Model save
Browse files- all_results.json +3 -3
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- runs/Jul03_06-05-57_n136-129-074/events.out.tfevents.1719957980.n136-129-074.2672781.0 +2 -2
- train_results.json +3 -3
- trainer_state.json +432 -432
- training_args.bin +1 -1
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 56236,
|
6 |
-
"train_samples_per_second": 8.
|
7 |
"train_steps_per_second": 0.064
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 1.0809600353240967,
|
4 |
+
"train_runtime": 6838.8864,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.223,
|
7 |
"train_steps_per_second": 0.064
|
8 |
}
|
model-00001-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4943162336
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7fed3152f0653eb477ecd4813f48e5b38badfc3072df7d0e51cd2dba373baab8
|
3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4999819336
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:768cb2fc066c755a551c8baf1f36264d28e8a73d92a2bbaac0c88d3f0093a81f
|
3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4540516344
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b6e1839f52689708dd2ac01be610d58ce047b76ccf162250f29b5556c599adad
|
3 |
size 4540516344
|
runs/Jul03_06-05-57_n136-129-074/events.out.tfevents.1719957980.n136-129-074.2672781.0
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:76ec4090046dbd4e9d2d0d3f01f40390a1d76adaad09b0d1b4f70c3f6ea27a05
|
3 |
+
size 35415
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 56236,
|
6 |
-
"train_samples_per_second": 8.
|
7 |
"train_steps_per_second": 0.064
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 1.0809600353240967,
|
4 |
+
"train_runtime": 6838.8864,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.223,
|
7 |
"train_steps_per_second": 0.064
|
8 |
}
|
trainer_state.json
CHANGED
@@ -10,13 +10,13 @@
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0,
|
13 |
-
"grad_norm":
|
14 |
"learning_rate": 2.2727272727272727e-09,
|
15 |
"logits/chosen": -1.6768856048583984,
|
16 |
"logits/rejected": -1.7259055376052856,
|
17 |
"logps/chosen": -1.2793102264404297,
|
18 |
"logps/rejected": -1.2162058353424072,
|
19 |
-
"loss":
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
@@ -25,656 +25,656 @@
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.02,
|
28 |
-
"grad_norm":
|
29 |
"learning_rate": 2.2727272727272725e-08,
|
30 |
-
"logits/chosen": -1.
|
31 |
-
"logits/rejected": -1.
|
32 |
-
"logps/chosen": -1.
|
33 |
-
"logps/rejected": -1.
|
34 |
-
"loss":
|
35 |
-
"rewards/accuracies": 0.
|
36 |
-
"rewards/chosen":
|
37 |
-
"rewards/margins": -
|
38 |
-
"rewards/rejected": 0.
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
-
"grad_norm":
|
44 |
"learning_rate": 4.545454545454545e-08,
|
45 |
-
"logits/chosen": -1.
|
46 |
-
"logits/rejected": -1.
|
47 |
-
"logps/chosen": -1.
|
48 |
-
"logps/rejected": -1.
|
49 |
-
"loss":
|
50 |
-
"rewards/accuracies": 0.
|
51 |
-
"rewards/chosen": -0.
|
52 |
-
"rewards/margins": 0.
|
53 |
-
"rewards/rejected": -0.
|
54 |
"step": 20
|
55 |
},
|
56 |
{
|
57 |
"epoch": 0.07,
|
58 |
-
"grad_norm":
|
59 |
"learning_rate": 6.818181818181817e-08,
|
60 |
-
"logits/chosen": -1.
|
61 |
-
"logits/rejected": -1.
|
62 |
-
"logps/chosen": -1.
|
63 |
-
"logps/rejected": -1.
|
64 |
-
"loss":
|
65 |
-
"rewards/accuracies": 0.
|
66 |
-
"rewards/chosen": -0.
|
67 |
-
"rewards/margins": 0.
|
68 |
-
"rewards/rejected": -0.
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.09,
|
73 |
-
"grad_norm":
|
74 |
"learning_rate": 9.09090909090909e-08,
|
75 |
-
"logits/chosen": -1.
|
76 |
-
"logits/rejected": -1.
|
77 |
-
"logps/chosen": -1.
|
78 |
-
"logps/rejected": -1.
|
79 |
-
"loss":
|
80 |
-
"rewards/accuracies": 0.
|
81 |
-
"rewards/chosen": -0.
|
82 |
-
"rewards/margins": 0.
|
83 |
-
"rewards/rejected": -0.
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.11,
|
88 |
-
"grad_norm":
|
89 |
"learning_rate": 9.994307990108962e-08,
|
90 |
-
"logits/chosen": -1.
|
91 |
-
"logits/rejected": -1.
|
92 |
-
"logps/chosen": -1.
|
93 |
-
"logps/rejected": -1.
|
94 |
-
"loss":
|
95 |
-
"rewards/accuracies": 0.
|
96 |
-
"rewards/chosen": -0.
|
97 |
-
"rewards/margins": 0.
|
98 |
-
"rewards/rejected": -0.
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.14,
|
103 |
-
"grad_norm":
|
104 |
"learning_rate": 9.959570405988094e-08,
|
105 |
-
"logits/chosen": -1.
|
106 |
-
"logits/rejected": -1.
|
107 |
-
"logps/chosen": -1.
|
108 |
-
"logps/rejected": -1.
|
109 |
-
"loss":
|
110 |
-
"rewards/accuracies": 0.
|
111 |
-
"rewards/chosen": -0.
|
112 |
-
"rewards/margins": 0.
|
113 |
-
"rewards/rejected": -0.
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.16,
|
118 |
-
"grad_norm":
|
119 |
"learning_rate": 9.893476820924666e-08,
|
120 |
-
"logits/chosen": -1.
|
121 |
-
"logits/rejected": -1.
|
122 |
-
"logps/chosen": -1.
|
123 |
-
"logps/rejected": -1.
|
124 |
-
"loss":
|
125 |
-
"rewards/accuracies": 0.
|
126 |
-
"rewards/chosen": -0.
|
127 |
-
"rewards/margins": 0.
|
128 |
-
"rewards/rejected": -0.
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.18,
|
133 |
-
"grad_norm":
|
134 |
"learning_rate": 9.796445099843647e-08,
|
135 |
-
"logits/chosen": -1.
|
136 |
-
"logits/rejected": -1.
|
137 |
-
"logps/chosen": -1.
|
138 |
-
"logps/rejected": -1.
|
139 |
-
"loss":
|
140 |
-
"rewards/accuracies": 0.
|
141 |
-
"rewards/chosen": -0.
|
142 |
-
"rewards/margins": 0.
|
143 |
-
"rewards/rejected": -
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2,
|
148 |
-
"grad_norm":
|
149 |
"learning_rate": 9.669088708527066e-08,
|
150 |
-
"logits/chosen": -1.
|
151 |
-
"logits/rejected": -1.
|
152 |
-
"logps/chosen": -1.
|
153 |
-
"logps/rejected": -1.
|
154 |
-
"loss":
|
155 |
-
"rewards/accuracies": 0.
|
156 |
-
"rewards/chosen": -
|
157 |
-
"rewards/margins": 0.
|
158 |
-
"rewards/rejected": -1.
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.23,
|
163 |
-
"grad_norm":
|
164 |
"learning_rate": 9.512212835085849e-08,
|
165 |
-
"logits/chosen": -1.
|
166 |
-
"logits/rejected": -1.
|
167 |
-
"logps/chosen": -1.
|
168 |
-
"logps/rejected": -1.
|
169 |
-
"loss":
|
170 |
-
"rewards/accuracies": 0.
|
171 |
-
"rewards/chosen": -
|
172 |
-
"rewards/margins": 0.
|
173 |
-
"rewards/rejected": -1.
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.25,
|
178 |
-
"grad_norm":
|
179 |
"learning_rate": 9.326809299301306e-08,
|
180 |
-
"logits/chosen": -1.
|
181 |
-
"logits/rejected": -1.
|
182 |
-
"logps/chosen": -1.
|
183 |
-
"logps/rejected": -
|
184 |
-
"loss":
|
185 |
-
"rewards/accuracies": 0.
|
186 |
-
"rewards/chosen": -1.
|
187 |
-
"rewards/margins": 0.
|
188 |
-
"rewards/rejected": -1.
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.27,
|
193 |
-
"grad_norm":
|
194 |
"learning_rate": 9.114050282021158e-08,
|
195 |
-
"logits/chosen": -1.
|
196 |
-
"logits/rejected": -1.
|
197 |
-
"logps/chosen": -1.
|
198 |
-
"logps/rejected": -
|
199 |
-
"loss":
|
200 |
-
"rewards/accuracies": 0.
|
201 |
-
"rewards/chosen": -1.
|
202 |
-
"rewards/margins": 0.
|
203 |
-
"rewards/rejected": -1.
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.3,
|
208 |
-
"grad_norm":
|
209 |
"learning_rate": 8.875280914254802e-08,
|
210 |
-
"logits/chosen": -1.
|
211 |
-
"logits/rejected": -1.
|
212 |
-
"logps/chosen": -
|
213 |
-
"logps/rejected": -
|
214 |
-
"loss":
|
215 |
-
"rewards/accuracies": 0.
|
216 |
-
"rewards/chosen": -1.
|
217 |
-
"rewards/margins": 0.
|
218 |
-
"rewards/rejected": -
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.32,
|
223 |
-
"grad_norm":
|
224 |
"learning_rate": 8.612010772821971e-08,
|
225 |
-
"logits/chosen": -1.
|
226 |
-
"logits/rejected": -1.
|
227 |
-
"logps/chosen": -
|
228 |
-
"logps/rejected": -
|
229 |
-
"loss":
|
230 |
-
"rewards/accuracies": 0.
|
231 |
-
"rewards/chosen": -1.
|
232 |
-
"rewards/margins": 0.
|
233 |
-
"rewards/rejected": -
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.34,
|
238 |
-
"grad_norm":
|
239 |
"learning_rate": 8.325904336322055e-08,
|
240 |
-
"logits/chosen": -1.
|
241 |
-
"logits/rejected": -1.
|
242 |
-
"logps/chosen": -
|
243 |
-
"logps/rejected": -
|
244 |
-
"loss":
|
245 |
-
"rewards/accuracies": 0.
|
246 |
-
"rewards/chosen": -
|
247 |
-
"rewards/margins": 0.
|
248 |
-
"rewards/rejected": -
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.36,
|
253 |
-
"grad_norm":
|
254 |
"learning_rate": 8.01877046176447e-08,
|
255 |
-
"logits/chosen": -1.
|
256 |
-
"logits/rejected": -1.
|
257 |
-
"logps/chosen": -
|
258 |
-
"logps/rejected": -
|
259 |
-
"loss":
|
260 |
-
"rewards/accuracies": 0.
|
261 |
-
"rewards/chosen": -
|
262 |
-
"rewards/margins": 0.
|
263 |
-
"rewards/rejected": -
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.39,
|
268 |
-
"grad_norm":
|
269 |
"learning_rate": 7.692550948392249e-08,
|
270 |
-
"logits/chosen": -1.
|
271 |
-
"logits/rejected": -1.
|
272 |
-
"logps/chosen": -
|
273 |
-
"logps/rejected": -
|
274 |
-
"loss":
|
275 |
-
"rewards/accuracies": 0.
|
276 |
-
"rewards/chosen": -
|
277 |
-
"rewards/margins": 0.
|
278 |
-
"rewards/rejected": -
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.41,
|
283 |
-
"grad_norm":
|
284 |
"learning_rate": 7.349308261002021e-08,
|
285 |
-
"logits/chosen": -1.
|
286 |
-
"logits/rejected": -1.
|
287 |
-
"logps/chosen": -
|
288 |
-
"logps/rejected": -
|
289 |
-
"loss":
|
290 |
-
"rewards/accuracies": 0.
|
291 |
-
"rewards/chosen": -
|
292 |
-
"rewards/margins": 0.
|
293 |
-
"rewards/rejected": -
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.43,
|
298 |
-
"grad_norm":
|
299 |
"learning_rate": 6.991212490377531e-08,
|
300 |
-
"logits/chosen": -1.
|
301 |
-
"logits/rejected": -1.
|
302 |
-
"logps/chosen": -
|
303 |
-
"logps/rejected": -
|
304 |
-
"loss":
|
305 |
-
"rewards/accuracies": 0.
|
306 |
-
"rewards/chosen": -
|
307 |
-
"rewards/margins": 0.
|
308 |
-
"rewards/rejected": -
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.46,
|
313 |
-
"grad_norm":
|
314 |
"learning_rate": 6.620527633276978e-08,
|
315 |
-
"logits/chosen": -1.
|
316 |
-
"logits/rejected": -1.
|
317 |
-
"logps/chosen": -
|
318 |
-
"logps/rejected": -
|
319 |
-
"loss":
|
320 |
-
"rewards/accuracies": 0.
|
321 |
-
"rewards/chosen": -
|
322 |
-
"rewards/margins": 0.
|
323 |
-
"rewards/rejected": -
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.48,
|
328 |
-
"grad_norm":
|
329 |
"learning_rate": 6.239597278716581e-08,
|
330 |
-
"logits/chosen": -1.
|
331 |
-
"logits/rejected": -1.
|
332 |
-
"logps/chosen": -
|
333 |
-
"logps/rejected": -
|
334 |
-
"loss":
|
335 |
-
"rewards/accuracies": 0.
|
336 |
-
"rewards/chosen": -
|
337 |
-
"rewards/margins": 0.
|
338 |
-
"rewards/rejected": -
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.5,
|
343 |
-
"grad_norm":
|
344 |
"learning_rate": 5.8508297910462456e-08,
|
345 |
-
"logits/chosen": -1.
|
346 |
-
"logits/rejected": -1.
|
347 |
-
"logps/chosen": -
|
348 |
-
"logps/rejected": -
|
349 |
-
"loss":
|
350 |
-
"rewards/accuracies": 0.
|
351 |
-
"rewards/chosen": -
|
352 |
-
"rewards/margins": 0.
|
353 |
-
"rewards/rejected": -
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.52,
|
358 |
-
"grad_norm":
|
359 |
"learning_rate": 5.456683083494731e-08,
|
360 |
-
"logits/chosen": -1.
|
361 |
-
"logits/rejected": -1.
|
362 |
-
"logps/chosen": -
|
363 |
-
"logps/rejected": -
|
364 |
-
"loss":
|
365 |
-
"rewards/accuracies": 0.
|
366 |
-
"rewards/chosen": -
|
367 |
-
"rewards/margins": 0.
|
368 |
-
"rewards/rejected": -
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.55,
|
373 |
-
"grad_norm": 50.
|
374 |
"learning_rate": 5.059649078450834e-08,
|
375 |
-
"logits/chosen": -1.
|
376 |
-
"logits/rejected": -1.
|
377 |
-
"logps/chosen": -
|
378 |
-
"logps/rejected": -
|
379 |
-
"loss":
|
380 |
-
"rewards/accuracies": 0.
|
381 |
-
"rewards/chosen": -
|
382 |
-
"rewards/margins": 0.
|
383 |
-
"rewards/rejected": -
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.57,
|
388 |
-
"grad_norm":
|
389 |
"learning_rate": 4.6622379527277186e-08,
|
390 |
-
"logits/chosen": -1.
|
391 |
-
"logits/rejected": -1.
|
392 |
-
"logps/chosen": -
|
393 |
-
"logps/rejected": -
|
394 |
-
"loss":
|
395 |
-
"rewards/accuracies": 0.
|
396 |
-
"rewards/chosen": -
|
397 |
-
"rewards/margins": 0.
|
398 |
-
"rewards/rejected": -
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.59,
|
403 |
-
"grad_norm":
|
404 |
"learning_rate": 4.26696226741691e-08,
|
405 |
-
"logits/chosen": -1.
|
406 |
-
"logits/rejected": -1.
|
407 |
-
"logps/chosen": -
|
408 |
-
"logps/rejected": -
|
409 |
-
"loss":
|
410 |
-
"rewards/accuracies": 0.
|
411 |
-
"rewards/chosen": -
|
412 |
-
"rewards/margins": 0.
|
413 |
-
"rewards/rejected": -
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.61,
|
418 |
-
"grad_norm":
|
419 |
"learning_rate": 3.876321082668098e-08,
|
420 |
-
"logits/chosen": -1.
|
421 |
-
"logits/rejected": -1.
|
422 |
-
"logps/chosen": -
|
423 |
-
"logps/rejected": -
|
424 |
-
"loss":
|
425 |
-
"rewards/accuracies": 0.
|
426 |
-
"rewards/chosen": -
|
427 |
-
"rewards/margins": 0.
|
428 |
-
"rewards/rejected": -
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.64,
|
433 |
-
"grad_norm":
|
434 |
"learning_rate": 3.492784157826244e-08,
|
435 |
-
"logits/chosen": -1.
|
436 |
-
"logits/rejected": -1.
|
437 |
-
"logps/chosen": -
|
438 |
-
"logps/rejected": -
|
439 |
-
"loss":
|
440 |
"rewards/accuracies": 0.71875,
|
441 |
-
"rewards/chosen": -
|
442 |
-
"rewards/margins": 0.
|
443 |
-
"rewards/rejected": -
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.66,
|
448 |
-
"grad_norm":
|
449 |
"learning_rate": 3.118776336817812e-08,
|
450 |
-
"logits/chosen": -1.
|
451 |
-
"logits/rejected": -1.
|
452 |
-
"logps/chosen": -
|
453 |
-
"logps/rejected": -
|
454 |
-
"loss": 0.
|
455 |
-
"rewards/accuracies": 0.
|
456 |
-
"rewards/chosen": -
|
457 |
-
"rewards/margins":
|
458 |
-
"rewards/rejected": -
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.68,
|
463 |
-
"grad_norm":
|
464 |
"learning_rate": 2.7566622175067443e-08,
|
465 |
-
"logits/chosen": -1.
|
466 |
-
"logits/rejected": -1.
|
467 |
-
"logps/chosen": -
|
468 |
-
"logps/rejected": -
|
469 |
-
"loss": 0.
|
470 |
-
"rewards/accuracies": 0.
|
471 |
-
"rewards/chosen": -
|
472 |
-
"rewards/margins":
|
473 |
-
"rewards/rejected": -
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.71,
|
478 |
-
"grad_norm": 47.
|
479 |
"learning_rate": 2.408731201945432e-08,
|
480 |
-
"logits/chosen": -1.
|
481 |
-
"logits/rejected": -1.
|
482 |
-
"logps/chosen": -
|
483 |
-
"logps/rejected": -
|
484 |
-
"loss":
|
485 |
"rewards/accuracies": 0.71875,
|
486 |
-
"rewards/chosen": -
|
487 |
-
"rewards/margins": 0.
|
488 |
-
"rewards/rejected": -
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.73,
|
493 |
-
"grad_norm":
|
494 |
"learning_rate": 2.0771830220378112e-08,
|
495 |
-
"logits/chosen": -1.
|
496 |
-
"logits/rejected": -1.
|
497 |
-
"logps/chosen": -
|
498 |
-
"logps/rejected": -
|
499 |
-
"loss":
|
500 |
-
"rewards/accuracies": 0.
|
501 |
-
"rewards/chosen": -
|
502 |
-
"rewards/margins": 0.
|
503 |
-
"rewards/rejected": -
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.75,
|
508 |
-
"grad_norm":
|
509 |
"learning_rate": 1.7641138321260257e-08,
|
510 |
-
"logits/chosen": -1.
|
511 |
-
"logits/rejected": -1.
|
512 |
-
"logps/chosen": -
|
513 |
-
"logps/rejected": -
|
514 |
-
"loss": 0.
|
515 |
-
"rewards/accuracies": 0.
|
516 |
-
"rewards/chosen": -
|
517 |
-
"rewards/margins":
|
518 |
-
"rewards/rejected": -
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.77,
|
523 |
-
"grad_norm":
|
524 |
"learning_rate": 1.4715029564277793e-08,
|
525 |
-
"logits/chosen": -1.
|
526 |
-
"logits/rejected": -1.
|
527 |
-
"logps/chosen": -
|
528 |
-
"logps/rejected": -
|
529 |
-
"loss":
|
530 |
-
"rewards/accuracies": 0.
|
531 |
-
"rewards/chosen": -
|
532 |
-
"rewards/margins":
|
533 |
-
"rewards/rejected": -
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.8,
|
538 |
-
"grad_norm":
|
539 |
"learning_rate": 1.2012003751113343e-08,
|
540 |
-
"logits/chosen": -1.
|
541 |
-
"logits/rejected": -1.
|
542 |
-
"logps/chosen": -
|
543 |
-
"logps/rejected": -
|
544 |
-
"loss": 0.
|
545 |
-
"rewards/accuracies": 0.
|
546 |
-
"rewards/chosen": -
|
547 |
-
"rewards/margins":
|
548 |
-
"rewards/rejected": -
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.82,
|
553 |
-
"grad_norm":
|
554 |
"learning_rate": 9.549150281252633e-09,
|
555 |
-
"logits/chosen": -1.
|
556 |
-
"logits/rejected": -1.
|
557 |
-
"logps/chosen": -
|
558 |
-
"logps/rejected": -
|
559 |
-
"loss": 0.
|
560 |
-
"rewards/accuracies": 0.
|
561 |
-
"rewards/chosen": -
|
562 |
-
"rewards/margins": 0.
|
563 |
-
"rewards/rejected": -
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.84,
|
568 |
-
"grad_norm":
|
569 |
"learning_rate": 7.3420401072985306e-09,
|
570 |
-
"logits/chosen": -1.
|
571 |
-
"logits/rejected": -1.
|
572 |
-
"logps/chosen": -
|
573 |
-
"logps/rejected": -
|
574 |
-
"loss": 0.
|
575 |
-
"rewards/accuracies": 0.
|
576 |
-
"rewards/chosen": -
|
577 |
-
"rewards/margins":
|
578 |
-
"rewards/rejected": -
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.86,
|
583 |
-
"grad_norm":
|
584 |
"learning_rate": 5.404627290395369e-09,
|
585 |
-
"logits/chosen": -1.
|
586 |
-
"logits/rejected": -1.
|
587 |
-
"logps/chosen": -
|
588 |
-
"logps/rejected": -
|
589 |
-
"loss": 0.
|
590 |
"rewards/accuracies": 0.737500011920929,
|
591 |
-
"rewards/chosen": -
|
592 |
-
"rewards/margins":
|
593 |
-
"rewards/rejected": -
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 0.89,
|
598 |
-
"grad_norm":
|
599 |
"learning_rate": 3.74916077816162e-09,
|
600 |
-
"logits/chosen": -1.
|
601 |
-
"logits/rejected": -1.
|
602 |
-
"logps/chosen": -
|
603 |
-
"logps/rejected": -
|
604 |
-
"loss":
|
605 |
-
"rewards/accuracies": 0.
|
606 |
-
"rewards/chosen": -
|
607 |
-
"rewards/margins": 0.
|
608 |
-
"rewards/rejected": -
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 0.91,
|
613 |
-
"grad_norm":
|
614 |
"learning_rate": 2.386106962899165e-09,
|
615 |
-
"logits/chosen": -1.
|
616 |
-
"logits/rejected": -1.
|
617 |
-
"logps/chosen": -
|
618 |
-
"logps/rejected": -
|
619 |
-
"loss": 0.
|
620 |
-
"rewards/accuracies": 0.
|
621 |
-
"rewards/chosen": -
|
622 |
-
"rewards/margins": 0.
|
623 |
-
"rewards/rejected": -
|
624 |
"step": 400
|
625 |
},
|
626 |
{
|
627 |
"epoch": 0.93,
|
628 |
-
"grad_norm":
|
629 |
"learning_rate": 1.3240835096913706e-09,
|
630 |
-
"logits/chosen": -1.
|
631 |
-
"logits/rejected": -1.
|
632 |
-
"logps/chosen": -
|
633 |
-
"logps/rejected": -
|
634 |
-
"loss":
|
635 |
-
"rewards/accuracies": 0.
|
636 |
-
"rewards/chosen": -
|
637 |
-
"rewards/margins":
|
638 |
-
"rewards/rejected": -
|
639 |
"step": 410
|
640 |
},
|
641 |
{
|
642 |
"epoch": 0.96,
|
643 |
-
"grad_norm":
|
644 |
"learning_rate": 5.698048727497462e-10,
|
645 |
-
"logits/chosen": -1.
|
646 |
-
"logits/rejected": -1.
|
647 |
-
"logps/chosen": -
|
648 |
-
"logps/rejected": -
|
649 |
-
"loss": 0.
|
650 |
-
"rewards/accuracies": 0.
|
651 |
-
"rewards/chosen": -
|
652 |
-
"rewards/margins":
|
653 |
-
"rewards/rejected": -
|
654 |
"step": 420
|
655 |
},
|
656 |
{
|
657 |
"epoch": 0.98,
|
658 |
-
"grad_norm":
|
659 |
"learning_rate": 1.2803984447259387e-10,
|
660 |
-
"logits/chosen": -1.
|
661 |
-
"logits/rejected": -1.
|
662 |
-
"logps/chosen": -
|
663 |
-
"logps/rejected": -
|
664 |
-
"loss": 0.
|
665 |
-
"rewards/accuracies": 0.
|
666 |
-
"rewards/chosen": -
|
667 |
-
"rewards/margins":
|
668 |
-
"rewards/rejected": -
|
669 |
"step": 430
|
670 |
},
|
671 |
{
|
672 |
"epoch": 1.0,
|
673 |
"step": 439,
|
674 |
"total_flos": 0.0,
|
675 |
-
"train_loss":
|
676 |
-
"train_runtime":
|
677 |
-
"train_samples_per_second": 8.
|
678 |
"train_steps_per_second": 0.064
|
679 |
}
|
680 |
],
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0,
|
13 |
+
"grad_norm": 33.30332403665468,
|
14 |
"learning_rate": 2.2727272727272727e-09,
|
15 |
"logits/chosen": -1.6768856048583984,
|
16 |
"logits/rejected": -1.7259055376052856,
|
17 |
"logps/chosen": -1.2793102264404297,
|
18 |
"logps/rejected": -1.2162058353424072,
|
19 |
+
"loss": 1.3133,
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.02,
|
28 |
+
"grad_norm": 37.64303926905523,
|
29 |
"learning_rate": 2.2727272727272725e-08,
|
30 |
+
"logits/chosen": -1.7033135890960693,
|
31 |
+
"logits/rejected": -1.668673038482666,
|
32 |
+
"logps/chosen": -1.2131016254425049,
|
33 |
+
"logps/rejected": -1.22050142288208,
|
34 |
+
"loss": 1.313,
|
35 |
+
"rewards/accuracies": 0.4513888955116272,
|
36 |
+
"rewards/chosen": 0.00040783319855108857,
|
37 |
+
"rewards/margins": -8.263149356935173e-05,
|
38 |
+
"rewards/rejected": 0.0004904646775685251,
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
+
"grad_norm": 38.69260337999141,
|
44 |
"learning_rate": 4.545454545454545e-08,
|
45 |
+
"logits/chosen": -1.7795250415802002,
|
46 |
+
"logits/rejected": -1.7348783016204834,
|
47 |
+
"logps/chosen": -1.1448484659194946,
|
48 |
+
"logps/rejected": -1.1852957010269165,
|
49 |
+
"loss": 1.3122,
|
50 |
+
"rewards/accuracies": 0.5062500238418579,
|
51 |
+
"rewards/chosen": -0.003251913469284773,
|
52 |
+
"rewards/margins": 0.0004202231648378074,
|
53 |
+
"rewards/rejected": -0.003672136692330241,
|
54 |
"step": 20
|
55 |
},
|
56 |
{
|
57 |
"epoch": 0.07,
|
58 |
+
"grad_norm": 44.09359407998382,
|
59 |
"learning_rate": 6.818181818181817e-08,
|
60 |
+
"logits/chosen": -1.7442439794540405,
|
61 |
+
"logits/rejected": -1.6752439737319946,
|
62 |
+
"logps/chosen": -1.1954559087753296,
|
63 |
+
"logps/rejected": -1.248280644416809,
|
64 |
+
"loss": 1.3059,
|
65 |
+
"rewards/accuracies": 0.612500011920929,
|
66 |
+
"rewards/chosen": -0.025108838453888893,
|
67 |
+
"rewards/margins": 0.01119022723287344,
|
68 |
+
"rewards/rejected": -0.03629906848073006,
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.09,
|
73 |
+
"grad_norm": 28.883029165176804,
|
74 |
"learning_rate": 9.09090909090909e-08,
|
75 |
+
"logits/chosen": -1.7305904626846313,
|
76 |
+
"logits/rejected": -1.6642875671386719,
|
77 |
+
"logps/chosen": -1.2533624172210693,
|
78 |
+
"logps/rejected": -1.3383153676986694,
|
79 |
+
"loss": 1.2922,
|
80 |
+
"rewards/accuracies": 0.637499988079071,
|
81 |
+
"rewards/chosen": -0.09282750636339188,
|
82 |
+
"rewards/margins": 0.06525905430316925,
|
83 |
+
"rewards/rejected": -0.15808656811714172,
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.11,
|
88 |
+
"grad_norm": 36.39900209589975,
|
89 |
"learning_rate": 9.994307990108962e-08,
|
90 |
+
"logits/chosen": -1.690720796585083,
|
91 |
+
"logits/rejected": -1.625451683998108,
|
92 |
+
"logps/chosen": -1.3044583797454834,
|
93 |
+
"logps/rejected": -1.3643444776535034,
|
94 |
+
"loss": 1.2643,
|
95 |
+
"rewards/accuracies": 0.71875,
|
96 |
+
"rewards/chosen": -0.19274269044399261,
|
97 |
+
"rewards/margins": 0.0795869454741478,
|
98 |
+
"rewards/rejected": -0.2723296284675598,
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.14,
|
103 |
+
"grad_norm": 43.081578827458706,
|
104 |
"learning_rate": 9.959570405988094e-08,
|
105 |
+
"logits/chosen": -1.71735417842865,
|
106 |
+
"logits/rejected": -1.6361076831817627,
|
107 |
+
"logps/chosen": -1.3119524717330933,
|
108 |
+
"logps/rejected": -1.4046932458877563,
|
109 |
+
"loss": 1.2541,
|
110 |
+
"rewards/accuracies": 0.6499999761581421,
|
111 |
+
"rewards/chosen": -0.4057086110115051,
|
112 |
+
"rewards/margins": 0.0816243588924408,
|
113 |
+
"rewards/rejected": -0.4873329699039459,
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.16,
|
118 |
+
"grad_norm": 33.06897498171632,
|
119 |
"learning_rate": 9.893476820924666e-08,
|
120 |
+
"logits/chosen": -1.7922325134277344,
|
121 |
+
"logits/rejected": -1.7017757892608643,
|
122 |
+
"logps/chosen": -1.5047810077667236,
|
123 |
+
"logps/rejected": -1.630091667175293,
|
124 |
+
"loss": 1.2355,
|
125 |
+
"rewards/accuracies": 0.675000011920929,
|
126 |
+
"rewards/chosen": -0.5838777422904968,
|
127 |
+
"rewards/margins": 0.1572917252779007,
|
128 |
+
"rewards/rejected": -0.7411695718765259,
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.18,
|
133 |
+
"grad_norm": 37.24284057004877,
|
134 |
"learning_rate": 9.796445099843647e-08,
|
135 |
+
"logits/chosen": -1.774518370628357,
|
136 |
+
"logits/rejected": -1.6856935024261475,
|
137 |
+
"logps/chosen": -1.5832115411758423,
|
138 |
+
"logps/rejected": -1.7514270544052124,
|
139 |
+
"loss": 1.232,
|
140 |
+
"rewards/accuracies": 0.668749988079071,
|
141 |
+
"rewards/chosen": -0.7524863481521606,
|
142 |
+
"rewards/margins": 0.21537098288536072,
|
143 |
+
"rewards/rejected": -0.9678572416305542,
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2,
|
148 |
+
"grad_norm": 45.064021238231845,
|
149 |
"learning_rate": 9.669088708527066e-08,
|
150 |
+
"logits/chosen": -1.7184202671051025,
|
151 |
+
"logits/rejected": -1.6467373371124268,
|
152 |
+
"logps/chosen": -1.7363929748535156,
|
153 |
+
"logps/rejected": -1.8083902597427368,
|
154 |
+
"loss": 1.2104,
|
155 |
+
"rewards/accuracies": 0.637499988079071,
|
156 |
+
"rewards/chosen": -1.0013912916183472,
|
157 |
+
"rewards/margins": 0.12990526854991913,
|
158 |
+
"rewards/rejected": -1.1312966346740723,
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.23,
|
163 |
+
"grad_norm": 44.286763175528534,
|
164 |
"learning_rate": 9.512212835085849e-08,
|
165 |
+
"logits/chosen": -1.757889986038208,
|
166 |
+
"logits/rejected": -1.6645339727401733,
|
167 |
+
"logps/chosen": -1.779813528060913,
|
168 |
+
"logps/rejected": -1.9353383779525757,
|
169 |
+
"loss": 1.1819,
|
170 |
+
"rewards/accuracies": 0.6937500238418579,
|
171 |
+
"rewards/chosen": -1.2227165699005127,
|
172 |
+
"rewards/margins": 0.2256297618150711,
|
173 |
+
"rewards/rejected": -1.448346495628357,
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.25,
|
178 |
+
"grad_norm": 43.36692624974112,
|
179 |
"learning_rate": 9.326809299301306e-08,
|
180 |
+
"logits/chosen": -1.761940360069275,
|
181 |
+
"logits/rejected": -1.6550146341323853,
|
182 |
+
"logps/chosen": -1.8854389190673828,
|
183 |
+
"logps/rejected": -2.1229450702667236,
|
184 |
+
"loss": 1.1674,
|
185 |
+
"rewards/accuracies": 0.699999988079071,
|
186 |
+
"rewards/chosen": -1.369593620300293,
|
187 |
+
"rewards/margins": 0.371805876493454,
|
188 |
+
"rewards/rejected": -1.7413995265960693,
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.27,
|
193 |
+
"grad_norm": 41.78554813342914,
|
194 |
"learning_rate": 9.114050282021158e-08,
|
195 |
+
"logits/chosen": -1.7491047382354736,
|
196 |
+
"logits/rejected": -1.6867637634277344,
|
197 |
+
"logps/chosen": -1.8475677967071533,
|
198 |
+
"logps/rejected": -2.0627474784851074,
|
199 |
+
"loss": 1.1591,
|
200 |
+
"rewards/accuracies": 0.731249988079071,
|
201 |
+
"rewards/chosen": -1.400010108947754,
|
202 |
+
"rewards/margins": 0.3355749249458313,
|
203 |
+
"rewards/rejected": -1.7355849742889404,
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.3,
|
208 |
+
"grad_norm": 40.42754129950971,
|
209 |
"learning_rate": 8.875280914254802e-08,
|
210 |
+
"logits/chosen": -1.737173080444336,
|
211 |
+
"logits/rejected": -1.644561529159546,
|
212 |
+
"logps/chosen": -2.0521700382232666,
|
213 |
+
"logps/rejected": -2.296677827835083,
|
214 |
+
"loss": 1.1348,
|
215 |
+
"rewards/accuracies": 0.668749988079071,
|
216 |
+
"rewards/chosen": -1.7186797857284546,
|
217 |
+
"rewards/margins": 0.43216562271118164,
|
218 |
+
"rewards/rejected": -2.1508452892303467,
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.32,
|
223 |
+
"grad_norm": 39.13812568144021,
|
224 |
"learning_rate": 8.612010772821971e-08,
|
225 |
+
"logits/chosen": -1.7612278461456299,
|
226 |
+
"logits/rejected": -1.715679407119751,
|
227 |
+
"logps/chosen": -2.0781049728393555,
|
228 |
+
"logps/rejected": -2.2759194374084473,
|
229 |
+
"loss": 1.127,
|
230 |
+
"rewards/accuracies": 0.737500011920929,
|
231 |
+
"rewards/chosen": -1.7284520864486694,
|
232 |
+
"rewards/margins": 0.4454485774040222,
|
233 |
+
"rewards/rejected": -2.173900604248047,
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.34,
|
238 |
+
"grad_norm": 37.596667789585375,
|
239 |
"learning_rate": 8.325904336322055e-08,
|
240 |
+
"logits/chosen": -1.735419511795044,
|
241 |
+
"logits/rejected": -1.6814868450164795,
|
242 |
+
"logps/chosen": -2.305412769317627,
|
243 |
+
"logps/rejected": -2.55448317527771,
|
244 |
+
"loss": 1.1399,
|
245 |
+
"rewards/accuracies": 0.65625,
|
246 |
+
"rewards/chosen": -2.2760961055755615,
|
247 |
+
"rewards/margins": 0.4037933945655823,
|
248 |
+
"rewards/rejected": -2.679889440536499,
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.36,
|
253 |
+
"grad_norm": 37.85769539137667,
|
254 |
"learning_rate": 8.01877046176447e-08,
|
255 |
+
"logits/chosen": -1.6751991510391235,
|
256 |
+
"logits/rejected": -1.6064836978912354,
|
257 |
+
"logps/chosen": -2.5598020553588867,
|
258 |
+
"logps/rejected": -2.8157076835632324,
|
259 |
+
"loss": 1.09,
|
260 |
+
"rewards/accuracies": 0.643750011920929,
|
261 |
+
"rewards/chosen": -2.763003349304199,
|
262 |
+
"rewards/margins": 0.40903931856155396,
|
263 |
+
"rewards/rejected": -3.1720428466796875,
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.39,
|
268 |
+
"grad_norm": 33.932219318133306,
|
269 |
"learning_rate": 7.692550948392249e-08,
|
270 |
+
"logits/chosen": -1.7231628894805908,
|
271 |
+
"logits/rejected": -1.6755987405776978,
|
272 |
+
"logps/chosen": -2.624762535095215,
|
273 |
+
"logps/rejected": -2.9136133193969727,
|
274 |
+
"loss": 1.1053,
|
275 |
+
"rewards/accuracies": 0.6625000238418579,
|
276 |
+
"rewards/chosen": -2.8290135860443115,
|
277 |
+
"rewards/margins": 0.5568121671676636,
|
278 |
+
"rewards/rejected": -3.3858256340026855,
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.41,
|
283 |
+
"grad_norm": 48.11500069751816,
|
284 |
"learning_rate": 7.349308261002021e-08,
|
285 |
+
"logits/chosen": -1.6858348846435547,
|
286 |
+
"logits/rejected": -1.6378986835479736,
|
287 |
+
"logps/chosen": -2.640817165374756,
|
288 |
+
"logps/rejected": -2.949113368988037,
|
289 |
+
"loss": 1.0837,
|
290 |
+
"rewards/accuracies": 0.706250011920929,
|
291 |
+
"rewards/chosen": -2.887462615966797,
|
292 |
+
"rewards/margins": 0.5172919034957886,
|
293 |
+
"rewards/rejected": -3.404754638671875,
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.43,
|
298 |
+
"grad_norm": 42.33388198011932,
|
299 |
"learning_rate": 6.991212490377531e-08,
|
300 |
+
"logits/chosen": -1.7423484325408936,
|
301 |
+
"logits/rejected": -1.7037559747695923,
|
302 |
+
"logps/chosen": -2.6472008228302,
|
303 |
+
"logps/rejected": -3.0077877044677734,
|
304 |
+
"loss": 1.0335,
|
305 |
+
"rewards/accuracies": 0.7562500238418579,
|
306 |
+
"rewards/chosen": -2.834195613861084,
|
307 |
+
"rewards/margins": 0.6780903339385986,
|
308 |
+
"rewards/rejected": -3.5122859477996826,
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.46,
|
313 |
+
"grad_norm": 49.65676542149092,
|
314 |
"learning_rate": 6.620527633276978e-08,
|
315 |
+
"logits/chosen": -1.6741564273834229,
|
316 |
+
"logits/rejected": -1.6151821613311768,
|
317 |
+
"logps/chosen": -2.735678195953369,
|
318 |
+
"logps/rejected": -3.225632905960083,
|
319 |
+
"loss": 1.0663,
|
320 |
+
"rewards/accuracies": 0.7250000238418579,
|
321 |
+
"rewards/chosen": -3.0824506282806396,
|
322 |
+
"rewards/margins": 0.80633145570755,
|
323 |
+
"rewards/rejected": -3.888781785964966,
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.48,
|
328 |
+
"grad_norm": 46.53275655997813,
|
329 |
"learning_rate": 6.239597278716581e-08,
|
330 |
+
"logits/chosen": -1.7146323919296265,
|
331 |
+
"logits/rejected": -1.6657183170318604,
|
332 |
+
"logps/chosen": -3.098931312561035,
|
333 |
+
"logps/rejected": -3.467923641204834,
|
334 |
+
"loss": 1.0287,
|
335 |
+
"rewards/accuracies": 0.71875,
|
336 |
+
"rewards/chosen": -3.707202911376953,
|
337 |
+
"rewards/margins": 0.7793115377426147,
|
338 |
+
"rewards/rejected": -4.486514091491699,
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.5,
|
343 |
+
"grad_norm": 47.77625681519385,
|
344 |
"learning_rate": 5.8508297910462456e-08,
|
345 |
+
"logits/chosen": -1.6560382843017578,
|
346 |
+
"logits/rejected": -1.5879056453704834,
|
347 |
+
"logps/chosen": -3.1243553161621094,
|
348 |
+
"logps/rejected": -3.6013519763946533,
|
349 |
+
"loss": 1.032,
|
350 |
+
"rewards/accuracies": 0.71875,
|
351 |
+
"rewards/chosen": -3.921679735183716,
|
352 |
+
"rewards/margins": 0.8384466171264648,
|
353 |
+
"rewards/rejected": -4.76012659072876,
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.52,
|
358 |
+
"grad_norm": 45.3152158322423,
|
359 |
"learning_rate": 5.456683083494731e-08,
|
360 |
+
"logits/chosen": -1.6423381567001343,
|
361 |
+
"logits/rejected": -1.6075971126556396,
|
362 |
+
"logps/chosen": -3.002626419067383,
|
363 |
+
"logps/rejected": -3.339411497116089,
|
364 |
+
"loss": 1.068,
|
365 |
+
"rewards/accuracies": 0.65625,
|
366 |
+
"rewards/chosen": -3.6176345348358154,
|
367 |
+
"rewards/margins": 0.6260865330696106,
|
368 |
+
"rewards/rejected": -4.2437214851379395,
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.55,
|
373 |
+
"grad_norm": 50.18712381426658,
|
374 |
"learning_rate": 5.059649078450834e-08,
|
375 |
+
"logits/chosen": -1.6221996545791626,
|
376 |
+
"logits/rejected": -1.587894082069397,
|
377 |
+
"logps/chosen": -2.9972426891326904,
|
378 |
+
"logps/rejected": -3.4068732261657715,
|
379 |
+
"loss": 1.0045,
|
380 |
+
"rewards/accuracies": 0.65625,
|
381 |
+
"rewards/chosen": -3.657778263092041,
|
382 |
+
"rewards/margins": 0.6951833963394165,
|
383 |
+
"rewards/rejected": -4.352961540222168,
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.57,
|
388 |
+
"grad_norm": 42.94625970616266,
|
389 |
"learning_rate": 4.6622379527277186e-08,
|
390 |
+
"logits/chosen": -1.6361802816390991,
|
391 |
+
"logits/rejected": -1.5883018970489502,
|
392 |
+
"logps/chosen": -3.0472984313964844,
|
393 |
+
"logps/rejected": -3.4067275524139404,
|
394 |
+
"loss": 1.0159,
|
395 |
+
"rewards/accuracies": 0.612500011920929,
|
396 |
+
"rewards/chosen": -3.816819429397583,
|
397 |
+
"rewards/margins": 0.65269935131073,
|
398 |
+
"rewards/rejected": -4.469518661499023,
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.59,
|
403 |
+
"grad_norm": 49.81189434860217,
|
404 |
"learning_rate": 4.26696226741691e-08,
|
405 |
+
"logits/chosen": -1.6441590785980225,
|
406 |
+
"logits/rejected": -1.5848346948623657,
|
407 |
+
"logps/chosen": -3.2412009239196777,
|
408 |
+
"logps/rejected": -3.667572021484375,
|
409 |
+
"loss": 1.0333,
|
410 |
+
"rewards/accuracies": 0.643750011920929,
|
411 |
+
"rewards/chosen": -4.0167555809021,
|
412 |
+
"rewards/margins": 0.8128072619438171,
|
413 |
+
"rewards/rejected": -4.829562664031982,
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.61,
|
418 |
+
"grad_norm": 46.698998113891435,
|
419 |
"learning_rate": 3.876321082668098e-08,
|
420 |
+
"logits/chosen": -1.6987736225128174,
|
421 |
+
"logits/rejected": -1.6376842260360718,
|
422 |
+
"logps/chosen": -3.1670312881469727,
|
423 |
+
"logps/rejected": -3.625418186187744,
|
424 |
+
"loss": 1.0046,
|
425 |
+
"rewards/accuracies": 0.7124999761581421,
|
426 |
+
"rewards/chosen": -3.857081174850464,
|
427 |
+
"rewards/margins": 0.8981560468673706,
|
428 |
+
"rewards/rejected": -4.755237579345703,
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.64,
|
433 |
+
"grad_norm": 54.35348471111713,
|
434 |
"learning_rate": 3.492784157826244e-08,
|
435 |
+
"logits/chosen": -1.63980233669281,
|
436 |
+
"logits/rejected": -1.552004337310791,
|
437 |
+
"logps/chosen": -3.2830092906951904,
|
438 |
+
"logps/rejected": -3.8152382373809814,
|
439 |
+
"loss": 1.0119,
|
440 |
"rewards/accuracies": 0.71875,
|
441 |
+
"rewards/chosen": -4.120265483856201,
|
442 |
+
"rewards/margins": 0.9936790466308594,
|
443 |
+
"rewards/rejected": -5.113945007324219,
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.66,
|
448 |
+
"grad_norm": 44.812750561614926,
|
449 |
"learning_rate": 3.118776336817812e-08,
|
450 |
+
"logits/chosen": -1.6625276803970337,
|
451 |
+
"logits/rejected": -1.6122783422470093,
|
452 |
+
"logps/chosen": -3.191256284713745,
|
453 |
+
"logps/rejected": -3.756882429122925,
|
454 |
+
"loss": 0.9859,
|
455 |
+
"rewards/accuracies": 0.71875,
|
456 |
+
"rewards/chosen": -3.97601580619812,
|
457 |
+
"rewards/margins": 1.1168193817138672,
|
458 |
+
"rewards/rejected": -5.092835426330566,
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.68,
|
463 |
+
"grad_norm": 45.89134253017904,
|
464 |
"learning_rate": 2.7566622175067443e-08,
|
465 |
+
"logits/chosen": -1.6413261890411377,
|
466 |
+
"logits/rejected": -1.5825086832046509,
|
467 |
+
"logps/chosen": -3.339484691619873,
|
468 |
+
"logps/rejected": -3.9588654041290283,
|
469 |
+
"loss": 0.994,
|
470 |
+
"rewards/accuracies": 0.768750011920929,
|
471 |
+
"rewards/chosen": -4.261802673339844,
|
472 |
+
"rewards/margins": 1.0827885866165161,
|
473 |
+
"rewards/rejected": -5.3445916175842285,
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.71,
|
478 |
+
"grad_norm": 47.840562340740895,
|
479 |
"learning_rate": 2.408731201945432e-08,
|
480 |
+
"logits/chosen": -1.64263117313385,
|
481 |
+
"logits/rejected": -1.6013950109481812,
|
482 |
+
"logps/chosen": -3.251277446746826,
|
483 |
+
"logps/rejected": -3.651395082473755,
|
484 |
+
"loss": 1.0008,
|
485 |
"rewards/accuracies": 0.71875,
|
486 |
+
"rewards/chosen": -4.11476469039917,
|
487 |
+
"rewards/margins": 0.7599252462387085,
|
488 |
+
"rewards/rejected": -4.874690532684326,
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.73,
|
493 |
+
"grad_norm": 51.79356167073485,
|
494 |
"learning_rate": 2.0771830220378112e-08,
|
495 |
+
"logits/chosen": -1.5991486310958862,
|
496 |
+
"logits/rejected": -1.5396713018417358,
|
497 |
+
"logps/chosen": -3.2509543895721436,
|
498 |
+
"logps/rejected": -3.6364498138427734,
|
499 |
+
"loss": 1.0066,
|
500 |
+
"rewards/accuracies": 0.6187499761581421,
|
501 |
+
"rewards/chosen": -4.056139945983887,
|
502 |
+
"rewards/margins": 0.7866916060447693,
|
503 |
+
"rewards/rejected": -4.842831611633301,
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.75,
|
508 |
+
"grad_norm": 43.99284684689101,
|
509 |
"learning_rate": 1.7641138321260257e-08,
|
510 |
+
"logits/chosen": -1.6334537267684937,
|
511 |
+
"logits/rejected": -1.5692901611328125,
|
512 |
+
"logps/chosen": -3.158041477203369,
|
513 |
+
"logps/rejected": -3.8241424560546875,
|
514 |
+
"loss": 0.9807,
|
515 |
+
"rewards/accuracies": 0.831250011920929,
|
516 |
+
"rewards/chosen": -3.9483726024627686,
|
517 |
+
"rewards/margins": 1.2771327495574951,
|
518 |
+
"rewards/rejected": -5.225505352020264,
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.77,
|
523 |
+
"grad_norm": 50.47520523412627,
|
524 |
"learning_rate": 1.4715029564277793e-08,
|
525 |
+
"logits/chosen": -1.6923463344573975,
|
526 |
+
"logits/rejected": -1.6500104665756226,
|
527 |
+
"logps/chosen": -3.0949554443359375,
|
528 |
+
"logps/rejected": -3.6954338550567627,
|
529 |
+
"loss": 1.0051,
|
530 |
+
"rewards/accuracies": 0.768750011920929,
|
531 |
+
"rewards/chosen": -3.7805895805358887,
|
532 |
+
"rewards/margins": 1.1268298625946045,
|
533 |
+
"rewards/rejected": -4.907419681549072,
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.8,
|
538 |
+
"grad_norm": 48.63775480340643,
|
539 |
"learning_rate": 1.2012003751113343e-08,
|
540 |
+
"logits/chosen": -1.6796951293945312,
|
541 |
+
"logits/rejected": -1.6264684200286865,
|
542 |
+
"logps/chosen": -3.3736748695373535,
|
543 |
+
"logps/rejected": -3.9650447368621826,
|
544 |
+
"loss": 0.9726,
|
545 |
+
"rewards/accuracies": 0.6875,
|
546 |
+
"rewards/chosen": -4.423010349273682,
|
547 |
+
"rewards/margins": 1.0823583602905273,
|
548 |
+
"rewards/rejected": -5.505368709564209,
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.82,
|
553 |
+
"grad_norm": 55.26630420954737,
|
554 |
"learning_rate": 9.549150281252633e-09,
|
555 |
+
"logits/chosen": -1.6259968280792236,
|
556 |
+
"logits/rejected": -1.5858485698699951,
|
557 |
+
"logps/chosen": -3.211542844772339,
|
558 |
+
"logps/rejected": -3.735614776611328,
|
559 |
+
"loss": 0.9729,
|
560 |
+
"rewards/accuracies": 0.6937500238418579,
|
561 |
+
"rewards/chosen": -4.078815460205078,
|
562 |
+
"rewards/margins": 0.9446828961372375,
|
563 |
+
"rewards/rejected": -5.02349853515625,
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.84,
|
568 |
+
"grad_norm": 51.52261591377872,
|
569 |
"learning_rate": 7.3420401072985306e-09,
|
570 |
+
"logits/chosen": -1.6755279302597046,
|
571 |
+
"logits/rejected": -1.6221554279327393,
|
572 |
+
"logps/chosen": -3.299112319946289,
|
573 |
+
"logps/rejected": -3.911120653152466,
|
574 |
+
"loss": 0.9649,
|
575 |
+
"rewards/accuracies": 0.7124999761581421,
|
576 |
+
"rewards/chosen": -4.168996810913086,
|
577 |
+
"rewards/margins": 1.1090896129608154,
|
578 |
+
"rewards/rejected": -5.2780866622924805,
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.86,
|
583 |
+
"grad_norm": 51.72886520205544,
|
584 |
"learning_rate": 5.404627290395369e-09,
|
585 |
+
"logits/chosen": -1.6374752521514893,
|
586 |
+
"logits/rejected": -1.5786619186401367,
|
587 |
+
"logps/chosen": -3.220484972000122,
|
588 |
+
"logps/rejected": -3.803584337234497,
|
589 |
+
"loss": 0.968,
|
590 |
"rewards/accuracies": 0.737500011920929,
|
591 |
+
"rewards/chosen": -4.061758518218994,
|
592 |
+
"rewards/margins": 1.0862071514129639,
|
593 |
+
"rewards/rejected": -5.147965431213379,
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 0.89,
|
598 |
+
"grad_norm": 55.56376010319163,
|
599 |
"learning_rate": 3.74916077816162e-09,
|
600 |
+
"logits/chosen": -1.6384235620498657,
|
601 |
+
"logits/rejected": -1.5836341381072998,
|
602 |
+
"logps/chosen": -3.2395005226135254,
|
603 |
+
"logps/rejected": -3.746983051300049,
|
604 |
+
"loss": 1.0011,
|
605 |
+
"rewards/accuracies": 0.6812499761581421,
|
606 |
+
"rewards/chosen": -4.179410934448242,
|
607 |
+
"rewards/margins": 0.982707142829895,
|
608 |
+
"rewards/rejected": -5.162117958068848,
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 0.91,
|
613 |
+
"grad_norm": 54.81143409505458,
|
614 |
"learning_rate": 2.386106962899165e-09,
|
615 |
+
"logits/chosen": -1.5698174238204956,
|
616 |
+
"logits/rejected": -1.5115009546279907,
|
617 |
+
"logps/chosen": -3.4176878929138184,
|
618 |
+
"logps/rejected": -3.958037853240967,
|
619 |
+
"loss": 0.9695,
|
620 |
+
"rewards/accuracies": 0.706250011920929,
|
621 |
+
"rewards/chosen": -4.382534027099609,
|
622 |
+
"rewards/margins": 0.9680202603340149,
|
623 |
+
"rewards/rejected": -5.350554466247559,
|
624 |
"step": 400
|
625 |
},
|
626 |
{
|
627 |
"epoch": 0.93,
|
628 |
+
"grad_norm": 56.75402221437199,
|
629 |
"learning_rate": 1.3240835096913706e-09,
|
630 |
+
"logits/chosen": -1.594696283340454,
|
631 |
+
"logits/rejected": -1.502890944480896,
|
632 |
+
"logps/chosen": -3.213305711746216,
|
633 |
+
"logps/rejected": -3.921264171600342,
|
634 |
+
"loss": 1.0286,
|
635 |
+
"rewards/accuracies": 0.6937500238418579,
|
636 |
+
"rewards/chosen": -4.083509922027588,
|
637 |
+
"rewards/margins": 1.2770874500274658,
|
638 |
+
"rewards/rejected": -5.360597133636475,
|
639 |
"step": 410
|
640 |
},
|
641 |
{
|
642 |
"epoch": 0.96,
|
643 |
+
"grad_norm": 47.085112169528884,
|
644 |
"learning_rate": 5.698048727497462e-10,
|
645 |
+
"logits/chosen": -1.6298091411590576,
|
646 |
+
"logits/rejected": -1.5658090114593506,
|
647 |
+
"logps/chosen": -3.3380351066589355,
|
648 |
+
"logps/rejected": -3.9660251140594482,
|
649 |
+
"loss": 0.983,
|
650 |
+
"rewards/accuracies": 0.7124999761581421,
|
651 |
+
"rewards/chosen": -4.293475151062012,
|
652 |
+
"rewards/margins": 1.1220663785934448,
|
653 |
+
"rewards/rejected": -5.415541648864746,
|
654 |
"step": 420
|
655 |
},
|
656 |
{
|
657 |
"epoch": 0.98,
|
658 |
+
"grad_norm": 46.70771599324875,
|
659 |
"learning_rate": 1.2803984447259387e-10,
|
660 |
+
"logits/chosen": -1.6368719339370728,
|
661 |
+
"logits/rejected": -1.5942411422729492,
|
662 |
+
"logps/chosen": -3.3361122608184814,
|
663 |
+
"logps/rejected": -3.98066782951355,
|
664 |
+
"loss": 0.9434,
|
665 |
+
"rewards/accuracies": 0.706250011920929,
|
666 |
+
"rewards/chosen": -4.3075456619262695,
|
667 |
+
"rewards/margins": 1.1938055753707886,
|
668 |
+
"rewards/rejected": -5.501351356506348,
|
669 |
"step": 430
|
670 |
},
|
671 |
{
|
672 |
"epoch": 1.0,
|
673 |
"step": 439,
|
674 |
"total_flos": 0.0,
|
675 |
+
"train_loss": 1.0809600353240967,
|
676 |
+
"train_runtime": 6838.8864,
|
677 |
+
"train_samples_per_second": 8.223,
|
678 |
"train_steps_per_second": 0.064
|
679 |
}
|
680 |
],
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6328
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:76320bb76e913f17a8744a245dfdf080870c3a71ef8331b90daacf6e96edfb04
|
3 |
size 6328
|