Hugging Face's logo

movefast
/
Qwen2.5-7B-Open-R1-Distill

Text Generation
Transformers
Safetensors
qwen2
Generated from Trainer
open-r1
trl
sft
conversational
text-generation-inference
Model card Files
xet
 Community
1
movefast commited on
Commit
14cc521
·
verified ·
1 Parent(s): b7ff3db

Model save

Browse files
Files changed (5) hide show
  1. README.md +58 -0
  2. all_results.json +8 -0
  3. generation_config.json +14 -0
  4. train_results.json +8 -0
  5. trainer_state.json +468 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-7B-Instruct
3
+ library_name: transformers
4
+ model_name: Qwen2.5-7B-Open-R1-Distill
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for Qwen2.5-7B-Open-R1-Distill
13
+
14
+ This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-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="movefast/Qwen2.5-7B-Open-R1-Distill", 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/liderek269/huggingface/runs/676qe9j2)
31
+
32
+
33
+ This model was trained with SFT.
34
+
35
+ ### Framework versions
36
+
37
+ - TRL: 0.16.0
38
+ - Transformers: 4.50.0
39
+ - Pytorch: 2.5.1
40
+ - Datasets: 3.3.2
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": 1.379511412069499e+18,
3
+ "train_loss": 0.45678649598093174,
4
+ "train_runtime": 12057.0545,
5
+ "train_samples": 93733,
6
+ "train_samples_per_second": 2.845,
7
+ "train_steps_per_second": 0.022
8
+ }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.50.0"
14
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 1.379511412069499e+18,
3
+ "train_loss": 0.45678649598093174,
4
+ "train_runtime": 12057.0545,
5
+ "train_samples": 93733,
6
+ "train_samples_per_second": 2.845,
7
+ "train_steps_per_second": 0.022
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,468 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9990680335507922,
6
+ "eval_steps": 500,
7
+ "global_step": 268,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.01863932898415657,
14
+ "grad_norm": 4.758676978183845,
15
+ "learning_rate": 1.785714285714286e-05,
16
+ "loss": 0.787,
17
+ "num_tokens": 10465392.0,
18
+ "step": 5
19
+ },
20
+ {
21
+ "epoch": 0.03727865796831314,
22
+ "grad_norm": 1.6688141341980978,
23
+ "learning_rate": 3.571428571428572e-05,
24
+ "loss": 0.6818,
25
+ "num_tokens": 20951152.0,
26
+ "step": 10
27
+ },
28
+ {
29
+ "epoch": 0.05591798695246971,
30
+ "grad_norm": 1.5315485414502277,
31
+ "learning_rate": 4.999827900623038e-05,
32
+ "loss": 0.5877,
33
+ "num_tokens": 31436912.0,
34
+ "step": 15
35
+ },
36
+ {
37
+ "epoch": 0.07455731593662628,
38
+ "grad_norm": 1.1907406620730119,
39
+ "learning_rate": 4.993807186343243e-05,
40
+ "loss": 0.5446,
41
+ "num_tokens": 41909269.0,
42
+ "step": 20
43
+ },
44
+ {
45
+ "epoch": 0.09319664492078285,
46
+ "grad_norm": 0.7640819097303133,
47
+ "learning_rate": 4.979207812402531e-05,
48
+ "loss": 0.5249,
49
+ "num_tokens": 52353719.0,
50
+ "step": 25
51
+ },
52
+ {
53
+ "epoch": 0.11183597390493942,
54
+ "grad_norm": 0.7051975732110988,
55
+ "learning_rate": 4.956085596012407e-05,
56
+ "loss": 0.5082,
57
+ "num_tokens": 62792124.0,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.13047530288909598,
62
+ "grad_norm": 0.5494431157725719,
63
+ "learning_rate": 4.924528939432311e-05,
64
+ "loss": 0.4926,
65
+ "num_tokens": 73260442.0,
66
+ "step": 35
67
+ },
68
+ {
69
+ "epoch": 0.14911463187325255,
70
+ "grad_norm": 0.4434538386895568,
71
+ "learning_rate": 4.884658491984735e-05,
72
+ "loss": 0.4775,
73
+ "num_tokens": 83727556.0,
74
+ "step": 40
75
+ },
76
+ {
77
+ "epoch": 0.16775396085740912,
78
+ "grad_norm": 0.36186373302288577,
79
+ "learning_rate": 4.8366266887814235e-05,
80
+ "loss": 0.4756,
81
+ "num_tokens": 94201938.0,
82
+ "step": 45
83
+ },
84
+ {
85
+ "epoch": 0.1863932898415657,
86
+ "grad_norm": 0.4278036185398598,
87
+ "learning_rate": 4.780617167924209e-05,
88
+ "loss": 0.4688,
89
+ "num_tokens": 104670740.0,
90
+ "step": 50
91
+ },
92
+ {
93
+ "epoch": 0.20503261882572227,
94
+ "grad_norm": 0.49396611069124025,
95
+ "learning_rate": 4.716844068408693e-05,
96
+ "loss": 0.4707,
97
+ "num_tokens": 115151933.0,
98
+ "step": 55
99
+ },
100
+ {
101
+ "epoch": 0.22367194780987884,
102
+ "grad_norm": 0.7061064770433652,
103
+ "learning_rate": 4.6455512114150546e-05,
104
+ "loss": 0.4636,
105
+ "num_tokens": 125607410.0,
106
+ "step": 60
107
+ },
108
+ {
109
+ "epoch": 0.2423112767940354,
110
+ "grad_norm": 0.5058313809026105,
111
+ "learning_rate": 4.5670111681161296e-05,
112
+ "loss": 0.4622,
113
+ "num_tokens": 136036466.0,
114
+ "step": 65
115
+ },
116
+ {
117
+ "epoch": 0.26095060577819196,
118
+ "grad_norm": 0.4448833534200374,
119
+ "learning_rate": 4.481524217566783e-05,
120
+ "loss": 0.4584,
121
+ "num_tokens": 146502251.0,
122
+ "step": 70
123
+ },
124
+ {
125
+ "epoch": 0.27958993476234856,
126
+ "grad_norm": 0.4081315243864177,
127
+ "learning_rate": 4.3894171986588217e-05,
128
+ "loss": 0.456,
129
+ "num_tokens": 156977308.0,
130
+ "step": 75
131
+ },
132
+ {
133
+ "epoch": 0.2982292637465051,
134
+ "grad_norm": 0.31799801669428307,
135
+ "learning_rate": 4.29104226053073e-05,
136
+ "loss": 0.451,
137
+ "num_tokens": 167423794.0,
138
+ "step": 80
139
+ },
140
+ {
141
+ "epoch": 0.3168685927306617,
142
+ "grad_norm": 0.25130630836972606,
143
+ "learning_rate": 4.186775516209732e-05,
144
+ "loss": 0.445,
145
+ "num_tokens": 177900466.0,
146
+ "step": 85
147
+ },
148
+ {
149
+ "epoch": 0.33550792171481825,
150
+ "grad_norm": 0.3766061697367993,
151
+ "learning_rate": 4.077015604633669e-05,
152
+ "loss": 0.4479,
153
+ "num_tokens": 188373813.0,
154
+ "step": 90
155
+ },
156
+ {
157
+ "epoch": 0.35414725069897485,
158
+ "grad_norm": 0.3619897065914863,
159
+ "learning_rate": 3.962182166550441e-05,
160
+ "loss": 0.4462,
161
+ "num_tokens": 198859382.0,
162
+ "step": 95
163
+ },
164
+ {
165
+ "epoch": 0.3727865796831314,
166
+ "grad_norm": 0.34068435938936203,
167
+ "learning_rate": 3.8427142401220634e-05,
168
+ "loss": 0.4441,
169
+ "num_tokens": 209345142.0,
170
+ "step": 100
171
+ },
172
+ {
173
+ "epoch": 0.391425908667288,
174
+ "grad_norm": 0.3381570974545347,
175
+ "learning_rate": 3.71906858236735e-05,
176
+ "loss": 0.4428,
177
+ "num_tokens": 219820636.0,
178
+ "step": 105
179
+ },
180
+ {
181
+ "epoch": 0.41006523765144454,
182
+ "grad_norm": 0.34544406185138454,
183
+ "learning_rate": 3.591717922860785e-05,
184
+ "loss": 0.4371,
185
+ "num_tokens": 230306396.0,
186
+ "step": 110
187
+ },
188
+ {
189
+ "epoch": 0.42870456663560114,
190
+ "grad_norm": 0.2955967687161345,
191
+ "learning_rate": 3.46114915636416e-05,
192
+ "loss": 0.4399,
193
+ "num_tokens": 240792156.0,
194
+ "step": 115
195
+ },
196
+ {
197
+ "epoch": 0.4473438956197577,
198
+ "grad_norm": 0.2952519756856542,
199
+ "learning_rate": 3.3278614813010034e-05,
200
+ "loss": 0.4377,
201
+ "num_tokens": 251259365.0,
202
+ "step": 120
203
+ },
204
+ {
205
+ "epoch": 0.4659832246039143,
206
+ "grad_norm": 0.33987091063444214,
207
+ "learning_rate": 3.1923644911909e-05,
208
+ "loss": 0.4409,
209
+ "num_tokens": 261736444.0,
210
+ "step": 125
211
+ },
212
+ {
213
+ "epoch": 0.4846225535880708,
214
+ "grad_norm": 0.2945130622298854,
215
+ "learning_rate": 3.0551762263406576e-05,
216
+ "loss": 0.4393,
217
+ "num_tokens": 272205866.0,
218
+ "step": 130
219
+ },
220
+ {
221
+ "epoch": 0.5032618825722274,
222
+ "grad_norm": 0.3070511037477627,
223
+ "learning_rate": 2.9168211932412042e-05,
224
+ "loss": 0.4361,
225
+ "num_tokens": 282683327.0,
226
+ "step": 135
227
+ },
228
+ {
229
+ "epoch": 0.5219012115563839,
230
+ "grad_norm": 0.31462438644621554,
231
+ "learning_rate": 2.777828359242567e-05,
232
+ "loss": 0.4378,
233
+ "num_tokens": 293122806.0,
234
+ "step": 140
235
+ },
236
+ {
237
+ "epoch": 0.5405405405405406,
238
+ "grad_norm": 0.28655205481310253,
239
+ "learning_rate": 2.6387291301738377e-05,
240
+ "loss": 0.4372,
241
+ "num_tokens": 303598947.0,
242
+ "step": 145
243
+ },
244
+ {
245
+ "epoch": 0.5591798695246971,
246
+ "grad_norm": 0.2732532198306148,
247
+ "learning_rate": 2.50005531864019e-05,
248
+ "loss": 0.435,
249
+ "num_tokens": 314077641.0,
250
+ "step": 150
251
+ },
252
+ {
253
+ "epoch": 0.5778191985088537,
254
+ "grad_norm": 0.2715460467250568,
255
+ "learning_rate": 2.362337110764688e-05,
256
+ "loss": 0.4273,
257
+ "num_tokens": 324545386.0,
258
+ "step": 155
259
+ },
260
+ {
261
+ "epoch": 0.5964585274930102,
262
+ "grad_norm": 0.25285042415640246,
263
+ "learning_rate": 2.226101039148557e-05,
264
+ "loss": 0.4367,
265
+ "num_tokens": 335031146.0,
266
+ "step": 160
267
+ },
268
+ {
269
+ "epoch": 0.6150978564771669,
270
+ "grad_norm": 0.2149192519078472,
271
+ "learning_rate": 2.0918679697998252e-05,
272
+ "loss": 0.4212,
273
+ "num_tokens": 345504904.0,
274
+ "step": 165
275
+ },
276
+ {
277
+ "epoch": 0.6337371854613234,
278
+ "grad_norm": 0.21785863501676977,
279
+ "learning_rate": 1.9601511107268255e-05,
280
+ "loss": 0.4284,
281
+ "num_tokens": 355956520.0,
282
+ "step": 170
283
+ },
284
+ {
285
+ "epoch": 0.65237651444548,
286
+ "grad_norm": 0.23310475937674308,
287
+ "learning_rate": 1.8314540498102216e-05,
288
+ "loss": 0.4285,
289
+ "num_tokens": 366434348.0,
290
+ "step": 175
291
+ },
292
+ {
293
+ "epoch": 0.6710158434296365,
294
+ "grad_norm": 0.20067515546839765,
295
+ "learning_rate": 1.7062688294552992e-05,
296
+ "loss": 0.4313,
297
+ "num_tokens": 376908695.0,
298
+ "step": 180
299
+ },
300
+ {
301
+ "epoch": 0.6896551724137931,
302
+ "grad_norm": 0.2271355571528628,
303
+ "learning_rate": 1.5850740653856096e-05,
304
+ "loss": 0.4325,
305
+ "num_tokens": 387384301.0,
306
+ "step": 185
307
+ },
308
+ {
309
+ "epoch": 0.7082945013979497,
310
+ "grad_norm": 0.2094480953288933,
311
+ "learning_rate": 1.4683331167703218e-05,
312
+ "loss": 0.4251,
313
+ "num_tokens": 397835323.0,
314
+ "step": 190
315
+ },
316
+ {
317
+ "epoch": 0.7269338303821062,
318
+ "grad_norm": 0.18502938113842718,
319
+ "learning_rate": 1.356492314681356e-05,
320
+ "loss": 0.4243,
321
+ "num_tokens": 408298322.0,
322
+ "step": 195
323
+ },
324
+ {
325
+ "epoch": 0.7455731593662628,
326
+ "grad_norm": 0.18093742204905666,
327
+ "learning_rate": 1.2499792556533716e-05,
328
+ "loss": 0.429,
329
+ "num_tokens": 418781566.0,
330
+ "step": 200
331
+ },
332
+ {
333
+ "epoch": 0.7642124883504194,
334
+ "grad_norm": 0.1807954132426552,
335
+ "learning_rate": 1.1492011668707753e-05,
336
+ "loss": 0.4253,
337
+ "num_tokens": 429246222.0,
338
+ "step": 205
339
+ },
340
+ {
341
+ "epoch": 0.782851817334576,
342
+ "grad_norm": 0.17659671150040587,
343
+ "learning_rate": 1.0545433492320603e-05,
344
+ "loss": 0.4201,
345
+ "num_tokens": 439717644.0,
346
+ "step": 210
347
+ },
348
+ {
349
+ "epoch": 0.8014911463187325,
350
+ "grad_norm": 0.18699345886220464,
351
+ "learning_rate": 9.663677042440537e-06,
352
+ "loss": 0.4274,
353
+ "num_tokens": 450170885.0,
354
+ "step": 215
355
+ },
356
+ {
357
+ "epoch": 0.8201304753028891,
358
+ "grad_norm": 0.17996961140043713,
359
+ "learning_rate": 8.850113503781367e-06,
360
+ "loss": 0.4245,
361
+ "num_tokens": 460656645.0,
362
+ "step": 220
363
+ },
364
+ {
365
+ "epoch": 0.8387698042870456,
366
+ "grad_norm": 0.19087861268470965,
367
+ "learning_rate": 8.107853341784671e-06,
368
+ "loss": 0.4184,
369
+ "num_tokens": 471142405.0,
370
+ "step": 225
371
+ },
372
+ {
373
+ "epoch": 0.8574091332712023,
374
+ "grad_norm": 0.1690921442521712,
375
+ "learning_rate": 7.439734410499752e-06,
376
+ "loss": 0.4212,
377
+ "num_tokens": 481616409.0,
378
+ "step": 230
379
+ },
380
+ {
381
+ "epoch": 0.8760484622553588,
382
+ "grad_norm": 0.18989008366244767,
383
+ "learning_rate": 6.848311102728011e-06,
384
+ "loss": 0.4211,
385
+ "num_tokens": 492097674.0,
386
+ "step": 235
387
+ },
388
+ {
389
+ "epoch": 0.8946877912395154,
390
+ "grad_norm": 0.16510657640464366,
391
+ "learning_rate": 6.335844583913515e-06,
392
+ "loss": 0.4188,
393
+ "num_tokens": 502555042.0,
394
+ "step": 240
395
+ },
396
+ {
397
+ "epoch": 0.9133271202236719,
398
+ "grad_norm": 0.17123253579463338,
399
+ "learning_rate": 5.904294147118193e-06,
400
+ "loss": 0.4171,
401
+ "num_tokens": 513031521.0,
402
+ "step": 245
403
+ },
404
+ {
405
+ "epoch": 0.9319664492078286,
406
+ "grad_norm": 0.17312646330002712,
407
+ "learning_rate": 5.555309722133842e-06,
408
+ "loss": 0.4237,
409
+ "num_tokens": 523517281.0,
410
+ "step": 250
411
+ },
412
+ {
413
+ "epoch": 0.9506057781919851,
414
+ "grad_norm": 0.1755636687629801,
415
+ "learning_rate": 5.290225567370509e-06,
416
+ "loss": 0.4206,
417
+ "num_tokens": 533998146.0,
418
+ "step": 255
419
+ },
420
+ {
421
+ "epoch": 0.9692451071761417,
422
+ "grad_norm": 0.1535618058215829,
423
+ "learning_rate": 5.110055168638854e-06,
424
+ "loss": 0.4173,
425
+ "num_tokens": 544470561.0,
426
+ "step": 260
427
+ },
428
+ {
429
+ "epoch": 0.9878844361602982,
430
+ "grad_norm": 0.15891062616343388,
431
+ "learning_rate": 5.0154873643297575e-06,
432
+ "loss": 0.4139,
433
+ "num_tokens": 554934577.0,
434
+ "step": 265
435
+ },
436
+ {
437
+ "epoch": 0.9990680335507922,
438
+ "num_tokens": 561226033.0,
439
+ "step": 268,
440
+ "total_flos": 1.379511412069499e+18,
441
+ "train_loss": 0.45678649598093174,
442
+ "train_runtime": 12057.0545,
443
+ "train_samples_per_second": 2.845,
444
+ "train_steps_per_second": 0.022
445
+ }
446
+ ],
447
+ "logging_steps": 5,
448
+ "max_steps": 268,
449
+ "num_input_tokens_seen": 0,
450
+ "num_train_epochs": 1,
451
+ "save_steps": 10,
452
+ "stateful_callbacks": {
453
+ "TrainerControl": {
454
+ "args": {
455
+ "should_epoch_stop": false,
456
+ "should_evaluate": false,
457
+ "should_log": false,
458
+ "should_save": true,
459
+ "should_training_stop": true
460
+ },
461
+ "attributes": {}
462
+ }
463
+ },
464
+ "total_flos": 1.379511412069499e+18,
465
+ "train_batch_size": 4,
466
+ "trial_name": null,
467
+ "trial_params": null
468
+ }