ZMC2019 commited on
Commit
c3a7aed
·
verified ·
1 Parent(s): 73e1dff

Model save

Browse files
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: agentica-org/DeepScaleR-1.5B-Preview
3
+ library_name: transformers
4
+ model_name: DeepScaleR-1.5B-L36-RollE6
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for DeepScaleR-1.5B-L36-RollE6
13
+
14
+ This model is a fine-tuned version of [agentica-org/DeepScaleR-1.5B-Preview](https://huggingface.co/agentica-org/DeepScaleR-1.5B-Preview).
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="ZMC2019/DeepScaleR-1.5B-L36-RollE6", 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/chenzhuoming911/huggingface/runs/a93zynx5)
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": 2.5527473493612954e+17,
3
+ "train_loss": 0.3212856304745714,
4
+ "train_runtime": 7845.5794,
5
+ "train_samples": 93733,
6
+ "train_samples_per_second": 1.451,
7
+ "train_steps_per_second": 0.091
8
+ }
generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 151646,
4
+ "do_sample": true,
5
+ "eos_token_id": 151643,
6
+ "temperature": 0.6,
7
+ "top_p": 0.95,
8
+ "transformers_version": "4.49.0"
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 2.5527473493612954e+17,
3
+ "train_loss": 0.3212856304745714,
4
+ "train_runtime": 7845.5794,
5
+ "train_samples": 93733,
6
+ "train_samples_per_second": 1.451,
7
+ "train_steps_per_second": 0.091
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1036 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 6.0,
5
+ "eval_steps": 500,
6
+ "global_step": 714,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04201680672268908,
13
+ "grad_norm": 7.36784279593653,
14
+ "learning_rate": 3.4722222222222224e-06,
15
+ "loss": 0.9203,
16
+ "step": 5
17
+ },
18
+ {
19
+ "epoch": 0.08403361344537816,
20
+ "grad_norm": 2.6817007224142193,
21
+ "learning_rate": 6.944444444444445e-06,
22
+ "loss": 0.8488,
23
+ "step": 10
24
+ },
25
+ {
26
+ "epoch": 0.12605042016806722,
27
+ "grad_norm": 1.1173297473773625,
28
+ "learning_rate": 1.0416666666666668e-05,
29
+ "loss": 0.6786,
30
+ "step": 15
31
+ },
32
+ {
33
+ "epoch": 0.16806722689075632,
34
+ "grad_norm": 0.5432821597326414,
35
+ "learning_rate": 1.388888888888889e-05,
36
+ "loss": 0.6023,
37
+ "step": 20
38
+ },
39
+ {
40
+ "epoch": 0.21008403361344538,
41
+ "grad_norm": 0.4253948970109724,
42
+ "learning_rate": 1.736111111111111e-05,
43
+ "loss": 0.5601,
44
+ "step": 25
45
+ },
46
+ {
47
+ "epoch": 0.25210084033613445,
48
+ "grad_norm": 0.31150432991851523,
49
+ "learning_rate": 2.0833333333333336e-05,
50
+ "loss": 0.5233,
51
+ "step": 30
52
+ },
53
+ {
54
+ "epoch": 0.29411764705882354,
55
+ "grad_norm": 0.32050134592917545,
56
+ "learning_rate": 2.4305555555555558e-05,
57
+ "loss": 0.506,
58
+ "step": 35
59
+ },
60
+ {
61
+ "epoch": 0.33613445378151263,
62
+ "grad_norm": 0.27149516786646455,
63
+ "learning_rate": 2.777777777777778e-05,
64
+ "loss": 0.4902,
65
+ "step": 40
66
+ },
67
+ {
68
+ "epoch": 0.37815126050420167,
69
+ "grad_norm": 0.256388828476645,
70
+ "learning_rate": 3.125e-05,
71
+ "loss": 0.466,
72
+ "step": 45
73
+ },
74
+ {
75
+ "epoch": 0.42016806722689076,
76
+ "grad_norm": 0.2604693465205412,
77
+ "learning_rate": 3.472222222222222e-05,
78
+ "loss": 0.4627,
79
+ "step": 50
80
+ },
81
+ {
82
+ "epoch": 0.46218487394957986,
83
+ "grad_norm": 0.23787838722566912,
84
+ "learning_rate": 3.8194444444444444e-05,
85
+ "loss": 0.443,
86
+ "step": 55
87
+ },
88
+ {
89
+ "epoch": 0.5042016806722689,
90
+ "grad_norm": 0.23524988298322033,
91
+ "learning_rate": 4.166666666666667e-05,
92
+ "loss": 0.4347,
93
+ "step": 60
94
+ },
95
+ {
96
+ "epoch": 0.5462184873949579,
97
+ "grad_norm": 0.720408504919723,
98
+ "learning_rate": 4.5138888888888894e-05,
99
+ "loss": 0.4325,
100
+ "step": 65
101
+ },
102
+ {
103
+ "epoch": 0.5882352941176471,
104
+ "grad_norm": 0.15502732828730537,
105
+ "learning_rate": 4.8611111111111115e-05,
106
+ "loss": 0.4092,
107
+ "step": 70
108
+ },
109
+ {
110
+ "epoch": 0.6302521008403361,
111
+ "grad_norm": 0.1681572620191539,
112
+ "learning_rate": 4.976635514018692e-05,
113
+ "loss": 0.4144,
114
+ "step": 75
115
+ },
116
+ {
117
+ "epoch": 0.6722689075630253,
118
+ "grad_norm": 0.14289664425952017,
119
+ "learning_rate": 4.937694704049845e-05,
120
+ "loss": 0.4101,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.7142857142857143,
125
+ "grad_norm": 0.1423836892341792,
126
+ "learning_rate": 4.898753894080997e-05,
127
+ "loss": 0.4028,
128
+ "step": 85
129
+ },
130
+ {
131
+ "epoch": 0.7563025210084033,
132
+ "grad_norm": 0.13921521879920523,
133
+ "learning_rate": 4.85981308411215e-05,
134
+ "loss": 0.3961,
135
+ "step": 90
136
+ },
137
+ {
138
+ "epoch": 0.7983193277310925,
139
+ "grad_norm": 0.1320808306944742,
140
+ "learning_rate": 4.820872274143303e-05,
141
+ "loss": 0.4057,
142
+ "step": 95
143
+ },
144
+ {
145
+ "epoch": 0.8403361344537815,
146
+ "grad_norm": 0.13511566705435527,
147
+ "learning_rate": 4.781931464174455e-05,
148
+ "loss": 0.3982,
149
+ "step": 100
150
+ },
151
+ {
152
+ "epoch": 0.8823529411764706,
153
+ "grad_norm": 0.12059831677330138,
154
+ "learning_rate": 4.742990654205608e-05,
155
+ "loss": 0.3926,
156
+ "step": 105
157
+ },
158
+ {
159
+ "epoch": 0.9243697478991597,
160
+ "grad_norm": 0.14025999739783343,
161
+ "learning_rate": 4.7040498442367604e-05,
162
+ "loss": 0.3863,
163
+ "step": 110
164
+ },
165
+ {
166
+ "epoch": 0.9663865546218487,
167
+ "grad_norm": 0.13908375301311737,
168
+ "learning_rate": 4.665109034267913e-05,
169
+ "loss": 0.3985,
170
+ "step": 115
171
+ },
172
+ {
173
+ "epoch": 1.0084033613445378,
174
+ "grad_norm": 0.1402995196474956,
175
+ "learning_rate": 4.6261682242990654e-05,
176
+ "loss": 0.3844,
177
+ "step": 120
178
+ },
179
+ {
180
+ "epoch": 1.050420168067227,
181
+ "grad_norm": 0.14896376645944973,
182
+ "learning_rate": 4.5872274143302186e-05,
183
+ "loss": 0.3711,
184
+ "step": 125
185
+ },
186
+ {
187
+ "epoch": 1.092436974789916,
188
+ "grad_norm": 0.1330752274361008,
189
+ "learning_rate": 4.548286604361371e-05,
190
+ "loss": 0.3627,
191
+ "step": 130
192
+ },
193
+ {
194
+ "epoch": 1.134453781512605,
195
+ "grad_norm": 0.12286078605816185,
196
+ "learning_rate": 4.5093457943925236e-05,
197
+ "loss": 0.3627,
198
+ "step": 135
199
+ },
200
+ {
201
+ "epoch": 1.1764705882352942,
202
+ "grad_norm": 0.1300040241362218,
203
+ "learning_rate": 4.470404984423676e-05,
204
+ "loss": 0.3586,
205
+ "step": 140
206
+ },
207
+ {
208
+ "epoch": 1.2184873949579833,
209
+ "grad_norm": 0.1212676350760631,
210
+ "learning_rate": 4.431464174454829e-05,
211
+ "loss": 0.3653,
212
+ "step": 145
213
+ },
214
+ {
215
+ "epoch": 1.2605042016806722,
216
+ "grad_norm": 0.13672573392631662,
217
+ "learning_rate": 4.392523364485982e-05,
218
+ "loss": 0.3601,
219
+ "step": 150
220
+ },
221
+ {
222
+ "epoch": 1.3025210084033614,
223
+ "grad_norm": 0.12866697257558982,
224
+ "learning_rate": 4.353582554517134e-05,
225
+ "loss": 0.3646,
226
+ "step": 155
227
+ },
228
+ {
229
+ "epoch": 1.3445378151260505,
230
+ "grad_norm": 0.12251791003688832,
231
+ "learning_rate": 4.314641744548287e-05,
232
+ "loss": 0.3554,
233
+ "step": 160
234
+ },
235
+ {
236
+ "epoch": 1.3865546218487395,
237
+ "grad_norm": 0.128140623459501,
238
+ "learning_rate": 4.27570093457944e-05,
239
+ "loss": 0.3597,
240
+ "step": 165
241
+ },
242
+ {
243
+ "epoch": 1.4285714285714286,
244
+ "grad_norm": 0.12517942956515188,
245
+ "learning_rate": 4.236760124610592e-05,
246
+ "loss": 0.3596,
247
+ "step": 170
248
+ },
249
+ {
250
+ "epoch": 1.4705882352941178,
251
+ "grad_norm": 0.1264068992533434,
252
+ "learning_rate": 4.197819314641745e-05,
253
+ "loss": 0.3523,
254
+ "step": 175
255
+ },
256
+ {
257
+ "epoch": 1.5126050420168067,
258
+ "grad_norm": 0.12844150586086567,
259
+ "learning_rate": 4.1588785046728974e-05,
260
+ "loss": 0.3611,
261
+ "step": 180
262
+ },
263
+ {
264
+ "epoch": 1.5546218487394958,
265
+ "grad_norm": 0.1463320479548455,
266
+ "learning_rate": 4.11993769470405e-05,
267
+ "loss": 0.3544,
268
+ "step": 185
269
+ },
270
+ {
271
+ "epoch": 1.596638655462185,
272
+ "grad_norm": 0.1387410614743581,
273
+ "learning_rate": 4.0809968847352024e-05,
274
+ "loss": 0.3506,
275
+ "step": 190
276
+ },
277
+ {
278
+ "epoch": 1.638655462184874,
279
+ "grad_norm": 0.11966374174214205,
280
+ "learning_rate": 4.0420560747663556e-05,
281
+ "loss": 0.3547,
282
+ "step": 195
283
+ },
284
+ {
285
+ "epoch": 1.680672268907563,
286
+ "grad_norm": 0.13041047515525458,
287
+ "learning_rate": 4.003115264797508e-05,
288
+ "loss": 0.3551,
289
+ "step": 200
290
+ },
291
+ {
292
+ "epoch": 1.7226890756302522,
293
+ "grad_norm": 0.4175865240152316,
294
+ "learning_rate": 3.9641744548286606e-05,
295
+ "loss": 0.359,
296
+ "step": 205
297
+ },
298
+ {
299
+ "epoch": 1.7647058823529411,
300
+ "grad_norm": 0.12341307943053885,
301
+ "learning_rate": 3.925233644859813e-05,
302
+ "loss": 0.3509,
303
+ "step": 210
304
+ },
305
+ {
306
+ "epoch": 1.8067226890756303,
307
+ "grad_norm": 0.11938805910700397,
308
+ "learning_rate": 3.886292834890966e-05,
309
+ "loss": 0.353,
310
+ "step": 215
311
+ },
312
+ {
313
+ "epoch": 1.8487394957983194,
314
+ "grad_norm": 0.11924644124050648,
315
+ "learning_rate": 3.847352024922119e-05,
316
+ "loss": 0.3543,
317
+ "step": 220
318
+ },
319
+ {
320
+ "epoch": 1.8907563025210083,
321
+ "grad_norm": 0.1192046521417864,
322
+ "learning_rate": 3.808411214953271e-05,
323
+ "loss": 0.3459,
324
+ "step": 225
325
+ },
326
+ {
327
+ "epoch": 1.9327731092436975,
328
+ "grad_norm": 0.11198889670651223,
329
+ "learning_rate": 3.769470404984424e-05,
330
+ "loss": 0.3526,
331
+ "step": 230
332
+ },
333
+ {
334
+ "epoch": 1.9747899159663866,
335
+ "grad_norm": 0.11064392419307578,
336
+ "learning_rate": 3.730529595015577e-05,
337
+ "loss": 0.3481,
338
+ "step": 235
339
+ },
340
+ {
341
+ "epoch": 2.0168067226890756,
342
+ "grad_norm": 0.14955656421111635,
343
+ "learning_rate": 3.691588785046729e-05,
344
+ "loss": 0.334,
345
+ "step": 240
346
+ },
347
+ {
348
+ "epoch": 2.0588235294117645,
349
+ "grad_norm": 0.1352971990563673,
350
+ "learning_rate": 3.652647975077882e-05,
351
+ "loss": 0.3192,
352
+ "step": 245
353
+ },
354
+ {
355
+ "epoch": 2.100840336134454,
356
+ "grad_norm": 0.16599609164840282,
357
+ "learning_rate": 3.6137071651090344e-05,
358
+ "loss": 0.3164,
359
+ "step": 250
360
+ },
361
+ {
362
+ "epoch": 2.142857142857143,
363
+ "grad_norm": 0.13656518038762977,
364
+ "learning_rate": 3.574766355140187e-05,
365
+ "loss": 0.3124,
366
+ "step": 255
367
+ },
368
+ {
369
+ "epoch": 2.184873949579832,
370
+ "grad_norm": 0.13474455104744834,
371
+ "learning_rate": 3.5358255451713394e-05,
372
+ "loss": 0.3119,
373
+ "step": 260
374
+ },
375
+ {
376
+ "epoch": 2.226890756302521,
377
+ "grad_norm": 0.12821245064355621,
378
+ "learning_rate": 3.4968847352024926e-05,
379
+ "loss": 0.3168,
380
+ "step": 265
381
+ },
382
+ {
383
+ "epoch": 2.26890756302521,
384
+ "grad_norm": 0.13307269371542046,
385
+ "learning_rate": 3.457943925233645e-05,
386
+ "loss": 0.3192,
387
+ "step": 270
388
+ },
389
+ {
390
+ "epoch": 2.310924369747899,
391
+ "grad_norm": 0.14021981893226773,
392
+ "learning_rate": 3.4190031152647976e-05,
393
+ "loss": 0.3159,
394
+ "step": 275
395
+ },
396
+ {
397
+ "epoch": 2.3529411764705883,
398
+ "grad_norm": 0.13437639361334078,
399
+ "learning_rate": 3.38006230529595e-05,
400
+ "loss": 0.3123,
401
+ "step": 280
402
+ },
403
+ {
404
+ "epoch": 2.3949579831932772,
405
+ "grad_norm": 0.1326899305498554,
406
+ "learning_rate": 3.341121495327103e-05,
407
+ "loss": 0.3177,
408
+ "step": 285
409
+ },
410
+ {
411
+ "epoch": 2.4369747899159666,
412
+ "grad_norm": 0.12293267360959077,
413
+ "learning_rate": 3.302180685358255e-05,
414
+ "loss": 0.3159,
415
+ "step": 290
416
+ },
417
+ {
418
+ "epoch": 2.4789915966386555,
419
+ "grad_norm": 0.12467837886737294,
420
+ "learning_rate": 3.263239875389408e-05,
421
+ "loss": 0.3152,
422
+ "step": 295
423
+ },
424
+ {
425
+ "epoch": 2.5210084033613445,
426
+ "grad_norm": 0.1245503385101852,
427
+ "learning_rate": 3.224299065420561e-05,
428
+ "loss": 0.3172,
429
+ "step": 300
430
+ },
431
+ {
432
+ "epoch": 2.5630252100840334,
433
+ "grad_norm": 0.12135305650404841,
434
+ "learning_rate": 3.185358255451714e-05,
435
+ "loss": 0.3131,
436
+ "step": 305
437
+ },
438
+ {
439
+ "epoch": 2.6050420168067228,
440
+ "grad_norm": 0.1196869916735729,
441
+ "learning_rate": 3.146417445482866e-05,
442
+ "loss": 0.3131,
443
+ "step": 310
444
+ },
445
+ {
446
+ "epoch": 2.6470588235294117,
447
+ "grad_norm": 0.1422717014600059,
448
+ "learning_rate": 3.107476635514019e-05,
449
+ "loss": 0.3122,
450
+ "step": 315
451
+ },
452
+ {
453
+ "epoch": 2.689075630252101,
454
+ "grad_norm": 0.12049177418433844,
455
+ "learning_rate": 3.0685358255451714e-05,
456
+ "loss": 0.3118,
457
+ "step": 320
458
+ },
459
+ {
460
+ "epoch": 2.73109243697479,
461
+ "grad_norm": 0.1218866413247397,
462
+ "learning_rate": 3.0295950155763243e-05,
463
+ "loss": 0.3093,
464
+ "step": 325
465
+ },
466
+ {
467
+ "epoch": 2.773109243697479,
468
+ "grad_norm": 0.12128404985597208,
469
+ "learning_rate": 2.9906542056074764e-05,
470
+ "loss": 0.3149,
471
+ "step": 330
472
+ },
473
+ {
474
+ "epoch": 2.815126050420168,
475
+ "grad_norm": 0.12165521196292105,
476
+ "learning_rate": 2.9517133956386296e-05,
477
+ "loss": 0.3152,
478
+ "step": 335
479
+ },
480
+ {
481
+ "epoch": 2.857142857142857,
482
+ "grad_norm": 0.11968176603091883,
483
+ "learning_rate": 2.9127725856697818e-05,
484
+ "loss": 0.3065,
485
+ "step": 340
486
+ },
487
+ {
488
+ "epoch": 2.899159663865546,
489
+ "grad_norm": 0.13174397154227724,
490
+ "learning_rate": 2.873831775700935e-05,
491
+ "loss": 0.3161,
492
+ "step": 345
493
+ },
494
+ {
495
+ "epoch": 2.9411764705882355,
496
+ "grad_norm": 0.12078966649058866,
497
+ "learning_rate": 2.834890965732087e-05,
498
+ "loss": 0.3143,
499
+ "step": 350
500
+ },
501
+ {
502
+ "epoch": 2.9831932773109244,
503
+ "grad_norm": 0.12196146741006934,
504
+ "learning_rate": 2.79595015576324e-05,
505
+ "loss": 0.32,
506
+ "step": 355
507
+ },
508
+ {
509
+ "epoch": 3.0252100840336134,
510
+ "grad_norm": 0.22204914838949716,
511
+ "learning_rate": 2.7570093457943924e-05,
512
+ "loss": 0.2892,
513
+ "step": 360
514
+ },
515
+ {
516
+ "epoch": 3.0672268907563027,
517
+ "grad_norm": 0.17224639019450644,
518
+ "learning_rate": 2.7180685358255453e-05,
519
+ "loss": 0.2803,
520
+ "step": 365
521
+ },
522
+ {
523
+ "epoch": 3.1092436974789917,
524
+ "grad_norm": 0.12584678266855343,
525
+ "learning_rate": 2.6791277258566978e-05,
526
+ "loss": 0.2762,
527
+ "step": 370
528
+ },
529
+ {
530
+ "epoch": 3.1512605042016806,
531
+ "grad_norm": 0.14018543572816114,
532
+ "learning_rate": 2.6401869158878506e-05,
533
+ "loss": 0.2791,
534
+ "step": 375
535
+ },
536
+ {
537
+ "epoch": 3.19327731092437,
538
+ "grad_norm": 0.12941623347853976,
539
+ "learning_rate": 2.601246105919003e-05,
540
+ "loss": 0.2749,
541
+ "step": 380
542
+ },
543
+ {
544
+ "epoch": 3.235294117647059,
545
+ "grad_norm": 0.12194288998302227,
546
+ "learning_rate": 2.562305295950156e-05,
547
+ "loss": 0.2803,
548
+ "step": 385
549
+ },
550
+ {
551
+ "epoch": 3.277310924369748,
552
+ "grad_norm": 0.12446866234019027,
553
+ "learning_rate": 2.5233644859813084e-05,
554
+ "loss": 0.2805,
555
+ "step": 390
556
+ },
557
+ {
558
+ "epoch": 3.3193277310924367,
559
+ "grad_norm": 0.12351722292105845,
560
+ "learning_rate": 2.4844236760124613e-05,
561
+ "loss": 0.2771,
562
+ "step": 395
563
+ },
564
+ {
565
+ "epoch": 3.361344537815126,
566
+ "grad_norm": 0.12653348786860177,
567
+ "learning_rate": 2.4454828660436138e-05,
568
+ "loss": 0.2781,
569
+ "step": 400
570
+ },
571
+ {
572
+ "epoch": 3.403361344537815,
573
+ "grad_norm": 0.13084716393704046,
574
+ "learning_rate": 2.4065420560747666e-05,
575
+ "loss": 0.2782,
576
+ "step": 405
577
+ },
578
+ {
579
+ "epoch": 3.4453781512605044,
580
+ "grad_norm": 0.12386845332574656,
581
+ "learning_rate": 2.367601246105919e-05,
582
+ "loss": 0.2789,
583
+ "step": 410
584
+ },
585
+ {
586
+ "epoch": 3.4873949579831933,
587
+ "grad_norm": 0.12966375682019396,
588
+ "learning_rate": 2.3286604361370716e-05,
589
+ "loss": 0.2859,
590
+ "step": 415
591
+ },
592
+ {
593
+ "epoch": 3.5294117647058822,
594
+ "grad_norm": 0.1244203255643183,
595
+ "learning_rate": 2.2897196261682244e-05,
596
+ "loss": 0.2808,
597
+ "step": 420
598
+ },
599
+ {
600
+ "epoch": 3.571428571428571,
601
+ "grad_norm": 0.12741940430874554,
602
+ "learning_rate": 2.250778816199377e-05,
603
+ "loss": 0.2787,
604
+ "step": 425
605
+ },
606
+ {
607
+ "epoch": 3.6134453781512605,
608
+ "grad_norm": 0.13192944689673486,
609
+ "learning_rate": 2.2118380062305298e-05,
610
+ "loss": 0.2844,
611
+ "step": 430
612
+ },
613
+ {
614
+ "epoch": 3.6554621848739495,
615
+ "grad_norm": 0.12331528955859135,
616
+ "learning_rate": 2.1728971962616822e-05,
617
+ "loss": 0.2825,
618
+ "step": 435
619
+ },
620
+ {
621
+ "epoch": 3.697478991596639,
622
+ "grad_norm": 0.13313354832203048,
623
+ "learning_rate": 2.133956386292835e-05,
624
+ "loss": 0.2816,
625
+ "step": 440
626
+ },
627
+ {
628
+ "epoch": 3.7394957983193278,
629
+ "grad_norm": 0.13549454769327543,
630
+ "learning_rate": 2.0950155763239876e-05,
631
+ "loss": 0.2817,
632
+ "step": 445
633
+ },
634
+ {
635
+ "epoch": 3.7815126050420167,
636
+ "grad_norm": 0.11803733674337212,
637
+ "learning_rate": 2.05607476635514e-05,
638
+ "loss": 0.28,
639
+ "step": 450
640
+ },
641
+ {
642
+ "epoch": 3.8235294117647056,
643
+ "grad_norm": 0.12012685963323684,
644
+ "learning_rate": 2.017133956386293e-05,
645
+ "loss": 0.2832,
646
+ "step": 455
647
+ },
648
+ {
649
+ "epoch": 3.865546218487395,
650
+ "grad_norm": 0.11875060997630844,
651
+ "learning_rate": 1.9781931464174454e-05,
652
+ "loss": 0.2806,
653
+ "step": 460
654
+ },
655
+ {
656
+ "epoch": 3.907563025210084,
657
+ "grad_norm": 0.13503311291347644,
658
+ "learning_rate": 1.9392523364485982e-05,
659
+ "loss": 0.2857,
660
+ "step": 465
661
+ },
662
+ {
663
+ "epoch": 3.9495798319327733,
664
+ "grad_norm": 0.13899315092229172,
665
+ "learning_rate": 1.9003115264797507e-05,
666
+ "loss": 0.2819,
667
+ "step": 470
668
+ },
669
+ {
670
+ "epoch": 3.991596638655462,
671
+ "grad_norm": 0.12553466656260837,
672
+ "learning_rate": 1.8613707165109036e-05,
673
+ "loss": 0.2781,
674
+ "step": 475
675
+ },
676
+ {
677
+ "epoch": 4.033613445378151,
678
+ "grad_norm": 0.2047499500117637,
679
+ "learning_rate": 1.822429906542056e-05,
680
+ "loss": 0.2559,
681
+ "step": 480
682
+ },
683
+ {
684
+ "epoch": 4.07563025210084,
685
+ "grad_norm": 0.16423755258674735,
686
+ "learning_rate": 1.7834890965732086e-05,
687
+ "loss": 0.2558,
688
+ "step": 485
689
+ },
690
+ {
691
+ "epoch": 4.117647058823529,
692
+ "grad_norm": 0.1348676429167651,
693
+ "learning_rate": 1.7445482866043614e-05,
694
+ "loss": 0.2527,
695
+ "step": 490
696
+ },
697
+ {
698
+ "epoch": 4.159663865546219,
699
+ "grad_norm": 0.14533312788167388,
700
+ "learning_rate": 1.705607476635514e-05,
701
+ "loss": 0.251,
702
+ "step": 495
703
+ },
704
+ {
705
+ "epoch": 4.201680672268908,
706
+ "grad_norm": 0.12637560985053983,
707
+ "learning_rate": 1.6666666666666667e-05,
708
+ "loss": 0.2519,
709
+ "step": 500
710
+ },
711
+ {
712
+ "epoch": 4.243697478991597,
713
+ "grad_norm": 0.12843775741491584,
714
+ "learning_rate": 1.6277258566978192e-05,
715
+ "loss": 0.2497,
716
+ "step": 505
717
+ },
718
+ {
719
+ "epoch": 4.285714285714286,
720
+ "grad_norm": 0.12903589040675842,
721
+ "learning_rate": 1.588785046728972e-05,
722
+ "loss": 0.2493,
723
+ "step": 510
724
+ },
725
+ {
726
+ "epoch": 4.3277310924369745,
727
+ "grad_norm": 0.1265086519312201,
728
+ "learning_rate": 1.5498442367601246e-05,
729
+ "loss": 0.2478,
730
+ "step": 515
731
+ },
732
+ {
733
+ "epoch": 4.369747899159664,
734
+ "grad_norm": 0.12474752965681195,
735
+ "learning_rate": 1.5109034267912772e-05,
736
+ "loss": 0.2552,
737
+ "step": 520
738
+ },
739
+ {
740
+ "epoch": 4.411764705882353,
741
+ "grad_norm": 0.12276341594269388,
742
+ "learning_rate": 1.4719626168224299e-05,
743
+ "loss": 0.2566,
744
+ "step": 525
745
+ },
746
+ {
747
+ "epoch": 4.453781512605042,
748
+ "grad_norm": 0.12799056315008822,
749
+ "learning_rate": 1.4330218068535826e-05,
750
+ "loss": 0.2544,
751
+ "step": 530
752
+ },
753
+ {
754
+ "epoch": 4.495798319327731,
755
+ "grad_norm": 0.1235959721069884,
756
+ "learning_rate": 1.394080996884735e-05,
757
+ "loss": 0.2573,
758
+ "step": 535
759
+ },
760
+ {
761
+ "epoch": 4.53781512605042,
762
+ "grad_norm": 0.12737316017685743,
763
+ "learning_rate": 1.3551401869158877e-05,
764
+ "loss": 0.251,
765
+ "step": 540
766
+ },
767
+ {
768
+ "epoch": 4.579831932773109,
769
+ "grad_norm": 0.12455494002416401,
770
+ "learning_rate": 1.3161993769470404e-05,
771
+ "loss": 0.2472,
772
+ "step": 545
773
+ },
774
+ {
775
+ "epoch": 4.621848739495798,
776
+ "grad_norm": 0.12200067774607738,
777
+ "learning_rate": 1.277258566978193e-05,
778
+ "loss": 0.2512,
779
+ "step": 550
780
+ },
781
+ {
782
+ "epoch": 4.663865546218488,
783
+ "grad_norm": 0.14205343587283586,
784
+ "learning_rate": 1.2383177570093459e-05,
785
+ "loss": 0.2518,
786
+ "step": 555
787
+ },
788
+ {
789
+ "epoch": 4.705882352941177,
790
+ "grad_norm": 0.12216097165620664,
791
+ "learning_rate": 1.1993769470404986e-05,
792
+ "loss": 0.252,
793
+ "step": 560
794
+ },
795
+ {
796
+ "epoch": 4.7478991596638656,
797
+ "grad_norm": 0.12513180450823716,
798
+ "learning_rate": 1.160436137071651e-05,
799
+ "loss": 0.2585,
800
+ "step": 565
801
+ },
802
+ {
803
+ "epoch": 4.7899159663865545,
804
+ "grad_norm": 0.12156552612032653,
805
+ "learning_rate": 1.1214953271028037e-05,
806
+ "loss": 0.2543,
807
+ "step": 570
808
+ },
809
+ {
810
+ "epoch": 4.831932773109243,
811
+ "grad_norm": 0.12936785318700886,
812
+ "learning_rate": 1.0825545171339564e-05,
813
+ "loss": 0.2511,
814
+ "step": 575
815
+ },
816
+ {
817
+ "epoch": 4.873949579831933,
818
+ "grad_norm": 0.12602274784345757,
819
+ "learning_rate": 1.043613707165109e-05,
820
+ "loss": 0.2514,
821
+ "step": 580
822
+ },
823
+ {
824
+ "epoch": 4.915966386554622,
825
+ "grad_norm": 0.16613194488301022,
826
+ "learning_rate": 1.0046728971962617e-05,
827
+ "loss": 0.2507,
828
+ "step": 585
829
+ },
830
+ {
831
+ "epoch": 4.957983193277311,
832
+ "grad_norm": 0.12043993253574951,
833
+ "learning_rate": 9.657320872274144e-06,
834
+ "loss": 0.2524,
835
+ "step": 590
836
+ },
837
+ {
838
+ "epoch": 5.0,
839
+ "grad_norm": 0.1359154617198902,
840
+ "learning_rate": 9.267912772585669e-06,
841
+ "loss": 0.2525,
842
+ "step": 595
843
+ },
844
+ {
845
+ "epoch": 5.042016806722689,
846
+ "grad_norm": 0.1645526234919548,
847
+ "learning_rate": 8.878504672897196e-06,
848
+ "loss": 0.2349,
849
+ "step": 600
850
+ },
851
+ {
852
+ "epoch": 5.084033613445378,
853
+ "grad_norm": 0.13565331505255587,
854
+ "learning_rate": 8.489096573208722e-06,
855
+ "loss": 0.226,
856
+ "step": 605
857
+ },
858
+ {
859
+ "epoch": 5.126050420168067,
860
+ "grad_norm": 0.12112957276832616,
861
+ "learning_rate": 8.099688473520249e-06,
862
+ "loss": 0.2323,
863
+ "step": 610
864
+ },
865
+ {
866
+ "epoch": 5.168067226890757,
867
+ "grad_norm": 0.13228103496620836,
868
+ "learning_rate": 7.710280373831776e-06,
869
+ "loss": 0.2333,
870
+ "step": 615
871
+ },
872
+ {
873
+ "epoch": 5.2100840336134455,
874
+ "grad_norm": 0.12755621341157677,
875
+ "learning_rate": 7.3208722741433015e-06,
876
+ "loss": 0.2281,
877
+ "step": 620
878
+ },
879
+ {
880
+ "epoch": 5.2521008403361344,
881
+ "grad_norm": 0.13414945242875928,
882
+ "learning_rate": 6.931464174454828e-06,
883
+ "loss": 0.2321,
884
+ "step": 625
885
+ },
886
+ {
887
+ "epoch": 5.294117647058823,
888
+ "grad_norm": 0.12052434335169325,
889
+ "learning_rate": 6.542056074766355e-06,
890
+ "loss": 0.2349,
891
+ "step": 630
892
+ },
893
+ {
894
+ "epoch": 5.336134453781512,
895
+ "grad_norm": 0.1251627138575762,
896
+ "learning_rate": 6.1526479750778815e-06,
897
+ "loss": 0.2343,
898
+ "step": 635
899
+ },
900
+ {
901
+ "epoch": 5.378151260504202,
902
+ "grad_norm": 0.1269368352207133,
903
+ "learning_rate": 5.763239875389408e-06,
904
+ "loss": 0.2336,
905
+ "step": 640
906
+ },
907
+ {
908
+ "epoch": 5.420168067226891,
909
+ "grad_norm": 0.12304767267974673,
910
+ "learning_rate": 5.373831775700935e-06,
911
+ "loss": 0.2279,
912
+ "step": 645
913
+ },
914
+ {
915
+ "epoch": 5.46218487394958,
916
+ "grad_norm": 0.12126938306506918,
917
+ "learning_rate": 4.9844236760124615e-06,
918
+ "loss": 0.2325,
919
+ "step": 650
920
+ },
921
+ {
922
+ "epoch": 5.504201680672269,
923
+ "grad_norm": 0.12430095329916237,
924
+ "learning_rate": 4.595015576323987e-06,
925
+ "loss": 0.2308,
926
+ "step": 655
927
+ },
928
+ {
929
+ "epoch": 5.546218487394958,
930
+ "grad_norm": 0.12206854649940266,
931
+ "learning_rate": 4.205607476635514e-06,
932
+ "loss": 0.2303,
933
+ "step": 660
934
+ },
935
+ {
936
+ "epoch": 5.588235294117647,
937
+ "grad_norm": 0.12336264199284398,
938
+ "learning_rate": 3.816199376947041e-06,
939
+ "loss": 0.2339,
940
+ "step": 665
941
+ },
942
+ {
943
+ "epoch": 5.630252100840336,
944
+ "grad_norm": 0.11965482655647582,
945
+ "learning_rate": 3.426791277258567e-06,
946
+ "loss": 0.2303,
947
+ "step": 670
948
+ },
949
+ {
950
+ "epoch": 5.6722689075630255,
951
+ "grad_norm": 0.11594681549175202,
952
+ "learning_rate": 3.0373831775700936e-06,
953
+ "loss": 0.2288,
954
+ "step": 675
955
+ },
956
+ {
957
+ "epoch": 5.714285714285714,
958
+ "grad_norm": 0.1194663204669251,
959
+ "learning_rate": 2.64797507788162e-06,
960
+ "loss": 0.2334,
961
+ "step": 680
962
+ },
963
+ {
964
+ "epoch": 5.756302521008403,
965
+ "grad_norm": 0.11767398574784753,
966
+ "learning_rate": 2.2585669781931465e-06,
967
+ "loss": 0.231,
968
+ "step": 685
969
+ },
970
+ {
971
+ "epoch": 5.798319327731092,
972
+ "grad_norm": 0.11390297708321843,
973
+ "learning_rate": 1.8691588785046728e-06,
974
+ "loss": 0.2344,
975
+ "step": 690
976
+ },
977
+ {
978
+ "epoch": 5.840336134453781,
979
+ "grad_norm": 0.12442153502768896,
980
+ "learning_rate": 1.4797507788161994e-06,
981
+ "loss": 0.2334,
982
+ "step": 695
983
+ },
984
+ {
985
+ "epoch": 5.882352941176471,
986
+ "grad_norm": 0.11827825723525778,
987
+ "learning_rate": 1.0903426791277259e-06,
988
+ "loss": 0.2334,
989
+ "step": 700
990
+ },
991
+ {
992
+ "epoch": 5.92436974789916,
993
+ "grad_norm": 0.12132650089651566,
994
+ "learning_rate": 7.009345794392523e-07,
995
+ "loss": 0.2351,
996
+ "step": 705
997
+ },
998
+ {
999
+ "epoch": 5.966386554621849,
1000
+ "grad_norm": 0.11532246430168183,
1001
+ "learning_rate": 3.1152647975077885e-07,
1002
+ "loss": 0.2337,
1003
+ "step": 710
1004
+ },
1005
+ {
1006
+ "epoch": 6.0,
1007
+ "step": 714,
1008
+ "total_flos": 2.5527473493612954e+17,
1009
+ "train_loss": 0.3212856304745714,
1010
+ "train_runtime": 7845.5794,
1011
+ "train_samples_per_second": 1.451,
1012
+ "train_steps_per_second": 0.091
1013
+ }
1014
+ ],
1015
+ "logging_steps": 5,
1016
+ "max_steps": 714,
1017
+ "num_input_tokens_seen": 0,
1018
+ "num_train_epochs": 6,
1019
+ "save_steps": 500,
1020
+ "stateful_callbacks": {
1021
+ "TrainerControl": {
1022
+ "args": {
1023
+ "should_epoch_stop": false,
1024
+ "should_evaluate": false,
1025
+ "should_log": false,
1026
+ "should_save": true,
1027
+ "should_training_stop": true
1028
+ },
1029
+ "attributes": {}
1030
+ }
1031
+ },
1032
+ "total_flos": 2.5527473493612954e+17,
1033
+ "train_batch_size": 1,
1034
+ "trial_name": null,
1035
+ "trial_params": null
1036
+ }