ "loss": 0.0317,
702
+ "step": 495
703
+ },
704
+ {
705
+ "epoch": 3.3783783783783785,
706
+ "grad_norm": 0.448289241118566,
707
+ "learning_rate": 1.6746995393144668e-05,
708
+ "loss": 0.0315,
709
+ "step": 500
710
+ },
711
+ {
712
+ "epoch": 3.4121621621621623,
713
+ "grad_norm": 0.8295893160466691,
714
+ "learning_rate": 1.6308101740348433e-05,
715
+ "loss": 0.0386,
716
+ "step": 505
717
+ },
718
+ {
719
+ "epoch": 3.445945945945946,
720
+ "grad_norm": 0.5295192301457442,
721
+ "learning_rate": 1.5874795552202773e-05,
722
+ "loss": 0.0388,
723
+ "step": 510
724
+ },
725
+ {
726
+ "epoch": 3.47972972972973,
727
+ "grad_norm": 0.5769268856553119,
728
+ "learning_rate": 1.5447293153316163e-05,
729
+ "loss": 0.0318,
730
+ "step": 515
731
+ },
732
+ {
733
+ "epoch": 3.5135135135135136,
734
+ "grad_norm": 1.7958739927680984,
735
+ "learning_rate": 1.5025807970802252e-05,
736
+ "loss": 0.0335,
737
+ "step": 520
738
+ },
739
+ {
740
+ "epoch": 3.5472972972972974,
741
+ "grad_norm": 0.5185543173255863,
742
+ "learning_rate": 1.4610550427728103e-05,
743
+ "loss": 0.0323,
744
+ "step": 525
745
+ },
746
+ {
747
+ "epoch": 3.581081081081081,
748
+ "grad_norm": 0.6094438977040642,
749
+ "learning_rate": 1.4201727838062181e-05,
750
+ "loss": 0.0382,
751
+ "step": 530
752
+ },
753
+ {
754
+ "epoch": 3.614864864864865,
755
+ "grad_norm": 0.9983111528521604,
756
+ "learning_rate": 1.3799544303174514e-05,
757
+ "loss": 0.043,
758
+ "step": 535
759
+ },
760
+ {
761
+ "epoch": 3.6486486486486487,
762
+ "grad_norm": 0.5800713261881625,
763
+ "learning_rate": 1.3404200609940754e-05,
764
+ "loss": 0.0357,
765
+ "step": 540
766
+ },
767
+ {
768
+ "epoch": 3.6824324324324325,
769
+ "grad_norm": 0.3822107351668231,
770
+ "learning_rate": 1.3015894130500977e-05,
771
+ "loss": 0.033,
772
+ "step": 545
773
+ },
774
+ {
775
+ "epoch": 3.7162162162162162,
776
+ "grad_norm": 0.5208532366640735,
777
+ "learning_rate": 1.2634818723723174e-05,
778
+ "loss": 0.0336,
779
+ "step": 550
780
+ },
781
+ {
782
+ "epoch": 3.75,
783
+ "grad_norm": 0.6114967345181951,
784
+ "learning_rate": 1.2261164638420832e-05,
785
+ "loss": 0.0311,
786
+ "step": 555
787
+ },
788
+ {
789
+ "epoch": 3.7837837837837838,
790
+ "grad_norm": 0.5437365394783563,
791
+ "learning_rate": 1.1895118418372734e-05,
792
+ "loss": 0.0354,
793
+ "step": 560
794
+ },
795
+ {
796
+ "epoch": 3.8175675675675675,
797
+ "grad_norm": 0.3780362895703916,
798
+ "learning_rate": 1.1536862809192518e-05,
799
+ "loss": 0.0346,
800
+ "step": 565
801
+ },
802
+ {
803
+ "epoch": 3.8513513513513513,
804
+ "grad_norm": 3.518297058582475,
805
+ "learning_rate": 1.1186576667094342e-05,
806
+ "loss": 0.0285,
807
+ "step": 570
808
+ },
809
+ {
810
+ "epoch": 3.885135135135135,
811
+ "grad_norm": 0.3660586029820935,
812
+ "learning_rate": 1.0844434869600428e-05,
813
+ "loss": 0.0312,
814
+ "step": 575
815
+ },
816
+ {
817
+ "epoch": 3.918918918918919,
818
+ "grad_norm": 0.4588190949336144,
819
+ "learning_rate": 1.0510608228234848e-05,
820
+ "loss": 0.0327,
821
+ "step": 580
822
+ },
823
+ {
824
+ "epoch": 3.9527027027027026,
825
+ "grad_norm": 0.46633602361920473,
826
+ "learning_rate": 1.0185263403247256e-05,
827
+ "loss": 0.0344,
828
+ "step": 585
829
+ },
830
+ {
831
+ "epoch": 3.9864864864864864,
832
+ "grad_norm": 0.48024481445354344,
833
+ "learning_rate": 9.868562820409103e-06,
834
+ "loss": 0.0312,
835
+ "step": 590
836
+ },
837
+ {
838
+ "epoch": 4.02027027027027,
839
+ "grad_norm": 0.3049063524706929,
840
+ "learning_rate": 9.560664589923895e-06,
841
+ "loss": 0.0236,
842
+ "step": 595
843
+ },
844
+ {
845
+ "epoch": 4.054054054054054,
846
+ "grad_norm": 0.36714803546431,
847
+ "learning_rate": 9.261722427491953e-06,
848
+ "loss": 0.0149,
849
+ "step": 600
850
+ },
851
+ {
852
+ "epoch": 4.087837837837838,
853
+ "grad_norm": 0.2893986636191623,
854
+ "learning_rate": 8.971885577569058e-06,
855
+ "loss": 0.0125,
856
+ "step": 605
857
+ },
858
+ {
859
+ "epoch": 4.121621621621622,
860
+ "grad_norm": 0.2225194619419826,
861
+ "learning_rate": 8.691298738857432e-06,
862
+ "loss": 0.0105,
863
+ "step": 610
864
+ },
865
+ {
866
+ "epoch": 4.155405405405405,
867
+ "grad_norm": 0.271632458626091,
868
+ "learning_rate": 8.420101992066028e-06,
869
+ "loss": 0.0132,
870
+ "step": 615
871
+ },
872
+ {
873
+ "epoch": 4.1891891891891895,
874
+ "grad_norm": 0.38853806983200195,
875
+ "learning_rate": 8.158430729976372e-06,
876
+ "loss": 0.0117,
877
+ "step": 620
878
+ },
879
+ {
880
+ "epoch": 4.222972972972973,
881
+ "grad_norm": 0.28121579810924724,
882
+ "learning_rate": 7.906415589848834e-06,
883
+ "loss": 0.0119,
884
+ "step": 625
885
+ },
886
+ {
887
+ "epoch": 4.256756756756757,
888
+ "grad_norm": 0.36950427724679225,
889
+ "learning_rate": 7.664182388203037e-06,
890
+ "loss": 0.0122,
891
+ "step": 630
892
+ },
893
+ {
894
+ "epoch": 4.29054054054054,
895
+ "grad_norm": 0.40982200614975045,
896
+ "learning_rate": 7.4318520580049444e-06,
897
+ "loss": 0.0144,
898
+ "step": 635
899
+ },
900
+ {
901
+ "epoch": 4.324324324324325,
902
+ "grad_norm": 0.30667616284588944,
903
+ "learning_rate": 7.209540588292083e-06,
904
+ "loss": 0.0149,
905
+ "step": 640
906
+ },
907
+ {
908
+ "epoch": 4.358108108108108,
909
+ "grad_norm": 0.27670270061009083,
910
+ "learning_rate": 6.9973589662669455e-06,
911
+ "loss": 0.0097,
912
+ "step": 645
913
+ },
914
+ {
915
+ "epoch": 4.391891891891892,
916
+ "grad_norm": 1.156032111625232,
917
+ "learning_rate": 6.7954131218875404e-06,
918
+ "loss": 0.0141,
919
+ "step": 650
920
+ },
921
+ {
922
+ "epoch": 4.425675675675675,
923
+ "grad_norm": 0.43940200076280117,
924
+ "learning_rate": 6.603803874982687e-06,
925
+ "loss": 0.0187,
926
+ "step": 655
927
+ },
928
+ {
929
+ "epoch": 4.45945945945946,
930
+ "grad_norm": 0.5834566768734173,
931
+ "learning_rate": 6.422626884918559e-06,
932
+ "loss": 0.0117,
933
+ "step": 660
934
+ },
935
+ {
936
+ "epoch": 4.493243243243243,
937
+ "grad_norm": 0.38705922881740046,
938
+ "learning_rate": 6.2519726028415145e-06,
939
+ "loss": 0.0138,
940
+ "step": 665
941
+ },
942
+ {
943
+ "epoch": 4.527027027027027,
944
+ "grad_norm": 0.28730538588311555,
945
+ "learning_rate": 6.091926226521089e-06,
946
+ "loss": 0.0127,
947
+ "step": 670
948
+ },
949
+ {
950
+ "epoch": 4.5608108108108105,
951
+ "grad_norm": 0.3749162518326673,
952
+ "learning_rate": 5.942567657815696e-06,
953
+ "loss": 0.012,
954
+ "step": 675
955
+ },
956
+ {
957
+ "epoch": 4.594594594594595,
958
+ "grad_norm": 0.2566159254010731,
959
+ "learning_rate": 5.8039714627822754e-06,
960
+ "loss": 0.0089,
961
+ "step": 680
962
+ },
963
+ {
964
+ "epoch": 4.628378378378378,
965
+ "grad_norm": 0.4339033109119829,
966
+ "learning_rate": 5.676206834449797e-06,
967
+ "loss": 0.0113,
968
+ "step": 685
969
+ },
970
+ {
971
+ "epoch": 4.662162162162162,
972
+ "grad_norm": 0.3802169399499887,
973
+ "learning_rate": 5.55933755827518e-06,
974
+ "loss": 0.0098,
975
+ "step": 690
976
+ },
977
+ {
978
+ "epoch": 4.695945945945946,
979
+ "grad_norm": 0.4733511291475961,
980
+ "learning_rate": 5.453421980298957e-06,
981
+ "loss": 0.0105,
982
+ "step": 695
983
+ },
984
+ {
985
+ "epoch": 4.72972972972973,
986
+ "grad_norm": 0.3236015566148624,
987
+ "learning_rate": 5.358512978016445e-06,
988
+ "loss": 0.0096,
989
+ "step": 700
990
+ },
991
+ {
992
+ "epoch": 4.763513513513513,
993
+ "grad_norm": 0.47392413673934136,
994
+ "learning_rate": 5.27465793397911e-06,
995
+ "loss": 0.0125,
996
+ "step": 705
997
+ },
998
+ {
999
+ "epoch": 4.797297297297297,
1000
+ "grad_norm": 0.24957782619954136,
1001
+ "learning_rate": 5.201898712139201e-06,
1002
+ "loss": 0.01,
1003
+ "step": 710
1004
+ },
1005
+ {
1006
+ "epoch": 4.831081081081081,
1007
+ "grad_norm": 0.23348067557635,
1008
+ "learning_rate": 5.1402716369495194e-06,
1009
+ "loss": 0.0089,
1010
+ "step": 715
1011
+ },
1012
+ {
1013
+ "epoch": 4.864864864864865,
1014
+ "grad_norm": 0.34646806534712726,
1015
+ "learning_rate": 5.089807475228711e-06,
1016
+ "loss": 0.0149,
1017
+ "step": 720
1018
+ },
1019
+ {
1020
+ "epoch": 4.898648648648649,
1021
+ "grad_norm": 0.27931985147442506,
1022
+ "learning_rate": 5.050531420801205e-06,
1023
+ "loss": 0.0098,
1024
+ "step": 725
1025
+ },
1026
+ {
1027
+ "epoch": 4.9324324324324325,
1028
+ "grad_norm": 0.38114606449587424,
1029
+ "learning_rate": 5.022463081919386e-06,
1030
+ "loss": 0.0141,
1031
+ "step": 730
1032
+ },
1033
+ {
1034
+ "epoch": 4.966216216216216,
1035
+ "grad_norm": 0.3191715681389194,
1036
+ "learning_rate": 5.005616471474332e-06,
1037
+ "loss": 0.0106,
1038
+ "step": 735
1039
+ },
1040
+ {
1041
+ "epoch": 5.0,
1042
+ "grad_norm": 0.29441671421186116,
1043
+ "learning_rate": 5e-06,
1044
+ "loss": 0.0107,
1045
+ "step": 740
1046
+ },
1047
+ {
1048
+ "epoch": 5.0,
1049
+ "step": 740,
1050
+ "total_flos": 435868890169344.0,
1051
+ "train_loss": 0.11167457006066232,
1052
+ "train_runtime": 12339.9122,
1053
+ "train_samples_per_second": 3.833,
1054
+ "train_steps_per_second": 0.06
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": 435868890169344.0,
1075
+ "train_batch_size": 8,
1076
+ "trial_name": null,
1077
+ "trial_params": null
1078
+ }