End of training
Browse files- all_results.json +3 -3
- model-00001-of-00002.safetensors +1 -1
- model-00002-of-00002.safetensors +1 -1
- train_results.json +3 -3
- trainer_state.json +511 -511
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"total_flos": 0.0,
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 140,
|
6 |
-
"train_samples_per_second": 1.
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
|
|
1 |
{
|
2 |
"total_flos": 0.0,
|
3 |
+
"train_loss": 296549.3844699778,
|
4 |
+
"train_runtime": 1501.8809,
|
5 |
"train_samples": 140,
|
6 |
+
"train_samples_per_second": 1.065,
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
model-00001-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4996670464
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7aaa9065df806024fdc8b6235eb388fe611870af778efd05d97eff23f4eb2a52
|
3 |
size 4996670464
|
model-00002-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1178224960
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3124cc9c290006d42f9e8ff285c36d3a31712f9601e7c939cc305a738cb9539c
|
3 |
size 1178224960
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"total_flos": 0.0,
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 140,
|
6 |
-
"train_samples_per_second": 1.
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
|
|
1 |
{
|
2 |
"total_flos": 0.0,
|
3 |
+
"train_loss": 296549.3844699778,
|
4 |
+
"train_runtime": 1501.8809,
|
5 |
"train_samples": 140,
|
6 |
+
"train_samples_per_second": 1.065,
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
trainer_state.json
CHANGED
@@ -10,962 +10,962 @@
|
|
10 |
"is_world_process_zero": true,
|
11 |
"log_history": [
|
12 |
{
|
13 |
-
"completion_length":
|
14 |
"epoch": 0.11428571428571428,
|
15 |
-
"grad_norm":
|
16 |
"kl": 0.0,
|
17 |
"learning_rate": 1.6666666666666665e-07,
|
18 |
"loss": 0.0,
|
19 |
-
"reward":
|
20 |
-
"reward_std":
|
21 |
-
"rewards/concensus_correctness_reward_func":
|
22 |
-
"rewards/consensus_reward_func": 1.
|
23 |
"rewards/cumulative_reward_2": 0.0,
|
24 |
-
"rewards/final_correctness_reward_func": 1.
|
25 |
-
"rewards/question_recreation_reward_func": 0.
|
26 |
"rewards/soft_format_reward_func": 0.0,
|
27 |
"rewards/strict_format_reward_func": 0.25,
|
28 |
-
"rewards/xmlcount_reward_func": 0.
|
29 |
"step": 2
|
30 |
},
|
31 |
{
|
32 |
-
"completion_length":
|
33 |
"epoch": 0.22857142857142856,
|
34 |
-
"grad_norm":
|
35 |
-
"kl": 0.
|
36 |
"learning_rate": 5e-07,
|
37 |
-
"loss": 0.
|
38 |
-
"reward":
|
39 |
-
"reward_std":
|
40 |
-
"rewards/concensus_correctness_reward_func":
|
41 |
-
"rewards/consensus_reward_func":
|
42 |
"rewards/cumulative_reward_2": 0.0,
|
43 |
-
"rewards/final_correctness_reward_func": 1.
|
44 |
-
"rewards/question_recreation_reward_func": 0.
|
45 |
"rewards/soft_format_reward_func": 0.0,
|
46 |
"rewards/strict_format_reward_func": 0.484375,
|
47 |
-
"rewards/xmlcount_reward_func": 1.
|
48 |
"step": 4
|
49 |
},
|
50 |
{
|
51 |
-
"completion_length":
|
52 |
"epoch": 0.34285714285714286,
|
53 |
-
"grad_norm":
|
54 |
-
"kl":
|
55 |
"learning_rate": 4.994757065594279e-07,
|
56 |
-
"loss":
|
57 |
-
"reward":
|
58 |
-
"reward_std": 0.
|
59 |
-
"rewards/concensus_correctness_reward_func":
|
60 |
-
"rewards/consensus_reward_func":
|
61 |
"rewards/cumulative_reward_2": 0.0,
|
62 |
-
"rewards/final_correctness_reward_func": 1.
|
63 |
-
"rewards/question_recreation_reward_func": 0.
|
64 |
"rewards/soft_format_reward_func": 0.0,
|
65 |
"rewards/strict_format_reward_func": 0.453125,
|
66 |
-
"rewards/xmlcount_reward_func": 1.
|
67 |
"step": 6
|
68 |
},
|
69 |
{
|
70 |
-
"completion_length":
|
71 |
"epoch": 0.45714285714285713,
|
72 |
-
"grad_norm":
|
73 |
-
"kl": 0.
|
74 |
"learning_rate": 4.979050253066063e-07,
|
75 |
-
"loss": 0.
|
76 |
-
"reward":
|
77 |
-
"reward_std": 0.
|
78 |
-
"rewards/concensus_correctness_reward_func":
|
79 |
"rewards/consensus_reward_func": 2.0,
|
80 |
"rewards/cumulative_reward_2": 0.0,
|
81 |
-
"rewards/final_correctness_reward_func": 1.
|
82 |
-
"rewards/question_recreation_reward_func": 0.
|
83 |
"rewards/soft_format_reward_func": 0.0,
|
84 |
-
"rewards/strict_format_reward_func": 0.
|
85 |
"rewards/xmlcount_reward_func": 1.25,
|
86 |
"step": 8
|
87 |
},
|
88 |
{
|
89 |
-
"completion_length":
|
90 |
"epoch": 0.5714285714285714,
|
91 |
-
"grad_norm":
|
92 |
-
"kl": 0.
|
93 |
"learning_rate": 4.952945442245597e-07,
|
94 |
-
"loss": 0.
|
95 |
-
"reward":
|
96 |
-
"reward_std": 0.
|
97 |
"rewards/concensus_correctness_reward_func": 18.75,
|
98 |
"rewards/consensus_reward_func": 2.0,
|
99 |
"rewards/cumulative_reward_2": 0.0,
|
100 |
-
"rewards/final_correctness_reward_func": 1.
|
101 |
-
"rewards/question_recreation_reward_func": 0.
|
102 |
"rewards/soft_format_reward_func": 0.0,
|
103 |
-
"rewards/strict_format_reward_func": 0.
|
104 |
-
"rewards/xmlcount_reward_func": 1.
|
105 |
"step": 10
|
106 |
},
|
107 |
{
|
108 |
-
"completion_length":
|
109 |
"epoch": 0.6857142857142857,
|
110 |
-
"grad_norm":
|
111 |
-
"kl":
|
112 |
"learning_rate": 4.916552125781528e-07,
|
113 |
-
"loss": 0.
|
114 |
-
"reward":
|
115 |
-
"reward_std": 0.
|
116 |
-
"rewards/concensus_correctness_reward_func":
|
117 |
"rewards/consensus_reward_func": 2.0,
|
118 |
"rewards/cumulative_reward_2": 0.0,
|
119 |
-
"rewards/final_correctness_reward_func":
|
120 |
-
"rewards/question_recreation_reward_func": 0.
|
121 |
"rewards/soft_format_reward_func": 0.0,
|
122 |
"rewards/strict_format_reward_func": 0.5,
|
123 |
"rewards/xmlcount_reward_func": 1.25,
|
124 |
"step": 12
|
125 |
},
|
126 |
{
|
127 |
-
"completion_length":
|
128 |
"epoch": 0.8,
|
129 |
-
"grad_norm":
|
130 |
-
"kl": 0.
|
131 |
"learning_rate": 4.870022949890676e-07,
|
132 |
-
"loss": 0.
|
133 |
-
"reward":
|
134 |
-
"reward_std": 0.
|
135 |
-
"rewards/concensus_correctness_reward_func":
|
136 |
"rewards/consensus_reward_func": 2.0,
|
137 |
"rewards/cumulative_reward_2": 0.0,
|
138 |
-
"rewards/final_correctness_reward_func": 1.
|
139 |
-
"rewards/question_recreation_reward_func": 0.
|
140 |
"rewards/soft_format_reward_func": 0.0,
|
141 |
"rewards/strict_format_reward_func": 0.484375,
|
142 |
-
"rewards/xmlcount_reward_func": 1.
|
143 |
"step": 14
|
144 |
},
|
145 |
{
|
146 |
-
"completion_length":
|
147 |
"epoch": 0.9142857142857143,
|
148 |
-
"grad_norm":
|
149 |
-
"kl":
|
150 |
"learning_rate": 4.81355307410676e-07,
|
151 |
-
"loss":
|
152 |
-
"reward":
|
153 |
-
"reward_std":
|
154 |
-
"rewards/concensus_correctness_reward_func":
|
155 |
-
"rewards/consensus_reward_func":
|
156 |
"rewards/cumulative_reward_2": 0.0,
|
157 |
-
"rewards/final_correctness_reward_func": 1.
|
158 |
-
"rewards/question_recreation_reward_func": 0.
|
159 |
"rewards/soft_format_reward_func": 0.0,
|
160 |
-
"rewards/strict_format_reward_func": 0.
|
161 |
-
"rewards/xmlcount_reward_func": 1.
|
162 |
"step": 16
|
163 |
},
|
164 |
{
|
165 |
-
"completion_length":
|
166 |
"epoch": 1.0,
|
167 |
-
"grad_norm":
|
168 |
-
"kl":
|
169 |
"learning_rate": 4.747379352713488e-07,
|
170 |
-
"loss":
|
171 |
-
"reward":
|
172 |
-
"reward_std":
|
173 |
-
"rewards/concensus_correctness_reward_func":
|
174 |
-
"rewards/consensus_reward_func":
|
175 |
"rewards/cumulative_reward_2": 0.0,
|
176 |
-
"rewards/final_correctness_reward_func": 1.
|
177 |
-
"rewards/question_recreation_reward_func": 0.
|
178 |
"rewards/soft_format_reward_func": 0.0,
|
179 |
"rewards/strict_format_reward_func": 0.4583333333333333,
|
180 |
-
"rewards/xmlcount_reward_func": 1.
|
181 |
"step": 18
|
182 |
},
|
183 |
{
|
184 |
-
"completion_length":
|
185 |
"epoch": 1.1142857142857143,
|
186 |
-
"grad_norm":
|
187 |
-
"kl": 0.
|
188 |
"learning_rate": 4.6717793412953776e-07,
|
189 |
-
"loss": 0.
|
190 |
-
"reward":
|
191 |
-
"reward_std": 0.
|
192 |
-
"rewards/concensus_correctness_reward_func":
|
193 |
"rewards/consensus_reward_func": 2.0,
|
194 |
"rewards/cumulative_reward_2": 0.0,
|
195 |
-
"rewards/final_correctness_reward_func": 1.
|
196 |
-
"rewards/question_recreation_reward_func": 0.
|
197 |
"rewards/soft_format_reward_func": 0.0,
|
198 |
-
"rewards/strict_format_reward_func": 0.
|
199 |
-
"rewards/xmlcount_reward_func": 1.
|
200 |
"step": 20
|
201 |
},
|
202 |
{
|
203 |
-
"completion_length":
|
204 |
"epoch": 1.2285714285714286,
|
205 |
-
"grad_norm":
|
206 |
-
"kl":
|
207 |
"learning_rate": 4.5870701325731773e-07,
|
208 |
-
"loss": 0.
|
209 |
-
"reward":
|
210 |
-
"reward_std": 0.
|
211 |
"rewards/concensus_correctness_reward_func": 18.75,
|
212 |
"rewards/consensus_reward_func": 2.0,
|
213 |
"rewards/cumulative_reward_2": 0.0,
|
214 |
-
"rewards/final_correctness_reward_func": 1.
|
215 |
-
"rewards/question_recreation_reward_func": 0.
|
216 |
"rewards/soft_format_reward_func": 0.0,
|
217 |
-
"rewards/strict_format_reward_func": 0.
|
218 |
-
"rewards/xmlcount_reward_func": 1.
|
219 |
"step": 22
|
220 |
},
|
221 |
{
|
222 |
-
"completion_length":
|
223 |
"epoch": 1.342857142857143,
|
224 |
-
"grad_norm":
|
225 |
-
"kl":
|
226 |
"learning_rate": 4.4936070264068016e-07,
|
227 |
-
"loss": 0.
|
228 |
-
"reward":
|
229 |
-
"reward_std": 0.
|
230 |
-
"rewards/concensus_correctness_reward_func":
|
231 |
"rewards/consensus_reward_func": 2.0,
|
232 |
"rewards/cumulative_reward_2": 0.0,
|
233 |
"rewards/final_correctness_reward_func": 1.75,
|
234 |
-
"rewards/question_recreation_reward_func": 0.
|
235 |
"rewards/soft_format_reward_func": 0.0,
|
236 |
-
"rewards/strict_format_reward_func": 0.
|
237 |
-
"rewards/xmlcount_reward_func": 1.
|
238 |
"step": 24
|
239 |
},
|
240 |
{
|
241 |
-
"completion_length":
|
242 |
"epoch": 1.457142857142857,
|
243 |
-
"grad_norm":
|
244 |
-
"kl":
|
245 |
"learning_rate": 4.391782039544238e-07,
|
246 |
-
"loss":
|
247 |
-
"reward":
|
248 |
-
"reward_std":
|
249 |
-
"rewards/concensus_correctness_reward_func": 16.
|
250 |
-
"rewards/consensus_reward_func":
|
251 |
"rewards/cumulative_reward_2": 0.0,
|
252 |
-
"rewards/final_correctness_reward_func": 1.
|
253 |
-
"rewards/question_recreation_reward_func": 0.
|
254 |
"rewards/soft_format_reward_func": 0.0,
|
255 |
-
"rewards/strict_format_reward_func": 0.
|
256 |
-
"rewards/xmlcount_reward_func": 1.
|
257 |
"step": 26
|
258 |
},
|
259 |
{
|
260 |
-
"completion_length":
|
261 |
"epoch": 1.5714285714285714,
|
262 |
-
"grad_norm":
|
263 |
-
"kl":
|
264 |
"learning_rate": 4.282022261367073e-07,
|
265 |
-
"loss":
|
266 |
-
"reward":
|
267 |
-
"reward_std": 0.
|
268 |
"rewards/concensus_correctness_reward_func": 15.0,
|
269 |
-
"rewards/consensus_reward_func":
|
270 |
"rewards/cumulative_reward_2": 0.0,
|
271 |
-
"rewards/final_correctness_reward_func": 1.
|
272 |
-
"rewards/question_recreation_reward_func": 0.
|
273 |
"rewards/soft_format_reward_func": 0.0,
|
274 |
-
"rewards/strict_format_reward_func": 0.
|
275 |
-
"rewards/xmlcount_reward_func": 1.
|
276 |
"step": 28
|
277 |
},
|
278 |
{
|
279 |
-
"completion_length":
|
280 |
"epoch": 1.6857142857142857,
|
281 |
-
"grad_norm":
|
282 |
-
"kl":
|
283 |
"learning_rate": 4.1647880625292027e-07,
|
284 |
-
"loss":
|
285 |
-
"reward":
|
286 |
-
"reward_std": 0.
|
287 |
-
"rewards/concensus_correctness_reward_func":
|
288 |
"rewards/consensus_reward_func": 2.0,
|
289 |
"rewards/cumulative_reward_2": 0.0,
|
290 |
-
"rewards/final_correctness_reward_func": 1.
|
291 |
-
"rewards/question_recreation_reward_func": 0.
|
292 |
"rewards/soft_format_reward_func": 0.0,
|
293 |
-
"rewards/strict_format_reward_func": 0.
|
294 |
-
"rewards/xmlcount_reward_func": 1.
|
295 |
"step": 30
|
296 |
},
|
297 |
{
|
298 |
-
"completion_length":
|
299 |
"epoch": 1.8,
|
300 |
-
"grad_norm":
|
301 |
-
"kl":
|
302 |
"learning_rate": 4.040571164002318e-07,
|
303 |
-
"loss":
|
304 |
-
"reward":
|
305 |
-
"reward_std": 0.
|
306 |
-
"rewards/concensus_correctness_reward_func":
|
307 |
-
"rewards/consensus_reward_func":
|
308 |
"rewards/cumulative_reward_2": 0.0,
|
309 |
-
"rewards/final_correctness_reward_func": 1.
|
310 |
-
"rewards/question_recreation_reward_func": 0.
|
311 |
"rewards/soft_format_reward_func": 0.0,
|
312 |
"rewards/strict_format_reward_func": 0.46875,
|
313 |
-
"rewards/xmlcount_reward_func": 1.
|
314 |
"step": 32
|
315 |
},
|
316 |
{
|
317 |
-
"completion_length":
|
318 |
"epoch": 1.9142857142857141,
|
319 |
-
"grad_norm":
|
320 |
-
"kl":
|
321 |
"learning_rate": 3.909892574627266e-07,
|
322 |
-
"loss":
|
323 |
-
"reward":
|
324 |
-
"reward_std": 0.
|
325 |
-
"rewards/concensus_correctness_reward_func":
|
326 |
"rewards/consensus_reward_func": 2.0,
|
327 |
"rewards/cumulative_reward_2": 0.0,
|
328 |
-
"rewards/final_correctness_reward_func": 1.
|
329 |
-
"rewards/question_recreation_reward_func": 0.
|
330 |
"rewards/soft_format_reward_func": 0.0,
|
331 |
"rewards/strict_format_reward_func": 0.5,
|
332 |
"rewards/xmlcount_reward_func": 1.25,
|
333 |
"step": 34
|
334 |
},
|
335 |
{
|
336 |
-
"completion_length":
|
337 |
"epoch": 2.0,
|
338 |
-
"grad_norm":
|
339 |
-
"kl":
|
340 |
"learning_rate": 3.773300405821908e-07,
|
341 |
-
"loss":
|
342 |
-
"reward": 22.
|
343 |
-
"reward_std": 0.
|
344 |
"rewards/concensus_correctness_reward_func": 16.666666666666668,
|
345 |
"rewards/consensus_reward_func": 2.0,
|
346 |
"rewards/cumulative_reward_2": 0.0,
|
347 |
-
"rewards/final_correctness_reward_func": 1.
|
348 |
-
"rewards/question_recreation_reward_func": 0.
|
349 |
"rewards/soft_format_reward_func": 0.0,
|
350 |
-
"rewards/strict_format_reward_func": 0.
|
351 |
-
"rewards/xmlcount_reward_func": 1.
|
352 |
"step": 36
|
353 |
},
|
354 |
{
|
355 |
-
"completion_length":
|
356 |
"epoch": 2.1142857142857143,
|
357 |
-
"grad_norm":
|
358 |
-
"kl":
|
359 |
"learning_rate": 3.6313675726113475e-07,
|
360 |
-
"loss":
|
361 |
-
"reward":
|
362 |
-
"reward_std": 0.
|
363 |
-
"rewards/concensus_correctness_reward_func":
|
364 |
"rewards/consensus_reward_func": 2.0,
|
365 |
"rewards/cumulative_reward_2": 0.0,
|
366 |
-
"rewards/final_correctness_reward_func": 1.
|
367 |
-
"rewards/question_recreation_reward_func": 0.
|
368 |
"rewards/soft_format_reward_func": 0.0,
|
369 |
"rewards/strict_format_reward_func": 0.5,
|
370 |
"rewards/xmlcount_reward_func": 1.25,
|
371 |
"step": 38
|
372 |
},
|
373 |
{
|
374 |
-
"completion_length":
|
375 |
"epoch": 2.2285714285714286,
|
376 |
-
"grad_norm":
|
377 |
-
"kl":
|
378 |
"learning_rate": 3.484689390623218e-07,
|
379 |
-
"loss": 0.
|
380 |
-
"reward":
|
381 |
-
"reward_std":
|
382 |
-
"rewards/concensus_correctness_reward_func":
|
383 |
-
"rewards/consensus_reward_func":
|
384 |
"rewards/cumulative_reward_2": 0.0,
|
385 |
-
"rewards/final_correctness_reward_func": 1.
|
386 |
-
"rewards/question_recreation_reward_func": 0.
|
387 |
"rewards/soft_format_reward_func": 0.0,
|
388 |
-
"rewards/strict_format_reward_func": 0.
|
389 |
-
"rewards/xmlcount_reward_func": 1.
|
390 |
"step": 40
|
391 |
},
|
392 |
{
|
393 |
-
"completion_length":
|
394 |
"epoch": 2.342857142857143,
|
395 |
-
"grad_norm":
|
396 |
-
"kl":
|
397 |
"learning_rate": 3.3338810791270517e-07,
|
398 |
-
"loss":
|
399 |
-
"reward": 19.
|
400 |
-
"reward_std": 0.
|
401 |
"rewards/concensus_correctness_reward_func": 13.75,
|
402 |
"rewards/consensus_reward_func": 2.0,
|
403 |
"rewards/cumulative_reward_2": 0.0,
|
404 |
-
"rewards/final_correctness_reward_func": 1.
|
405 |
-
"rewards/question_recreation_reward_func": 0.
|
406 |
"rewards/soft_format_reward_func": 0.0,
|
407 |
-
"rewards/strict_format_reward_func": 0.
|
408 |
-
"rewards/xmlcount_reward_func": 1.
|
409 |
"step": 42
|
410 |
},
|
411 |
{
|
412 |
-
"completion_length":
|
413 |
"epoch": 2.4571428571428573,
|
414 |
-
"grad_norm":
|
415 |
-
"kl":
|
416 |
"learning_rate": 3.179575180590857e-07,
|
417 |
-
"loss": 0.
|
418 |
-
"reward":
|
419 |
-
"reward_std":
|
420 |
-
"rewards/concensus_correctness_reward_func":
|
421 |
-
"rewards/consensus_reward_func":
|
422 |
"rewards/cumulative_reward_2": 0.0,
|
423 |
"rewards/final_correctness_reward_func": 1.375,
|
424 |
-
"rewards/question_recreation_reward_func": 0.
|
425 |
"rewards/soft_format_reward_func": 0.0,
|
426 |
"rewards/strict_format_reward_func": 0.5,
|
427 |
"rewards/xmlcount_reward_func": 1.25,
|
428 |
"step": 44
|
429 |
},
|
430 |
{
|
431 |
-
"completion_length":
|
432 |
"epoch": 2.571428571428571,
|
433 |
-
"grad_norm":
|
434 |
-
"kl":
|
435 |
"learning_rate": 3.022418907578188e-07,
|
436 |
-
"loss": 0.
|
437 |
-
"reward": 22.
|
438 |
-
"reward_std": 0.
|
439 |
"rewards/concensus_correctness_reward_func": 16.25,
|
440 |
-
"rewards/consensus_reward_func":
|
441 |
"rewards/cumulative_reward_2": 0.0,
|
442 |
-
"rewards/final_correctness_reward_func": 1.
|
443 |
-
"rewards/question_recreation_reward_func": 0.
|
444 |
"rewards/soft_format_reward_func": 0.0,
|
445 |
-
"rewards/strict_format_reward_func": 0.
|
446 |
"rewards/xmlcount_reward_func": 1.25,
|
447 |
"step": 46
|
448 |
},
|
449 |
{
|
450 |
-
"completion_length":
|
451 |
"epoch": 2.685714285714286,
|
452 |
-
"grad_norm":
|
453 |
-
"kl":
|
454 |
"learning_rate": 2.863071428113726e-07,
|
455 |
-
"loss":
|
456 |
-
"reward": 21.
|
457 |
-
"reward_std": 1.
|
458 |
"rewards/concensus_correctness_reward_func": 15.625,
|
459 |
"rewards/consensus_reward_func": 1.9375,
|
460 |
"rewards/cumulative_reward_2": 0.0,
|
461 |
-
"rewards/final_correctness_reward_func": 1.
|
462 |
-
"rewards/question_recreation_reward_func": 0.
|
463 |
"rewards/soft_format_reward_func": 0.0,
|
464 |
"rewards/strict_format_reward_func": 0.5,
|
465 |
-
"rewards/xmlcount_reward_func": 1.
|
466 |
"step": 48
|
467 |
},
|
468 |
{
|
469 |
-
"completion_length": 288.
|
470 |
"epoch": 2.8,
|
471 |
-
"grad_norm":
|
472 |
-
"kl":
|
473 |
"learning_rate": 2.7022011009035107e-07,
|
474 |
-
"loss": 0.
|
475 |
-
"reward": 23.
|
476 |
-
"reward_std": 0.
|
477 |
"rewards/concensus_correctness_reward_func": 17.5,
|
478 |
-
"rewards/consensus_reward_func":
|
479 |
"rewards/cumulative_reward_2": 0.0,
|
480 |
-
"rewards/final_correctness_reward_func": 1.
|
481 |
-
"rewards/question_recreation_reward_func": 0.
|
482 |
"rewards/soft_format_reward_func": 0.0,
|
483 |
"rewards/strict_format_reward_func": 0.46875,
|
484 |
"rewards/xmlcount_reward_func": 1.25,
|
485 |
"step": 50
|
486 |
},
|
487 |
{
|
488 |
-
"completion_length":
|
489 |
"epoch": 2.914285714285714,
|
490 |
-
"grad_norm":
|
491 |
-
"kl":
|
492 |
"learning_rate": 2.540482672006254e-07,
|
493 |
-
"loss":
|
494 |
-
"reward":
|
495 |
-
"reward_std": 0.
|
496 |
-
"rewards/concensus_correctness_reward_func":
|
497 |
"rewards/consensus_reward_func": 2.0,
|
498 |
"rewards/cumulative_reward_2": 0.0,
|
499 |
-
"rewards/final_correctness_reward_func": 1.
|
500 |
-
"rewards/question_recreation_reward_func": 0.
|
501 |
"rewards/soft_format_reward_func": 0.0,
|
502 |
-
"rewards/strict_format_reward_func": 0.
|
503 |
-
"rewards/xmlcount_reward_func": 1.
|
504 |
"step": 52
|
505 |
},
|
506 |
{
|
507 |
-
"completion_length":
|
508 |
"epoch": 3.0,
|
509 |
-
"grad_norm":
|
510 |
-
"kl":
|
511 |
"learning_rate": 2.37859444471388e-07,
|
512 |
-
"loss":
|
513 |
-
"reward":
|
514 |
-
"reward_std":
|
515 |
-
"rewards/concensus_correctness_reward_func": 15.
|
516 |
-
"rewards/consensus_reward_func":
|
517 |
"rewards/cumulative_reward_2": 0.0,
|
518 |
-
"rewards/final_correctness_reward_func": 1.
|
519 |
-
"rewards/question_recreation_reward_func": 0.
|
520 |
"rewards/soft_format_reward_func": 0.0,
|
521 |
"rewards/strict_format_reward_func": 0.5,
|
522 |
"rewards/xmlcount_reward_func": 1.25,
|
523 |
"step": 54
|
524 |
},
|
525 |
{
|
526 |
-
"completion_length":
|
527 |
"epoch": 3.1142857142857143,
|
528 |
-
"grad_norm":
|
529 |
-
"kl":
|
530 |
"learning_rate": 2.2172154345117894e-07,
|
531 |
-
"loss": 0.
|
532 |
-
"reward": 23.
|
533 |
-
"reward_std": 0.
|
534 |
"rewards/concensus_correctness_reward_func": 17.5,
|
535 |
"rewards/consensus_reward_func": 2.0,
|
536 |
"rewards/cumulative_reward_2": 0.0,
|
537 |
"rewards/final_correctness_reward_func": 1.6875,
|
538 |
-
"rewards/question_recreation_reward_func": 0.
|
539 |
"rewards/soft_format_reward_func": 0.0,
|
540 |
"rewards/strict_format_reward_func": 0.5,
|
541 |
"rewards/xmlcount_reward_func": 1.25,
|
542 |
"step": 56
|
543 |
},
|
544 |
{
|
545 |
-
"completion_length":
|
546 |
"epoch": 3.2285714285714286,
|
547 |
-
"grad_norm":
|
548 |
-
"kl": 0.
|
549 |
"learning_rate": 2.0570225210519433e-07,
|
550 |
-
"loss": 0.
|
551 |
-
"reward": 21.
|
552 |
-
"reward_std": 0.
|
553 |
"rewards/concensus_correctness_reward_func": 15.0,
|
554 |
"rewards/consensus_reward_func": 2.0,
|
555 |
"rewards/cumulative_reward_2": 0.0,
|
556 |
-
"rewards/final_correctness_reward_func": 1.
|
557 |
-
"rewards/question_recreation_reward_func": 0.
|
558 |
"rewards/soft_format_reward_func": 0.0,
|
559 |
-
"rewards/strict_format_reward_func": 0.
|
560 |
-
"rewards/xmlcount_reward_func": 1.
|
561 |
"step": 58
|
562 |
},
|
563 |
{
|
564 |
-
"completion_length":
|
565 |
"epoch": 3.342857142857143,
|
566 |
-
"grad_norm":
|
567 |
-
"kl":
|
568 |
"learning_rate": 1.8986876090843664e-07,
|
569 |
-
"loss":
|
570 |
-
"reward":
|
571 |
-
"reward_std": 0.
|
572 |
-
"rewards/concensus_correctness_reward_func":
|
573 |
-
"rewards/consensus_reward_func":
|
574 |
"rewards/cumulative_reward_2": 0.0,
|
575 |
-
"rewards/final_correctness_reward_func": 1.
|
576 |
-
"rewards/question_recreation_reward_func": 0.
|
577 |
"rewards/soft_format_reward_func": 0.0,
|
578 |
-
"rewards/strict_format_reward_func": 0.
|
579 |
"rewards/xmlcount_reward_func": 1.25,
|
580 |
"step": 60
|
581 |
},
|
582 |
{
|
583 |
-
"completion_length":
|
584 |
"epoch": 3.4571428571428573,
|
585 |
-
"grad_norm":
|
586 |
-
"kl":
|
587 |
"learning_rate": 1.7428748102551234e-07,
|
588 |
-
"loss": 0.
|
589 |
-
"reward": 21.
|
590 |
-
"reward_std": 0.
|
591 |
"rewards/concensus_correctness_reward_func": 15.0,
|
592 |
-
"rewards/consensus_reward_func":
|
593 |
"rewards/cumulative_reward_2": 0.0,
|
594 |
-
"rewards/final_correctness_reward_func": 1.
|
595 |
-
"rewards/question_recreation_reward_func": 0.
|
596 |
"rewards/soft_format_reward_func": 0.0,
|
597 |
-
"rewards/strict_format_reward_func": 0.
|
598 |
"rewards/xmlcount_reward_func": 1.25,
|
599 |
"step": 62
|
600 |
},
|
601 |
{
|
602 |
-
"completion_length":
|
603 |
"epoch": 3.571428571428571,
|
604 |
-
"grad_norm":
|
605 |
-
"kl":
|
606 |
"learning_rate": 1.5902376575912814e-07,
|
607 |
-
"loss": 0.
|
608 |
-
"reward": 22.
|
609 |
-
"reward_std": 0.
|
610 |
"rewards/concensus_correctness_reward_func": 16.25,
|
611 |
"rewards/consensus_reward_func": 2.0,
|
612 |
"rewards/cumulative_reward_2": 0.0,
|
613 |
-
"rewards/final_correctness_reward_func": 1.
|
614 |
-
"rewards/question_recreation_reward_func": 0.
|
615 |
"rewards/soft_format_reward_func": 0.0,
|
616 |
"rewards/strict_format_reward_func": 0.5,
|
617 |
"rewards/xmlcount_reward_func": 1.25,
|
618 |
"step": 64
|
619 |
},
|
620 |
{
|
621 |
-
"completion_length":
|
622 |
"epoch": 3.685714285714286,
|
623 |
-
"grad_norm":
|
624 |
-
"kl": 0.
|
625 |
"learning_rate": 1.4414163643562753e-07,
|
626 |
-
"loss": 0.
|
627 |
-
"reward": 22.
|
628 |
-
"reward_std": 0.
|
629 |
"rewards/concensus_correctness_reward_func": 16.25,
|
630 |
"rewards/consensus_reward_func": 2.0,
|
631 |
"rewards/cumulative_reward_2": 0.0,
|
632 |
-
"rewards/final_correctness_reward_func": 1.
|
633 |
-
"rewards/question_recreation_reward_func": 0.
|
634 |
"rewards/soft_format_reward_func": 0.0,
|
635 |
"rewards/strict_format_reward_func": 0.5,
|
636 |
"rewards/xmlcount_reward_func": 1.25,
|
637 |
"step": 66
|
638 |
},
|
639 |
{
|
640 |
-
"completion_length":
|
641 |
"epoch": 3.8,
|
642 |
-
"grad_norm":
|
643 |
-
"kl":
|
644 |
"learning_rate": 1.2970351387729872e-07,
|
645 |
-
"loss":
|
646 |
-
"reward":
|
647 |
-
"reward_std": 0.
|
648 |
-
"rewards/concensus_correctness_reward_func":
|
649 |
"rewards/consensus_reward_func": 2.0,
|
650 |
"rewards/cumulative_reward_2": 0.0,
|
651 |
"rewards/final_correctness_reward_func": 1.4375,
|
652 |
-
"rewards/question_recreation_reward_func": 0.
|
653 |
"rewards/soft_format_reward_func": 0.0,
|
654 |
-
"rewards/strict_format_reward_func": 0.
|
655 |
-
"rewards/xmlcount_reward_func": 1.
|
656 |
"step": 68
|
657 |
},
|
658 |
{
|
659 |
-
"completion_length":
|
660 |
"epoch": 3.914285714285714,
|
661 |
-
"grad_norm": 5.
|
662 |
-
"kl":
|
663 |
"learning_rate": 1.1576995658775404e-07,
|
664 |
-
"loss":
|
665 |
-
"reward": 24.
|
666 |
-
"reward_std":
|
667 |
-
"rewards/concensus_correctness_reward_func": 18.
|
668 |
-
"rewards/consensus_reward_func":
|
669 |
"rewards/cumulative_reward_2": 0.0,
|
670 |
-
"rewards/final_correctness_reward_func": 1.
|
671 |
-
"rewards/question_recreation_reward_func": 0.
|
672 |
"rewards/soft_format_reward_func": 0.0,
|
673 |
"rewards/strict_format_reward_func": 0.46875,
|
674 |
-
"rewards/xmlcount_reward_func": 1.
|
675 |
"step": 70
|
676 |
},
|
677 |
{
|
678 |
-
"completion_length":
|
679 |
"epoch": 4.0,
|
680 |
-
"grad_norm": 4.
|
681 |
-
"kl":
|
682 |
"learning_rate": 1.0239940674851941e-07,
|
683 |
-
"loss":
|
684 |
-
"reward": 24.
|
685 |
-
"reward_std": 0.
|
686 |
"rewards/concensus_correctness_reward_func": 18.333333333333332,
|
687 |
"rewards/consensus_reward_func": 2.0,
|
688 |
"rewards/cumulative_reward_2": 0.0,
|
689 |
-
"rewards/final_correctness_reward_func": 1.
|
690 |
-
"rewards/question_recreation_reward_func": 0.
|
691 |
"rewards/soft_format_reward_func": 0.0,
|
692 |
-
"rewards/strict_format_reward_func": 0.
|
693 |
-
"rewards/xmlcount_reward_func": 1.
|
694 |
"step": 72
|
695 |
},
|
696 |
{
|
697 |
-
"completion_length":
|
698 |
"epoch": 4.114285714285714,
|
699 |
-
"grad_norm":
|
700 |
-
"kl":
|
701 |
"learning_rate": 8.964794509221507e-08,
|
702 |
-
"loss":
|
703 |
-
"reward":
|
704 |
-
"reward_std":
|
705 |
-
"rewards/concensus_correctness_reward_func":
|
706 |
-
"rewards/consensus_reward_func":
|
707 |
"rewards/cumulative_reward_2": 0.0,
|
708 |
-
"rewards/final_correctness_reward_func": 1.
|
709 |
-
"rewards/question_recreation_reward_func": 0.
|
710 |
"rewards/soft_format_reward_func": 0.0,
|
711 |
"rewards/strict_format_reward_func": 0.5,
|
712 |
"rewards/xmlcount_reward_func": 1.25,
|
713 |
"step": 74
|
714 |
},
|
715 |
{
|
716 |
-
"completion_length":
|
717 |
"epoch": 4.228571428571429,
|
718 |
-
"grad_norm":
|
719 |
-
"kl":
|
720 |
"learning_rate": 7.756905568047392e-08,
|
721 |
-
"loss":
|
722 |
-
"reward":
|
723 |
-
"reward_std":
|
724 |
-
"rewards/concensus_correctness_reward_func":
|
725 |
-
"rewards/consensus_reward_func":
|
726 |
"rewards/cumulative_reward_2": 0.0,
|
727 |
-
"rewards/final_correctness_reward_func": 1.
|
728 |
-
"rewards/question_recreation_reward_func": 0.
|
729 |
"rewards/soft_format_reward_func": 0.0,
|
730 |
-
"rewards/strict_format_reward_func": 0.
|
731 |
-
"rewards/xmlcount_reward_func": 1.
|
732 |
"step": 76
|
733 |
},
|
734 |
{
|
735 |
-
"completion_length":
|
736 |
"epoch": 4.3428571428571425,
|
737 |
-
"grad_norm": 5.
|
738 |
-
"kl":
|
739 |
"learning_rate": 6.621340157319996e-08,
|
740 |
-
"loss": 0.
|
741 |
-
"reward": 20.
|
742 |
-
"reward_std": 0.
|
743 |
"rewards/concensus_correctness_reward_func": 15.0,
|
744 |
"rewards/consensus_reward_func": 2.0,
|
745 |
"rewards/cumulative_reward_2": 0.0,
|
746 |
-
"rewards/final_correctness_reward_func": 1.
|
747 |
-
"rewards/question_recreation_reward_func": 0.
|
748 |
"rewards/soft_format_reward_func": 0.0,
|
749 |
"rewards/strict_format_reward_func": 0.453125,
|
750 |
-
"rewards/xmlcount_reward_func": 1.
|
751 |
"step": 78
|
752 |
},
|
753 |
{
|
754 |
-
"completion_length":
|
755 |
"epoch": 4.457142857142857,
|
756 |
-
"grad_norm":
|
757 |
-
"kl":
|
758 |
"learning_rate": 5.5628612330087724e-08,
|
759 |
-
"loss": 0.
|
760 |
-
"reward":
|
761 |
-
"reward_std": 0.
|
762 |
-
"rewards/concensus_correctness_reward_func":
|
763 |
"rewards/consensus_reward_func": 2.0,
|
764 |
"rewards/cumulative_reward_2": 0.0,
|
765 |
-
"rewards/final_correctness_reward_func":
|
766 |
-
"rewards/question_recreation_reward_func": 0.
|
767 |
"rewards/soft_format_reward_func": 0.0,
|
768 |
-
"rewards/strict_format_reward_func": 0.
|
769 |
-
"rewards/xmlcount_reward_func": 1.
|
770 |
"step": 80
|
771 |
},
|
772 |
{
|
773 |
-
"completion_length":
|
774 |
"epoch": 4.571428571428571,
|
775 |
-
"grad_norm":
|
776 |
-
"kl":
|
777 |
"learning_rate": 4.5859084235697235e-08,
|
778 |
-
"loss": 0.
|
779 |
-
"reward":
|
780 |
-
"reward_std":
|
781 |
-
"rewards/concensus_correctness_reward_func":
|
782 |
-
"rewards/consensus_reward_func":
|
783 |
"rewards/cumulative_reward_2": 0.0,
|
784 |
"rewards/final_correctness_reward_func": 1.25,
|
785 |
-
"rewards/question_recreation_reward_func": 0.
|
786 |
"rewards/soft_format_reward_func": 0.0,
|
787 |
-
"rewards/strict_format_reward_func": 0.
|
788 |
"rewards/xmlcount_reward_func": 1.25,
|
789 |
"step": 82
|
790 |
},
|
791 |
{
|
792 |
-
"completion_length":
|
793 |
"epoch": 4.685714285714286,
|
794 |
-
"grad_norm":
|
795 |
-
"kl":
|
796 |
"learning_rate": 3.6945794086007705e-08,
|
797 |
-
"loss": 0.
|
798 |
-
"reward": 22.
|
799 |
-
"reward_std": 0.
|
800 |
"rewards/concensus_correctness_reward_func": 16.25,
|
801 |
-
"rewards/consensus_reward_func":
|
802 |
"rewards/cumulative_reward_2": 0.0,
|
803 |
-
"rewards/final_correctness_reward_func": 1.
|
804 |
-
"rewards/question_recreation_reward_func": 0.
|
805 |
"rewards/soft_format_reward_func": 0.0,
|
806 |
-
"rewards/strict_format_reward_func": 0.
|
807 |
-
"rewards/xmlcount_reward_func": 1.
|
808 |
"step": 84
|
809 |
},
|
810 |
{
|
811 |
-
"completion_length":
|
812 |
"epoch": 4.8,
|
813 |
-
"grad_norm":
|
814 |
-
"kl": 0.
|
815 |
"learning_rate": 2.892612731749414e-08,
|
816 |
-
"loss": 0.
|
817 |
-
"reward":
|
818 |
-
"reward_std": 0.
|
819 |
-
"rewards/concensus_correctness_reward_func":
|
820 |
"rewards/consensus_reward_func": 2.0,
|
821 |
"rewards/cumulative_reward_2": 0.0,
|
822 |
-
"rewards/final_correctness_reward_func": 1.
|
823 |
-
"rewards/question_recreation_reward_func": 0.
|
824 |
"rewards/soft_format_reward_func": 0.0,
|
825 |
-
"rewards/strict_format_reward_func": 0.
|
826 |
-
"rewards/xmlcount_reward_func": 1.
|
827 |
"step": 86
|
828 |
},
|
829 |
{
|
830 |
-
"completion_length":
|
831 |
"epoch": 4.914285714285715,
|
832 |
-
"grad_norm":
|
833 |
-
"kl":
|
834 |
"learning_rate": 2.183372119961499e-08,
|
835 |
-
"loss": 0.
|
836 |
-
"reward": 23.
|
837 |
-
"reward_std": 0.
|
838 |
"rewards/concensus_correctness_reward_func": 17.5,
|
839 |
"rewards/consensus_reward_func": 2.0,
|
840 |
"rewards/cumulative_reward_2": 0.0,
|
841 |
-
"rewards/final_correctness_reward_func": 1.
|
842 |
-
"rewards/question_recreation_reward_func": 0.
|
843 |
"rewards/soft_format_reward_func": 0.0,
|
844 |
"rewards/strict_format_reward_func": 0.5,
|
845 |
"rewards/xmlcount_reward_func": 1.25,
|
846 |
"step": 88
|
847 |
},
|
848 |
{
|
849 |
-
"completion_length":
|
850 |
"epoch": 5.0,
|
851 |
-
"grad_norm":
|
852 |
-
"kl":
|
853 |
"learning_rate": 1.5698323748414122e-08,
|
854 |
-
"loss": 0.
|
855 |
-
"reward":
|
856 |
-
"reward_std": 0.
|
857 |
-
"rewards/concensus_correctness_reward_func":
|
858 |
"rewards/consensus_reward_func": 2.0,
|
859 |
"rewards/cumulative_reward_2": 0.0,
|
860 |
-
"rewards/final_correctness_reward_func": 1.
|
861 |
-
"rewards/question_recreation_reward_func": 0.
|
862 |
"rewards/soft_format_reward_func": 0.0,
|
863 |
-
"rewards/strict_format_reward_func": 0.
|
864 |
"rewards/xmlcount_reward_func": 1.25,
|
865 |
"step": 90
|
866 |
},
|
867 |
{
|
868 |
-
"completion_length":
|
869 |
"epoch": 5.114285714285714,
|
870 |
-
"grad_norm":
|
871 |
-
"kl":
|
872 |
"learning_rate": 1.054566895300324e-08,
|
873 |
-
"loss": 0.
|
874 |
-
"reward":
|
875 |
-
"reward_std":
|
876 |
-
"rewards/concensus_correctness_reward_func": 18.
|
877 |
-
"rewards/consensus_reward_func":
|
878 |
"rewards/cumulative_reward_2": 0.0,
|
879 |
-
"rewards/final_correctness_reward_func": 1.
|
880 |
-
"rewards/question_recreation_reward_func": 0.
|
881 |
"rewards/soft_format_reward_func": 0.0,
|
882 |
-
"rewards/strict_format_reward_func": 0.
|
883 |
-
"rewards/xmlcount_reward_func": 1.
|
884 |
"step": 92
|
885 |
},
|
886 |
{
|
887 |
-
"completion_length":
|
888 |
"epoch": 5.228571428571429,
|
889 |
-
"grad_norm":
|
890 |
-
"kl":
|
891 |
"learning_rate": 6.397368838268496e-09,
|
892 |
-
"loss": 0.
|
893 |
-
"reward":
|
894 |
-
"reward_std": 0.
|
895 |
-
"rewards/concensus_correctness_reward_func":
|
896 |
-
"rewards/consensus_reward_func":
|
897 |
"rewards/cumulative_reward_2": 0.0,
|
898 |
-
"rewards/final_correctness_reward_func": 1.
|
899 |
-
"rewards/question_recreation_reward_func": 0.
|
900 |
"rewards/soft_format_reward_func": 0.0,
|
901 |
"rewards/strict_format_reward_func": 0.46875,
|
902 |
-
"rewards/xmlcount_reward_func": 1.
|
903 |
"step": 94
|
904 |
},
|
905 |
{
|
906 |
-
"completion_length":
|
907 |
"epoch": 5.3428571428571425,
|
908 |
-
"grad_norm":
|
909 |
-
"kl":
|
910 |
"learning_rate": 3.2708228165273244e-09,
|
911 |
-
"loss": 0.
|
912 |
-
"reward":
|
913 |
-
"reward_std": 0.
|
914 |
-
"rewards/concensus_correctness_reward_func":
|
915 |
"rewards/consensus_reward_func": 2.0,
|
916 |
"rewards/cumulative_reward_2": 0.0,
|
917 |
-
"rewards/final_correctness_reward_func": 1.
|
918 |
-
"rewards/question_recreation_reward_func": 0.
|
919 |
"rewards/soft_format_reward_func": 0.0,
|
920 |
-
"rewards/strict_format_reward_func": 0.
|
921 |
"rewards/xmlcount_reward_func": 1.25,
|
922 |
"step": 96
|
923 |
},
|
924 |
{
|
925 |
-
"completion_length":
|
926 |
"epoch": 5.457142857142857,
|
927 |
-
"grad_norm": 2.
|
928 |
-
"kl":
|
929 |
"learning_rate": 1.1791447083465133e-09,
|
930 |
-
"loss": 0.
|
931 |
-
"reward": 21.
|
932 |
-
"reward_std": 0.
|
933 |
"rewards/concensus_correctness_reward_func": 15.0,
|
934 |
-
"rewards/consensus_reward_func":
|
935 |
"rewards/cumulative_reward_2": 0.0,
|
936 |
-
"rewards/final_correctness_reward_func": 1.
|
937 |
-
"rewards/question_recreation_reward_func": 0.
|
938 |
"rewards/soft_format_reward_func": 0.0,
|
939 |
-
"rewards/strict_format_reward_func": 0.
|
940 |
-
"rewards/xmlcount_reward_func": 1.
|
941 |
"step": 98
|
942 |
},
|
943 |
{
|
944 |
-
"completion_length":
|
945 |
"epoch": 5.571428571428571,
|
946 |
-
"grad_norm":
|
947 |
-
"kl": 0.
|
948 |
"learning_rate": 1.3110773862126667e-10,
|
949 |
-
"loss": 0.
|
950 |
-
"reward": 26.
|
951 |
-
"reward_std": 0.
|
952 |
"rewards/concensus_correctness_reward_func": 20.0,
|
953 |
"rewards/consensus_reward_func": 2.0,
|
954 |
"rewards/cumulative_reward_2": 0.0,
|
955 |
-
"rewards/final_correctness_reward_func": 1.
|
956 |
-
"rewards/question_recreation_reward_func": 0.
|
957 |
"rewards/soft_format_reward_func": 0.0,
|
958 |
-
"rewards/strict_format_reward_func": 0.
|
959 |
-
"rewards/xmlcount_reward_func": 1.
|
960 |
"step": 100
|
961 |
},
|
962 |
{
|
963 |
"epoch": 5.571428571428571,
|
964 |
"step": 100,
|
965 |
"total_flos": 0.0,
|
966 |
-
"train_loss":
|
967 |
-
"train_runtime":
|
968 |
-
"train_samples_per_second": 1.
|
969 |
"train_steps_per_second": 0.067
|
970 |
}
|
971 |
],
|
|
|
10 |
"is_world_process_zero": true,
|
11 |
"log_history": [
|
12 |
{
|
13 |
+
"completion_length": 258.78125,
|
14 |
"epoch": 0.11428571428571428,
|
15 |
+
"grad_norm": 2.054772138595581,
|
16 |
"kl": 0.0,
|
17 |
"learning_rate": 1.6666666666666665e-07,
|
18 |
"loss": 0.0,
|
19 |
+
"reward": 17.02327122539282,
|
20 |
+
"reward_std": 4.40558909252286,
|
21 |
+
"rewards/concensus_correctness_reward_func": 12.5,
|
22 |
+
"rewards/consensus_reward_func": 1.25,
|
23 |
"rewards/cumulative_reward_2": 0.0,
|
24 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
25 |
+
"rewards/question_recreation_reward_func": 0.7546773618087173,
|
26 |
"rewards/soft_format_reward_func": 0.0,
|
27 |
"rewards/strict_format_reward_func": 0.25,
|
28 |
+
"rewards/xmlcount_reward_func": 0.8310937508940697,
|
29 |
"step": 2
|
30 |
},
|
31 |
{
|
32 |
+
"completion_length": 279.90625,
|
33 |
"epoch": 0.22857142857142856,
|
34 |
+
"grad_norm": 6.912466049194336,
|
35 |
+
"kl": 0.01094070820545312,
|
36 |
"learning_rate": 5e-07,
|
37 |
+
"loss": 0.0,
|
38 |
+
"reward": 23.899269431829453,
|
39 |
+
"reward_std": 0.04921933158766478,
|
40 |
+
"rewards/concensus_correctness_reward_func": 17.5,
|
41 |
+
"rewards/consensus_reward_func": 2.0,
|
42 |
"rewards/cumulative_reward_2": 0.0,
|
43 |
+
"rewards/final_correctness_reward_func": 1.75,
|
44 |
+
"rewards/question_recreation_reward_func": 0.9344256119802594,
|
45 |
"rewards/soft_format_reward_func": 0.0,
|
46 |
"rewards/strict_format_reward_func": 0.484375,
|
47 |
+
"rewards/xmlcount_reward_func": 1.23046875,
|
48 |
"step": 4
|
49 |
},
|
50 |
{
|
51 |
+
"completion_length": 318.5625,
|
52 |
"epoch": 0.34285714285714286,
|
53 |
+
"grad_norm": 5.637369632720947,
|
54 |
+
"kl": 83747384.09356707,
|
55 |
"learning_rate": 4.994757065594279e-07,
|
56 |
+
"loss": 83747.3828,
|
57 |
+
"reward": 18.383561819791794,
|
58 |
+
"reward_std": 0.45949314400786534,
|
59 |
+
"rewards/concensus_correctness_reward_func": 12.5,
|
60 |
+
"rewards/consensus_reward_func": 1.8125,
|
61 |
"rewards/cumulative_reward_2": 0.0,
|
62 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
63 |
+
"rewards/question_recreation_reward_func": 0.9421556144952774,
|
64 |
"rewards/soft_format_reward_func": 0.0,
|
65 |
"rewards/strict_format_reward_func": 0.453125,
|
66 |
+
"rewards/xmlcount_reward_func": 1.23828125,
|
67 |
"step": 6
|
68 |
},
|
69 |
{
|
70 |
+
"completion_length": 293.25,
|
71 |
"epoch": 0.45714285714285713,
|
72 |
+
"grad_norm": 6.48016881942749,
|
73 |
+
"kl": 0.10146056092344224,
|
74 |
"learning_rate": 4.979050253066063e-07,
|
75 |
+
"loss": 0.0001,
|
76 |
+
"reward": 24.90006184577942,
|
77 |
+
"reward_std": 0.250804515555501,
|
78 |
+
"rewards/concensus_correctness_reward_func": 18.75,
|
79 |
"rewards/consensus_reward_func": 2.0,
|
80 |
"rewards/cumulative_reward_2": 0.0,
|
81 |
+
"rewards/final_correctness_reward_func": 1.5,
|
82 |
+
"rewards/question_recreation_reward_func": 0.9156866073608398,
|
83 |
"rewards/soft_format_reward_func": 0.0,
|
84 |
+
"rewards/strict_format_reward_func": 0.484375,
|
85 |
"rewards/xmlcount_reward_func": 1.25,
|
86 |
"step": 8
|
87 |
},
|
88 |
{
|
89 |
+
"completion_length": 294.90625,
|
90 |
"epoch": 0.5714285714285714,
|
91 |
+
"grad_norm": 16.164697647094727,
|
92 |
+
"kl": 0.09648153139278293,
|
93 |
"learning_rate": 4.952945442245597e-07,
|
94 |
+
"loss": 0.0001,
|
95 |
+
"reward": 25.115739941596985,
|
96 |
+
"reward_std": 0.14409029902890325,
|
97 |
"rewards/concensus_correctness_reward_func": 18.75,
|
98 |
"rewards/consensus_reward_func": 2.0,
|
99 |
"rewards/cumulative_reward_2": 0.0,
|
100 |
+
"rewards/final_correctness_reward_func": 1.6875,
|
101 |
+
"rewards/question_recreation_reward_func": 0.9477710761129856,
|
102 |
"rewards/soft_format_reward_func": 0.0,
|
103 |
+
"rewards/strict_format_reward_func": 0.484375,
|
104 |
+
"rewards/xmlcount_reward_func": 1.24609375,
|
105 |
"step": 10
|
106 |
},
|
107 |
{
|
108 |
+
"completion_length": 285.53125,
|
109 |
"epoch": 0.6857142857142857,
|
110 |
+
"grad_norm": 0.182708278298378,
|
111 |
+
"kl": 0.16675707220565528,
|
112 |
"learning_rate": 4.916552125781528e-07,
|
113 |
+
"loss": 0.0002,
|
114 |
+
"reward": 26.597843527793884,
|
115 |
+
"reward_std": 0.002685512910829857,
|
116 |
+
"rewards/concensus_correctness_reward_func": 20.0,
|
117 |
"rewards/consensus_reward_func": 2.0,
|
118 |
"rewards/cumulative_reward_2": 0.0,
|
119 |
+
"rewards/final_correctness_reward_func": 1.875,
|
120 |
+
"rewards/question_recreation_reward_func": 0.9728433899581432,
|
121 |
"rewards/soft_format_reward_func": 0.0,
|
122 |
"rewards/strict_format_reward_func": 0.5,
|
123 |
"rewards/xmlcount_reward_func": 1.25,
|
124 |
"step": 12
|
125 |
},
|
126 |
{
|
127 |
+
"completion_length": 308.71875,
|
128 |
"epoch": 0.8,
|
129 |
+
"grad_norm": 6.084612846374512,
|
130 |
+
"kl": 0.21913069766014814,
|
131 |
"learning_rate": 4.870022949890676e-07,
|
132 |
+
"loss": 0.0002,
|
133 |
+
"reward": 21.113146513700485,
|
134 |
+
"reward_std": 0.13240044617487,
|
135 |
+
"rewards/concensus_correctness_reward_func": 15.0,
|
136 |
"rewards/consensus_reward_func": 2.0,
|
137 |
"rewards/cumulative_reward_2": 0.0,
|
138 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
139 |
+
"rewards/question_recreation_reward_func": 0.9442716278135777,
|
140 |
"rewards/soft_format_reward_func": 0.0,
|
141 |
"rewards/strict_format_reward_func": 0.484375,
|
142 |
+
"rewards/xmlcount_reward_func": 1.2470000013709068,
|
143 |
"step": 14
|
144 |
},
|
145 |
{
|
146 |
+
"completion_length": 333.25,
|
147 |
"epoch": 0.9142857142857143,
|
148 |
+
"grad_norm": 6157868.5,
|
149 |
+
"kl": 288729.23049705196,
|
150 |
"learning_rate": 4.81355307410676e-07,
|
151 |
+
"loss": 288.7293,
|
152 |
+
"reward": 19.089906871318817,
|
153 |
+
"reward_std": 1.0611610219348222,
|
154 |
+
"rewards/concensus_correctness_reward_func": 13.125,
|
155 |
+
"rewards/consensus_reward_func": 1.8125,
|
156 |
"rewards/cumulative_reward_2": 0.0,
|
157 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
158 |
+
"rewards/question_recreation_reward_func": 0.9974069446325302,
|
159 |
"rewards/soft_format_reward_func": 0.0,
|
160 |
+
"rewards/strict_format_reward_func": 0.46875,
|
161 |
+
"rewards/xmlcount_reward_func": 1.248750001192093,
|
162 |
"step": 16
|
163 |
},
|
164 |
{
|
165 |
+
"completion_length": 285.5,
|
166 |
"epoch": 1.0,
|
167 |
+
"grad_norm": 78.33558654785156,
|
168 |
+
"kl": 4402.47263197725,
|
169 |
"learning_rate": 4.747379352713488e-07,
|
170 |
+
"loss": 3.3019,
|
171 |
+
"reward": 20.285069465637207,
|
172 |
+
"reward_std": 1.364617843956997,
|
173 |
+
"rewards/concensus_correctness_reward_func": 14.166666666666666,
|
174 |
+
"rewards/consensus_reward_func": 1.9166666666666667,
|
175 |
"rewards/cumulative_reward_2": 0.0,
|
176 |
+
"rewards/final_correctness_reward_func": 1.5,
|
177 |
+
"rewards/question_recreation_reward_func": 0.9986111124356588,
|
178 |
"rewards/soft_format_reward_func": 0.0,
|
179 |
"rewards/strict_format_reward_func": 0.4583333333333333,
|
180 |
+
"rewards/xmlcount_reward_func": 1.2447916666666667,
|
181 |
"step": 18
|
182 |
},
|
183 |
{
|
184 |
+
"completion_length": 302.59375,
|
185 |
"epoch": 1.1142857142857143,
|
186 |
+
"grad_norm": 3.6053833961486816,
|
187 |
+
"kl": 0.13355864235199988,
|
188 |
"learning_rate": 4.6717793412953776e-07,
|
189 |
+
"loss": 0.0001,
|
190 |
+
"reward": 19.868769019842148,
|
191 |
+
"reward_std": 0.008435817871941254,
|
192 |
+
"rewards/concensus_correctness_reward_func": 13.75,
|
193 |
"rewards/consensus_reward_func": 2.0,
|
194 |
"rewards/cumulative_reward_2": 0.0,
|
195 |
+
"rewards/final_correctness_reward_func": 1.375,
|
196 |
+
"rewards/question_recreation_reward_func": 0.9937690384685993,
|
197 |
"rewards/soft_format_reward_func": 0.0,
|
198 |
+
"rewards/strict_format_reward_func": 0.5,
|
199 |
+
"rewards/xmlcount_reward_func": 1.25,
|
200 |
"step": 20
|
201 |
},
|
202 |
{
|
203 |
+
"completion_length": 298.9375,
|
204 |
"epoch": 1.2285714285714286,
|
205 |
+
"grad_norm": 4.034364700317383,
|
206 |
+
"kl": 0.284994178917259,
|
207 |
"learning_rate": 4.5870701325731773e-07,
|
208 |
+
"loss": 0.0003,
|
209 |
+
"reward": 25.217251539230347,
|
210 |
+
"reward_std": 0.1768817222182406,
|
211 |
"rewards/concensus_correctness_reward_func": 18.75,
|
212 |
"rewards/consensus_reward_func": 2.0,
|
213 |
"rewards/cumulative_reward_2": 0.0,
|
214 |
+
"rewards/final_correctness_reward_func": 1.75,
|
215 |
+
"rewards/question_recreation_reward_func": 0.9985014833509922,
|
216 |
"rewards/soft_format_reward_func": 0.0,
|
217 |
+
"rewards/strict_format_reward_func": 0.46875,
|
218 |
+
"rewards/xmlcount_reward_func": 1.25,
|
219 |
"step": 22
|
220 |
},
|
221 |
{
|
222 |
+
"completion_length": 295.875,
|
223 |
"epoch": 1.342857142857143,
|
224 |
+
"grad_norm": 19.670671463012695,
|
225 |
+
"kl": 93.65684759290889,
|
226 |
"learning_rate": 4.4936070264068016e-07,
|
227 |
+
"loss": 0.0937,
|
228 |
+
"reward": 25.219618678092957,
|
229 |
+
"reward_std": 0.2108936388976872,
|
230 |
+
"rewards/concensus_correctness_reward_func": 18.75,
|
231 |
"rewards/consensus_reward_func": 2.0,
|
232 |
"rewards/cumulative_reward_2": 0.0,
|
233 |
"rewards/final_correctness_reward_func": 1.75,
|
234 |
+
"rewards/question_recreation_reward_func": 0.9891498424112797,
|
235 |
"rewards/soft_format_reward_func": 0.0,
|
236 |
+
"rewards/strict_format_reward_func": 0.484375,
|
237 |
+
"rewards/xmlcount_reward_func": 1.24609375,
|
238 |
"step": 24
|
239 |
},
|
240 |
{
|
241 |
+
"completion_length": 289.5,
|
242 |
"epoch": 1.457142857142857,
|
243 |
+
"grad_norm": 118071795712.0,
|
244 |
+
"kl": 14642561024.712172,
|
245 |
"learning_rate": 4.391782039544238e-07,
|
246 |
+
"loss": 14642560.0,
|
247 |
+
"reward": 23.393190026283264,
|
248 |
+
"reward_std": 1.1594447159441188,
|
249 |
+
"rewards/concensus_correctness_reward_func": 16.875,
|
250 |
+
"rewards/consensus_reward_func": 1.9375,
|
251 |
"rewards/cumulative_reward_2": 0.0,
|
252 |
+
"rewards/final_correctness_reward_func": 1.875,
|
253 |
+
"rewards/question_recreation_reward_func": 0.987252376973629,
|
254 |
"rewards/soft_format_reward_func": 0.0,
|
255 |
+
"rewards/strict_format_reward_func": 0.46875,
|
256 |
+
"rewards/xmlcount_reward_func": 1.2496875077486038,
|
257 |
"step": 26
|
258 |
},
|
259 |
{
|
260 |
+
"completion_length": 290.4375,
|
261 |
"epoch": 1.5714285714285714,
|
262 |
+
"grad_norm": 985.9644165039062,
|
263 |
+
"kl": 47.3938269498758,
|
264 |
"learning_rate": 4.282022261367073e-07,
|
265 |
+
"loss": 0.0474,
|
266 |
+
"reward": 20.982531785964966,
|
267 |
+
"reward_std": 0.09451750945299864,
|
268 |
"rewards/concensus_correctness_reward_func": 15.0,
|
269 |
+
"rewards/consensus_reward_func": 1.875,
|
270 |
"rewards/cumulative_reward_2": 0.0,
|
271 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
272 |
+
"rewards/question_recreation_reward_func": 0.9512817114591599,
|
273 |
"rewards/soft_format_reward_func": 0.0,
|
274 |
+
"rewards/strict_format_reward_func": 0.46875,
|
275 |
+
"rewards/xmlcount_reward_func": 1.25,
|
276 |
"step": 28
|
277 |
},
|
278 |
{
|
279 |
+
"completion_length": 291.75,
|
280 |
"epoch": 1.6857142857142857,
|
281 |
+
"grad_norm": 14.219393730163574,
|
282 |
+
"kl": 5697.919686453417,
|
283 |
"learning_rate": 4.1647880625292027e-07,
|
284 |
+
"loss": 5.6979,
|
285 |
+
"reward": 24.938138127326965,
|
286 |
+
"reward_std": 0.5503595303162001,
|
287 |
+
"rewards/concensus_correctness_reward_func": 18.75,
|
288 |
"rewards/consensus_reward_func": 2.0,
|
289 |
"rewards/cumulative_reward_2": 0.0,
|
290 |
+
"rewards/final_correctness_reward_func": 1.5625,
|
291 |
+
"rewards/question_recreation_reward_func": 0.9167943075299263,
|
292 |
"rewards/soft_format_reward_func": 0.0,
|
293 |
+
"rewards/strict_format_reward_func": 0.484375,
|
294 |
+
"rewards/xmlcount_reward_func": 1.2244687490165234,
|
295 |
"step": 30
|
296 |
},
|
297 |
{
|
298 |
+
"completion_length": 320.78125,
|
299 |
"epoch": 1.8,
|
300 |
+
"grad_norm": 111691.4140625,
|
301 |
+
"kl": 9744.625750856474,
|
302 |
"learning_rate": 4.040571164002318e-07,
|
303 |
+
"loss": 9.7446,
|
304 |
+
"reward": 23.627057909965515,
|
305 |
+
"reward_std": 0.3214102545171045,
|
306 |
+
"rewards/concensus_correctness_reward_func": 17.5,
|
307 |
+
"rewards/consensus_reward_func": 1.875,
|
308 |
"rewards/cumulative_reward_2": 0.0,
|
309 |
+
"rewards/final_correctness_reward_func": 1.6875,
|
310 |
+
"rewards/question_recreation_reward_func": 0.9628704935312271,
|
311 |
"rewards/soft_format_reward_func": 0.0,
|
312 |
"rewards/strict_format_reward_func": 0.46875,
|
313 |
+
"rewards/xmlcount_reward_func": 1.1329374983906746,
|
314 |
"step": 32
|
315 |
},
|
316 |
{
|
317 |
+
"completion_length": 288.90625,
|
318 |
"epoch": 1.9142857142857141,
|
319 |
+
"grad_norm": 5.710204124450684,
|
320 |
+
"kl": 35.09870433434844,
|
321 |
"learning_rate": 3.909892574627266e-07,
|
322 |
+
"loss": 0.0351,
|
323 |
+
"reward": 19.93251895904541,
|
324 |
+
"reward_std": 0.09295251258299686,
|
325 |
+
"rewards/concensus_correctness_reward_func": 13.75,
|
326 |
"rewards/consensus_reward_func": 2.0,
|
327 |
"rewards/cumulative_reward_2": 0.0,
|
328 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
329 |
+
"rewards/question_recreation_reward_func": 0.9950189106166363,
|
330 |
"rewards/soft_format_reward_func": 0.0,
|
331 |
"rewards/strict_format_reward_func": 0.5,
|
332 |
"rewards/xmlcount_reward_func": 1.25,
|
333 |
"step": 34
|
334 |
},
|
335 |
{
|
336 |
+
"completion_length": 312.9583333333333,
|
337 |
"epoch": 2.0,
|
338 |
+
"grad_norm": 186.2655029296875,
|
339 |
+
"kl": 8339.249363450954,
|
340 |
"learning_rate": 3.773300405821908e-07,
|
341 |
+
"loss": 6.2544,
|
342 |
+
"reward": 22.710465768973034,
|
343 |
+
"reward_std": 0.40588310376430553,
|
344 |
"rewards/concensus_correctness_reward_func": 16.666666666666668,
|
345 |
"rewards/consensus_reward_func": 2.0,
|
346 |
"rewards/cumulative_reward_2": 0.0,
|
347 |
+
"rewards/final_correctness_reward_func": 1.5,
|
348 |
+
"rewards/question_recreation_reward_func": 0.8562991668780645,
|
349 |
"rewards/soft_format_reward_func": 0.0,
|
350 |
+
"rewards/strict_format_reward_func": 0.4583333333333333,
|
351 |
+
"rewards/xmlcount_reward_func": 1.2291666666666667,
|
352 |
"step": 36
|
353 |
},
|
354 |
{
|
355 |
+
"completion_length": 286.53125,
|
356 |
"epoch": 2.1142857142857143,
|
357 |
+
"grad_norm": 10.10280704498291,
|
358 |
+
"kl": 1.4235679027624428,
|
359 |
"learning_rate": 3.6313675726113475e-07,
|
360 |
+
"loss": 0.0014,
|
361 |
+
"reward": 25.085916996002197,
|
362 |
+
"reward_std": 0.01389117946382612,
|
363 |
+
"rewards/concensus_correctness_reward_func": 18.75,
|
364 |
"rewards/consensus_reward_func": 2.0,
|
365 |
"rewards/cumulative_reward_2": 0.0,
|
366 |
+
"rewards/final_correctness_reward_func": 1.625,
|
367 |
+
"rewards/question_recreation_reward_func": 0.9609169252216816,
|
368 |
"rewards/soft_format_reward_func": 0.0,
|
369 |
"rewards/strict_format_reward_func": 0.5,
|
370 |
"rewards/xmlcount_reward_func": 1.25,
|
371 |
"step": 38
|
372 |
},
|
373 |
{
|
374 |
+
"completion_length": 307.84375,
|
375 |
"epoch": 2.2285714285714286,
|
376 |
+
"grad_norm": 301.5379333496094,
|
377 |
+
"kl": 31.047572038602084,
|
378 |
"learning_rate": 3.484689390623218e-07,
|
379 |
+
"loss": 0.031,
|
380 |
+
"reward": 24.409271508455276,
|
381 |
+
"reward_std": 1.3443972589448094,
|
382 |
+
"rewards/concensus_correctness_reward_func": 18.125,
|
383 |
+
"rewards/consensus_reward_func": 1.9375,
|
384 |
"rewards/cumulative_reward_2": 0.0,
|
385 |
+
"rewards/final_correctness_reward_func": 1.75,
|
386 |
+
"rewards/question_recreation_reward_func": 0.9250839278101921,
|
387 |
"rewards/soft_format_reward_func": 0.0,
|
388 |
+
"rewards/strict_format_reward_func": 0.453125,
|
389 |
+
"rewards/xmlcount_reward_func": 1.2185625061392784,
|
390 |
"step": 40
|
391 |
},
|
392 |
{
|
393 |
+
"completion_length": 321.125,
|
394 |
"epoch": 2.342857142857143,
|
395 |
+
"grad_norm": 9.44746208190918,
|
396 |
+
"kl": 99998651.09828581,
|
397 |
"learning_rate": 3.3338810791270517e-07,
|
398 |
+
"loss": 99998.6562,
|
399 |
+
"reward": 19.91300740838051,
|
400 |
+
"reward_std": 0.06640663610232878,
|
401 |
"rewards/concensus_correctness_reward_func": 13.75,
|
402 |
"rewards/consensus_reward_func": 2.0,
|
403 |
"rewards/cumulative_reward_2": 0.0,
|
404 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
405 |
+
"rewards/question_recreation_reward_func": 0.9950386732816696,
|
406 |
"rewards/soft_format_reward_func": 0.0,
|
407 |
+
"rewards/strict_format_reward_func": 0.484375,
|
408 |
+
"rewards/xmlcount_reward_func": 1.24609375,
|
409 |
"step": 42
|
410 |
},
|
411 |
{
|
412 |
+
"completion_length": 297.0,
|
413 |
"epoch": 2.4571428571428573,
|
414 |
+
"grad_norm": 4.234458923339844,
|
415 |
+
"kl": 26.058129234821536,
|
416 |
"learning_rate": 3.179575180590857e-07,
|
417 |
+
"loss": 0.0261,
|
418 |
+
"reward": 20.36888739466667,
|
419 |
+
"reward_std": 1.2362263132818043,
|
420 |
+
"rewards/concensus_correctness_reward_func": 14.375,
|
421 |
+
"rewards/consensus_reward_func": 1.9375,
|
422 |
"rewards/cumulative_reward_2": 0.0,
|
423 |
"rewards/final_correctness_reward_func": 1.375,
|
424 |
+
"rewards/question_recreation_reward_func": 0.9313873276114464,
|
425 |
"rewards/soft_format_reward_func": 0.0,
|
426 |
"rewards/strict_format_reward_func": 0.5,
|
427 |
"rewards/xmlcount_reward_func": 1.25,
|
428 |
"step": 44
|
429 |
},
|
430 |
{
|
431 |
+
"completion_length": 293.375,
|
432 |
"epoch": 2.571428571428571,
|
433 |
+
"grad_norm": 3.131242513656616,
|
434 |
+
"kl": 0.3580109770409763,
|
435 |
"learning_rate": 3.022418907578188e-07,
|
436 |
+
"loss": 0.0004,
|
437 |
+
"reward": 22.51879781484604,
|
438 |
+
"reward_std": 0.1030593290925026,
|
439 |
"rewards/concensus_correctness_reward_func": 16.25,
|
440 |
+
"rewards/consensus_reward_func": 1.875,
|
441 |
"rewards/cumulative_reward_2": 0.0,
|
442 |
+
"rewards/final_correctness_reward_func": 1.6875,
|
443 |
+
"rewards/question_recreation_reward_func": 0.9875477105379105,
|
444 |
"rewards/soft_format_reward_func": 0.0,
|
445 |
+
"rewards/strict_format_reward_func": 0.46875,
|
446 |
"rewards/xmlcount_reward_func": 1.25,
|
447 |
"step": 46
|
448 |
},
|
449 |
{
|
450 |
+
"completion_length": 303.6875,
|
451 |
"epoch": 2.685714285714286,
|
452 |
+
"grad_norm": 7157400.5,
|
453 |
+
"kl": 723253.104146285,
|
454 |
"learning_rate": 2.863071428113726e-07,
|
455 |
+
"loss": 723.2531,
|
456 |
+
"reward": 21.64236208796501,
|
457 |
+
"reward_std": 1.3292767723178258,
|
458 |
"rewards/concensus_correctness_reward_func": 15.625,
|
459 |
"rewards/consensus_reward_func": 1.9375,
|
460 |
"rewards/cumulative_reward_2": 0.0,
|
461 |
+
"rewards/final_correctness_reward_func": 1.375,
|
462 |
+
"rewards/question_recreation_reward_func": 0.9548620991408825,
|
463 |
"rewards/soft_format_reward_func": 0.0,
|
464 |
"rewards/strict_format_reward_func": 0.5,
|
465 |
+
"rewards/xmlcount_reward_func": 1.25,
|
466 |
"step": 48
|
467 |
},
|
468 |
{
|
469 |
+
"completion_length": 288.4375,
|
470 |
"epoch": 2.8,
|
471 |
+
"grad_norm": 3.812501907348633,
|
472 |
+
"kl": 0.1941389513667673,
|
473 |
"learning_rate": 2.7022011009035107e-07,
|
474 |
+
"loss": 0.0002,
|
475 |
+
"reward": 23.754722118377686,
|
476 |
+
"reward_std": 0.13451411115238443,
|
477 |
"rewards/concensus_correctness_reward_func": 17.5,
|
478 |
+
"rewards/consensus_reward_func": 1.875,
|
479 |
"rewards/cumulative_reward_2": 0.0,
|
480 |
+
"rewards/final_correctness_reward_func": 1.6875,
|
481 |
+
"rewards/question_recreation_reward_func": 0.9734721444547176,
|
482 |
"rewards/soft_format_reward_func": 0.0,
|
483 |
"rewards/strict_format_reward_func": 0.46875,
|
484 |
"rewards/xmlcount_reward_func": 1.25,
|
485 |
"step": 50
|
486 |
},
|
487 |
{
|
488 |
+
"completion_length": 273.25,
|
489 |
"epoch": 2.914285714285714,
|
490 |
+
"grad_norm": 64386.68359375,
|
491 |
+
"kl": 4703.566962761339,
|
492 |
"learning_rate": 2.540482672006254e-07,
|
493 |
+
"loss": 4.7036,
|
494 |
+
"reward": 23.86061406135559,
|
495 |
+
"reward_std": 0.11497652903926792,
|
496 |
+
"rewards/concensus_correctness_reward_func": 17.5,
|
497 |
"rewards/consensus_reward_func": 2.0,
|
498 |
"rewards/cumulative_reward_2": 0.0,
|
499 |
+
"rewards/final_correctness_reward_func": 1.6875,
|
500 |
+
"rewards/question_recreation_reward_func": 0.9387390580959618,
|
501 |
"rewards/soft_format_reward_func": 0.0,
|
502 |
+
"rewards/strict_format_reward_func": 0.484375,
|
503 |
+
"rewards/xmlcount_reward_func": 1.25,
|
504 |
"step": 52
|
505 |
},
|
506 |
{
|
507 |
+
"completion_length": 283.375,
|
508 |
"epoch": 3.0,
|
509 |
+
"grad_norm": 4.72637939453125,
|
510 |
+
"kl": 0.41162373684346676,
|
511 |
"learning_rate": 2.37859444471388e-07,
|
512 |
+
"loss": 0.0003,
|
513 |
+
"reward": 22.2163348197937,
|
514 |
+
"reward_std": 1.4251218068723877,
|
515 |
+
"rewards/concensus_correctness_reward_func": 15.833333333333334,
|
516 |
+
"rewards/consensus_reward_func": 1.9166666666666667,
|
517 |
"rewards/cumulative_reward_2": 0.0,
|
518 |
+
"rewards/final_correctness_reward_func": 1.75,
|
519 |
+
"rewards/question_recreation_reward_func": 0.9663348992665609,
|
520 |
"rewards/soft_format_reward_func": 0.0,
|
521 |
"rewards/strict_format_reward_func": 0.5,
|
522 |
"rewards/xmlcount_reward_func": 1.25,
|
523 |
"step": 54
|
524 |
},
|
525 |
{
|
526 |
+
"completion_length": 288.875,
|
527 |
"epoch": 3.1142857142857143,
|
528 |
+
"grad_norm": 8.240327835083008,
|
529 |
+
"kl": 0.11788215674459934,
|
530 |
"learning_rate": 2.2172154345117894e-07,
|
531 |
+
"loss": 0.0001,
|
532 |
+
"reward": 23.904327988624573,
|
533 |
+
"reward_std": 0.30620702836313285,
|
534 |
"rewards/concensus_correctness_reward_func": 17.5,
|
535 |
"rewards/consensus_reward_func": 2.0,
|
536 |
"rewards/cumulative_reward_2": 0.0,
|
537 |
"rewards/final_correctness_reward_func": 1.6875,
|
538 |
+
"rewards/question_recreation_reward_func": 0.9668280705809593,
|
539 |
"rewards/soft_format_reward_func": 0.0,
|
540 |
"rewards/strict_format_reward_func": 0.5,
|
541 |
"rewards/xmlcount_reward_func": 1.25,
|
542 |
"step": 56
|
543 |
},
|
544 |
{
|
545 |
+
"completion_length": 286.1875,
|
546 |
"epoch": 3.2285714285714286,
|
547 |
+
"grad_norm": 5.542027950286865,
|
548 |
+
"kl": 0.11034615035168827,
|
549 |
"learning_rate": 2.0570225210519433e-07,
|
550 |
+
"loss": 0.0001,
|
551 |
+
"reward": 21.113164007663727,
|
552 |
+
"reward_std": 0.19153568047477165,
|
553 |
"rewards/concensus_correctness_reward_func": 15.0,
|
554 |
"rewards/consensus_reward_func": 2.0,
|
555 |
"rewards/cumulative_reward_2": 0.0,
|
556 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
557 |
+
"rewards/question_recreation_reward_func": 0.941288948059082,
|
558 |
"rewards/soft_format_reward_func": 0.0,
|
559 |
+
"rewards/strict_format_reward_func": 0.484375,
|
560 |
+
"rewards/xmlcount_reward_func": 1.25,
|
561 |
"step": 58
|
562 |
},
|
563 |
{
|
564 |
+
"completion_length": 293.96875,
|
565 |
"epoch": 3.342857142857143,
|
566 |
+
"grad_norm": 89873.234375,
|
567 |
+
"kl": 15506.7046823774,
|
568 |
"learning_rate": 1.8986876090843664e-07,
|
569 |
+
"loss": 15.5067,
|
570 |
+
"reward": 24.321931093931198,
|
571 |
+
"reward_std": 0.9067758475139271,
|
572 |
+
"rewards/concensus_correctness_reward_func": 18.238062500953674,
|
573 |
+
"rewards/consensus_reward_func": 1.9375,
|
574 |
"rewards/cumulative_reward_2": 0.0,
|
575 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
576 |
+
"rewards/question_recreation_reward_func": 0.9588685538619757,
|
577 |
"rewards/soft_format_reward_func": 0.0,
|
578 |
+
"rewards/strict_format_reward_func": 0.5,
|
579 |
"rewards/xmlcount_reward_func": 1.25,
|
580 |
"step": 60
|
581 |
},
|
582 |
{
|
583 |
+
"completion_length": 292.65625,
|
584 |
"epoch": 3.4571428571428573,
|
585 |
+
"grad_norm": 1.29377019405365,
|
586 |
+
"kl": 2.4666087054647505,
|
587 |
"learning_rate": 1.7428748102551234e-07,
|
588 |
+
"loss": 0.0025,
|
589 |
+
"reward": 21.342330873012543,
|
590 |
+
"reward_std": 0.17704137448163237,
|
591 |
"rewards/concensus_correctness_reward_func": 15.0,
|
592 |
+
"rewards/consensus_reward_func": 1.875,
|
593 |
"rewards/cumulative_reward_2": 0.0,
|
594 |
+
"rewards/final_correctness_reward_func": 1.75,
|
595 |
+
"rewards/question_recreation_reward_func": 0.9985808059573174,
|
596 |
"rewards/soft_format_reward_func": 0.0,
|
597 |
+
"rewards/strict_format_reward_func": 0.46875,
|
598 |
"rewards/xmlcount_reward_func": 1.25,
|
599 |
"step": 62
|
600 |
},
|
601 |
{
|
602 |
+
"completion_length": 296.125,
|
603 |
"epoch": 3.571428571428571,
|
604 |
+
"grad_norm": 11.036781311035156,
|
605 |
+
"kl": 1.3272000113502145,
|
606 |
"learning_rate": 1.5902376575912814e-07,
|
607 |
+
"loss": 0.0013,
|
608 |
+
"reward": 22.402726590633392,
|
609 |
+
"reward_std": 0.0906209279637551,
|
610 |
"rewards/concensus_correctness_reward_func": 16.25,
|
611 |
"rewards/consensus_reward_func": 2.0,
|
612 |
"rewards/cumulative_reward_2": 0.0,
|
613 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
614 |
+
"rewards/question_recreation_reward_func": 0.9652266465127468,
|
615 |
"rewards/soft_format_reward_func": 0.0,
|
616 |
"rewards/strict_format_reward_func": 0.5,
|
617 |
"rewards/xmlcount_reward_func": 1.25,
|
618 |
"step": 64
|
619 |
},
|
620 |
{
|
621 |
+
"completion_length": 311.25,
|
622 |
"epoch": 3.685714285714286,
|
623 |
+
"grad_norm": 3.4472224712371826,
|
624 |
+
"kl": 0.8945652563124895,
|
625 |
"learning_rate": 1.4414163643562753e-07,
|
626 |
+
"loss": 0.0009,
|
627 |
+
"reward": 22.61855262517929,
|
628 |
+
"reward_std": 0.18265954998787493,
|
629 |
"rewards/concensus_correctness_reward_func": 16.25,
|
630 |
"rewards/consensus_reward_func": 2.0,
|
631 |
"rewards/cumulative_reward_2": 0.0,
|
632 |
+
"rewards/final_correctness_reward_func": 1.625,
|
633 |
+
"rewards/question_recreation_reward_func": 0.9935526624321938,
|
634 |
"rewards/soft_format_reward_func": 0.0,
|
635 |
"rewards/strict_format_reward_func": 0.5,
|
636 |
"rewards/xmlcount_reward_func": 1.25,
|
637 |
"step": 66
|
638 |
},
|
639 |
{
|
640 |
+
"completion_length": 322.75,
|
641 |
"epoch": 3.8,
|
642 |
+
"grad_norm": 6.844115734100342,
|
643 |
+
"kl": 3.179185423068702,
|
644 |
"learning_rate": 1.2970351387729872e-07,
|
645 |
+
"loss": 0.0032,
|
646 |
+
"reward": 21.179388463497162,
|
647 |
+
"reward_std": 0.09523822067421861,
|
648 |
+
"rewards/concensus_correctness_reward_func": 15.0,
|
649 |
"rewards/consensus_reward_func": 2.0,
|
650 |
"rewards/cumulative_reward_2": 0.0,
|
651 |
"rewards/final_correctness_reward_func": 1.4375,
|
652 |
+
"rewards/question_recreation_reward_func": 0.9918884485960007,
|
653 |
"rewards/soft_format_reward_func": 0.0,
|
654 |
+
"rewards/strict_format_reward_func": 0.5,
|
655 |
+
"rewards/xmlcount_reward_func": 1.25,
|
656 |
"step": 68
|
657 |
},
|
658 |
{
|
659 |
+
"completion_length": 280.5,
|
660 |
"epoch": 3.914285714285714,
|
661 |
+
"grad_norm": 5.8618083000183105,
|
662 |
+
"kl": 7725.971832116949,
|
663 |
"learning_rate": 1.1576995658775404e-07,
|
664 |
+
"loss": 7.726,
|
665 |
+
"reward": 24.525205433368683,
|
666 |
+
"reward_std": 1.1531662662309827,
|
667 |
+
"rewards/concensus_correctness_reward_func": 18.125,
|
668 |
+
"rewards/consensus_reward_func": 1.9375,
|
669 |
"rewards/cumulative_reward_2": 0.0,
|
670 |
+
"rewards/final_correctness_reward_func": 1.8125,
|
671 |
+
"rewards/question_recreation_reward_func": 0.935486514121294,
|
672 |
"rewards/soft_format_reward_func": 0.0,
|
673 |
"rewards/strict_format_reward_func": 0.46875,
|
674 |
+
"rewards/xmlcount_reward_func": 1.2459687516093254,
|
675 |
"step": 70
|
676 |
},
|
677 |
{
|
678 |
+
"completion_length": 295.375,
|
679 |
"epoch": 4.0,
|
680 |
+
"grad_norm": 4.205502986907959,
|
681 |
+
"kl": 5170.305763981615,
|
682 |
"learning_rate": 1.0239940674851941e-07,
|
683 |
+
"loss": 3.8777,
|
684 |
+
"reward": 24.69401526451111,
|
685 |
+
"reward_std": 0.29310186601166305,
|
686 |
"rewards/concensus_correctness_reward_func": 18.333333333333332,
|
687 |
"rewards/consensus_reward_func": 2.0,
|
688 |
"rewards/cumulative_reward_2": 0.0,
|
689 |
+
"rewards/final_correctness_reward_func": 1.6666666666666667,
|
690 |
+
"rewards/question_recreation_reward_func": 0.9440152446428934,
|
691 |
"rewards/soft_format_reward_func": 0.0,
|
692 |
+
"rewards/strict_format_reward_func": 0.5,
|
693 |
+
"rewards/xmlcount_reward_func": 1.25,
|
694 |
"step": 72
|
695 |
},
|
696 |
{
|
697 |
+
"completion_length": 322.75,
|
698 |
"epoch": 4.114285714285714,
|
699 |
+
"grad_norm": 853897.4375,
|
700 |
+
"kl": 93995.61682448094,
|
701 |
"learning_rate": 8.964794509221507e-08,
|
702 |
+
"loss": 93.9956,
|
703 |
+
"reward": 21.796031087636948,
|
704 |
+
"reward_std": 1.1461751838796772,
|
705 |
+
"rewards/concensus_correctness_reward_func": 15.625,
|
706 |
+
"rewards/consensus_reward_func": 1.9375,
|
707 |
"rewards/cumulative_reward_2": 0.0,
|
708 |
+
"rewards/final_correctness_reward_func": 1.5,
|
709 |
+
"rewards/question_recreation_reward_func": 0.9835311211645603,
|
710 |
"rewards/soft_format_reward_func": 0.0,
|
711 |
"rewards/strict_format_reward_func": 0.5,
|
712 |
"rewards/xmlcount_reward_func": 1.25,
|
713 |
"step": 74
|
714 |
},
|
715 |
{
|
716 |
+
"completion_length": 310.625,
|
717 |
"epoch": 4.228571428571429,
|
718 |
+
"grad_norm": 6.561783790588379,
|
719 |
+
"kl": 0.6868871822953224,
|
720 |
"learning_rate": 7.756905568047392e-08,
|
721 |
+
"loss": 0.0007,
|
722 |
+
"reward": 23.608940601348877,
|
723 |
+
"reward_std": 0.1481841884815367,
|
724 |
+
"rewards/concensus_correctness_reward_func": 17.5,
|
725 |
+
"rewards/consensus_reward_func": 2.0,
|
726 |
"rewards/cumulative_reward_2": 0.0,
|
727 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
728 |
+
"rewards/question_recreation_reward_func": 0.9526904746890068,
|
729 |
"rewards/soft_format_reward_func": 0.0,
|
730 |
+
"rewards/strict_format_reward_func": 0.46875,
|
731 |
+
"rewards/xmlcount_reward_func": 1.25,
|
732 |
"step": 76
|
733 |
},
|
734 |
{
|
735 |
+
"completion_length": 300.6875,
|
736 |
"epoch": 4.3428571428571425,
|
737 |
+
"grad_norm": 5.61808967590332,
|
738 |
+
"kl": 0.17497208586428314,
|
739 |
"learning_rate": 6.621340157319996e-08,
|
740 |
+
"loss": 0.0002,
|
741 |
+
"reward": 20.960274010896683,
|
742 |
+
"reward_std": 0.12753743107896298,
|
743 |
"rewards/concensus_correctness_reward_func": 15.0,
|
744 |
"rewards/consensus_reward_func": 2.0,
|
745 |
"rewards/cumulative_reward_2": 0.0,
|
746 |
+
"rewards/final_correctness_reward_func": 1.375,
|
747 |
+
"rewards/question_recreation_reward_func": 0.9174302890896797,
|
748 |
"rewards/soft_format_reward_func": 0.0,
|
749 |
"rewards/strict_format_reward_func": 0.453125,
|
750 |
+
"rewards/xmlcount_reward_func": 1.2147187516093254,
|
751 |
"step": 78
|
752 |
},
|
753 |
{
|
754 |
+
"completion_length": 269.59375,
|
755 |
"epoch": 4.457142857142857,
|
756 |
+
"grad_norm": 58.665802001953125,
|
757 |
+
"kl": 5.802749904803932,
|
758 |
"learning_rate": 5.5628612330087724e-08,
|
759 |
+
"loss": 0.0058,
|
760 |
+
"reward": 26.747634530067444,
|
761 |
+
"reward_std": 0.0016583941760472953,
|
762 |
+
"rewards/concensus_correctness_reward_func": 20.0,
|
763 |
"rewards/consensus_reward_func": 2.0,
|
764 |
"rewards/cumulative_reward_2": 0.0,
|
765 |
+
"rewards/final_correctness_reward_func": 2.0,
|
766 |
+
"rewards/question_recreation_reward_func": 0.9976345337927341,
|
767 |
"rewards/soft_format_reward_func": 0.0,
|
768 |
+
"rewards/strict_format_reward_func": 0.5,
|
769 |
+
"rewards/xmlcount_reward_func": 1.25,
|
770 |
"step": 80
|
771 |
},
|
772 |
{
|
773 |
+
"completion_length": 315.21875,
|
774 |
"epoch": 4.571428571428571,
|
775 |
+
"grad_norm": 0.19868668913841248,
|
776 |
+
"kl": 131.63193395012058,
|
777 |
"learning_rate": 4.5859084235697235e-08,
|
778 |
+
"loss": 0.1316,
|
779 |
+
"reward": 16.382025599479675,
|
780 |
+
"reward_std": 1.1490484923124313,
|
781 |
+
"rewards/concensus_correctness_reward_func": 10.625,
|
782 |
+
"rewards/consensus_reward_func": 1.8125,
|
783 |
"rewards/cumulative_reward_2": 0.0,
|
784 |
"rewards/final_correctness_reward_func": 1.25,
|
785 |
+
"rewards/question_recreation_reward_func": 0.9757755696773529,
|
786 |
"rewards/soft_format_reward_func": 0.0,
|
787 |
+
"rewards/strict_format_reward_func": 0.46875,
|
788 |
"rewards/xmlcount_reward_func": 1.25,
|
789 |
"step": 82
|
790 |
},
|
791 |
{
|
792 |
+
"completion_length": 291.96875,
|
793 |
"epoch": 4.685714285714286,
|
794 |
+
"grad_norm": 1.1480889320373535,
|
795 |
+
"kl": 0.19505003839731216,
|
796 |
"learning_rate": 3.6945794086007705e-08,
|
797 |
+
"loss": 0.0002,
|
798 |
+
"reward": 22.737855315208435,
|
799 |
+
"reward_std": 0.0,
|
800 |
"rewards/concensus_correctness_reward_func": 16.25,
|
801 |
+
"rewards/consensus_reward_func": 1.875,
|
802 |
"rewards/cumulative_reward_2": 0.0,
|
803 |
+
"rewards/final_correctness_reward_func": 1.875,
|
804 |
+
"rewards/question_recreation_reward_func": 0.9878552854061127,
|
805 |
"rewards/soft_format_reward_func": 0.0,
|
806 |
+
"rewards/strict_format_reward_func": 0.5,
|
807 |
+
"rewards/xmlcount_reward_func": 1.25,
|
808 |
"step": 84
|
809 |
},
|
810 |
{
|
811 |
+
"completion_length": 284.625,
|
812 |
"epoch": 4.8,
|
813 |
+
"grad_norm": 9.016759872436523,
|
814 |
+
"kl": 0.27029623091220856,
|
815 |
"learning_rate": 2.892612731749414e-08,
|
816 |
+
"loss": 0.0003,
|
817 |
+
"reward": 25.095996618270874,
|
818 |
+
"reward_std": 0.38329675246495754,
|
819 |
+
"rewards/concensus_correctness_reward_func": 18.75,
|
820 |
"rewards/consensus_reward_func": 2.0,
|
821 |
"rewards/cumulative_reward_2": 0.0,
|
822 |
+
"rewards/final_correctness_reward_func": 1.625,
|
823 |
+
"rewards/question_recreation_reward_func": 0.9905278831720352,
|
824 |
"rewards/soft_format_reward_func": 0.0,
|
825 |
+
"rewards/strict_format_reward_func": 0.484375,
|
826 |
+
"rewards/xmlcount_reward_func": 1.24609375,
|
827 |
"step": 86
|
828 |
},
|
829 |
{
|
830 |
+
"completion_length": 282.96875,
|
831 |
"epoch": 4.914285714285715,
|
832 |
+
"grad_norm": 10.288511276245117,
|
833 |
+
"kl": 1.007630041684024,
|
834 |
"learning_rate": 2.183372119961499e-08,
|
835 |
+
"loss": 0.001,
|
836 |
+
"reward": 23.806805968284607,
|
837 |
+
"reward_std": 0.08782566487207077,
|
838 |
"rewards/concensus_correctness_reward_func": 17.5,
|
839 |
"rewards/consensus_reward_func": 2.0,
|
840 |
"rewards/cumulative_reward_2": 0.0,
|
841 |
+
"rewards/final_correctness_reward_func": 1.5625,
|
842 |
+
"rewards/question_recreation_reward_func": 0.9943058528006077,
|
843 |
"rewards/soft_format_reward_func": 0.0,
|
844 |
"rewards/strict_format_reward_func": 0.5,
|
845 |
"rewards/xmlcount_reward_func": 1.25,
|
846 |
"step": 88
|
847 |
},
|
848 |
{
|
849 |
+
"completion_length": 288.9166666666667,
|
850 |
"epoch": 5.0,
|
851 |
+
"grad_norm": 5.103769779205322,
|
852 |
+
"kl": 0.5060152485966682,
|
853 |
"learning_rate": 1.5698323748414122e-08,
|
854 |
+
"loss": 0.0004,
|
855 |
+
"reward": 24.75821030139923,
|
856 |
+
"reward_std": 0.1651654613087885,
|
857 |
+
"rewards/concensus_correctness_reward_func": 18.333333333333332,
|
858 |
"rewards/consensus_reward_func": 2.0,
|
859 |
"rewards/cumulative_reward_2": 0.0,
|
860 |
+
"rewards/final_correctness_reward_func": 1.75,
|
861 |
+
"rewards/question_recreation_reward_func": 0.9873770674069723,
|
862 |
"rewards/soft_format_reward_func": 0.0,
|
863 |
+
"rewards/strict_format_reward_func": 0.4375,
|
864 |
"rewards/xmlcount_reward_func": 1.25,
|
865 |
"step": 90
|
866 |
},
|
867 |
{
|
868 |
+
"completion_length": 312.25,
|
869 |
"epoch": 5.114285714285714,
|
870 |
+
"grad_norm": 4.315695762634277,
|
871 |
+
"kl": 0.2543930097017437,
|
872 |
"learning_rate": 1.054566895300324e-08,
|
873 |
+
"loss": 0.0003,
|
874 |
+
"reward": 24.557986617088318,
|
875 |
+
"reward_std": 1.1104227881878614,
|
876 |
+
"rewards/concensus_correctness_reward_func": 18.125,
|
877 |
+
"rewards/consensus_reward_func": 1.9375,
|
878 |
"rewards/cumulative_reward_2": 0.0,
|
879 |
+
"rewards/final_correctness_reward_func": 1.8125,
|
880 |
+
"rewards/question_recreation_reward_func": 0.9994240589439869,
|
881 |
"rewards/soft_format_reward_func": 0.0,
|
882 |
+
"rewards/strict_format_reward_func": 0.453125,
|
883 |
+
"rewards/xmlcount_reward_func": 1.2304375022649765,
|
884 |
"step": 92
|
885 |
},
|
886 |
{
|
887 |
+
"completion_length": 297.375,
|
888 |
"epoch": 5.228571428571429,
|
889 |
+
"grad_norm": 4.21738338470459,
|
890 |
+
"kl": 0.24928702949546278,
|
891 |
"learning_rate": 6.397368838268496e-09,
|
892 |
+
"loss": 0.0002,
|
893 |
+
"reward": 19.507496863603592,
|
894 |
+
"reward_std": 0.33748881984502077,
|
895 |
+
"rewards/concensus_correctness_reward_func": 13.75,
|
896 |
+
"rewards/consensus_reward_func": 1.875,
|
897 |
"rewards/cumulative_reward_2": 0.0,
|
898 |
+
"rewards/final_correctness_reward_func": 1.25,
|
899 |
+
"rewards/question_recreation_reward_func": 0.9494031816720963,
|
900 |
"rewards/soft_format_reward_func": 0.0,
|
901 |
"rewards/strict_format_reward_func": 0.46875,
|
902 |
+
"rewards/xmlcount_reward_func": 1.214343748986721,
|
903 |
"step": 94
|
904 |
},
|
905 |
{
|
906 |
+
"completion_length": 286.0,
|
907 |
"epoch": 5.3428571428571425,
|
908 |
+
"grad_norm": 4.452499866485596,
|
909 |
+
"kl": 7.8268896921072155,
|
910 |
"learning_rate": 3.2708228165273244e-09,
|
911 |
+
"loss": 0.0078,
|
912 |
+
"reward": 22.384023070335388,
|
913 |
+
"reward_std": 0.30442836828297004,
|
914 |
+
"rewards/concensus_correctness_reward_func": 16.25,
|
915 |
"rewards/consensus_reward_func": 2.0,
|
916 |
"rewards/cumulative_reward_2": 0.0,
|
917 |
+
"rewards/final_correctness_reward_func": 1.4375,
|
918 |
+
"rewards/question_recreation_reward_func": 0.9465231820940971,
|
919 |
"rewards/soft_format_reward_func": 0.0,
|
920 |
+
"rewards/strict_format_reward_func": 0.5,
|
921 |
"rewards/xmlcount_reward_func": 1.25,
|
922 |
"step": 96
|
923 |
},
|
924 |
{
|
925 |
+
"completion_length": 264.4375,
|
926 |
"epoch": 5.457142857142857,
|
927 |
+
"grad_norm": 2.327564001083374,
|
928 |
+
"kl": 0.21026335144415498,
|
929 |
"learning_rate": 1.1791447083465133e-09,
|
930 |
+
"loss": 0.0002,
|
931 |
+
"reward": 21.16586148738861,
|
932 |
+
"reward_std": 0.29173441627062857,
|
933 |
"rewards/concensus_correctness_reward_func": 15.0,
|
934 |
+
"rewards/consensus_reward_func": 1.875,
|
935 |
"rewards/cumulative_reward_2": 0.0,
|
936 |
+
"rewards/final_correctness_reward_func": 1.625,
|
937 |
+
"rewards/question_recreation_reward_func": 0.968142680823803,
|
938 |
"rewards/soft_format_reward_func": 0.0,
|
939 |
+
"rewards/strict_format_reward_func": 0.484375,
|
940 |
+
"rewards/xmlcount_reward_func": 1.2133437506854534,
|
941 |
"step": 98
|
942 |
},
|
943 |
{
|
944 |
+
"completion_length": 284.6875,
|
945 |
"epoch": 5.571428571428571,
|
946 |
+
"grad_norm": 7.0255889892578125,
|
947 |
+
"kl": 0.2286427365615964,
|
948 |
"learning_rate": 1.3110773862126667e-10,
|
949 |
+
"loss": 0.0002,
|
950 |
+
"reward": 26.42949390411377,
|
951 |
+
"reward_std": 0.40112065349239856,
|
952 |
"rewards/concensus_correctness_reward_func": 20.0,
|
953 |
"rewards/consensus_reward_func": 2.0,
|
954 |
"rewards/cumulative_reward_2": 0.0,
|
955 |
+
"rewards/final_correctness_reward_func": 1.75,
|
956 |
+
"rewards/question_recreation_reward_func": 0.9490251876413822,
|
957 |
"rewards/soft_format_reward_func": 0.0,
|
958 |
+
"rewards/strict_format_reward_func": 0.484375,
|
959 |
+
"rewards/xmlcount_reward_func": 1.24609375,
|
960 |
"step": 100
|
961 |
},
|
962 |
{
|
963 |
"epoch": 5.571428571428571,
|
964 |
"step": 100,
|
965 |
"total_flos": 0.0,
|
966 |
+
"train_loss": 296549.3844699778,
|
967 |
+
"train_runtime": 1501.8809,
|
968 |
+
"train_samples_per_second": 1.065,
|
969 |
"train_steps_per_second": 0.067
|
970 |
}
|
971 |
],
|