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/Jul02_16-21-45_n136-100-194/events.out.tfevents.1719908528.n136-100-194.1658191.0 +2 -2
- train_results.json +3 -3
- trainer_state.json +431 -431
- 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": 0.9518976428785737,
|
4 |
+
"train_runtime": 6868.0311,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.188,
|
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 sha256:7ca818714ff132e49fe3efaad8395277eee1abf8aa194efda60c3b54df9ba68e
|
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 sha256:120903dde67d78a96fffd5cc95eb23ca4bea822546ee32c239aa2250162c5c51
|
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 sha256:0e5b9a3b73d64098d70d2ccb7c88d04b96265090da8b4e12a2a4a554d6bf47cd
|
3 |
size 4540516344
|
runs/Jul02_16-21-45_n136-100-194/events.out.tfevents.1719908528.n136-100-194.1658191.0
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:d2e53420504b8bb3aa71e5bad9fc8ac592a57562f2a77ac53161e87aca885873
|
3 |
+
size 35423
|
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": 0.9518976428785737,
|
4 |
+
"train_runtime": 6868.0311,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.188,
|
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": 0.
|
37 |
-
"rewards/margins": -0.
|
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":
|
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": -
|
67 |
-
"rewards/margins": 0.
|
68 |
-
"rewards/rejected": -1.
|
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": 0.643750011920929,
|
81 |
-
"rewards/chosen": -
|
82 |
-
"rewards/margins":
|
83 |
-
"rewards/rejected": -
|
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": 0.6937500238418579,
|
96 |
-
"rewards/chosen": -
|
97 |
-
"rewards/margins":
|
98 |
-
"rewards/rejected": -
|
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": 0.7124999761581421,
|
111 |
-
"rewards/chosen": -
|
112 |
-
"rewards/margins":
|
113 |
-
"rewards/rejected": -
|
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": -
|
127 |
-
"rewards/margins":
|
128 |
-
"rewards/rejected": -
|
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": -
|
142 |
-
"rewards/margins":
|
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":
|
158 |
-
"rewards/rejected": -
|
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":
|
173 |
-
"rewards/rejected": -
|
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": -1.
|
184 |
-
"loss":
|
185 |
-
"rewards/accuracies": 0.
|
186 |
-
"rewards/chosen": -
|
187 |
-
"rewards/margins":
|
188 |
-
"rewards/rejected": -
|
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": -1.
|
199 |
-
"loss":
|
200 |
-
"rewards/accuracies": 0.
|
201 |
-
"rewards/chosen": -
|
202 |
-
"rewards/margins":
|
203 |
-
"rewards/rejected": -
|
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": -1.
|
213 |
-
"logps/rejected": -1.
|
214 |
-
"loss":
|
215 |
-
"rewards/accuracies": 0.
|
216 |
-
"rewards/chosen": -
|
217 |
-
"rewards/margins":
|
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": -1.
|
228 |
-
"logps/rejected": -1.
|
229 |
-
"loss":
|
230 |
-
"rewards/accuracies": 0.
|
231 |
-
"rewards/chosen": -
|
232 |
-
"rewards/margins":
|
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": -1.
|
243 |
-
"logps/rejected": -1.
|
244 |
-
"loss":
|
245 |
-
"rewards/accuracies": 0.
|
246 |
-
"rewards/chosen": -
|
247 |
-
"rewards/margins":
|
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": -1.
|
258 |
-
"logps/rejected": -1.
|
259 |
-
"loss":
|
260 |
-
"rewards/accuracies": 0.
|
261 |
-
"rewards/chosen": -
|
262 |
-
"rewards/margins":
|
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": -1.
|
273 |
-
"logps/rejected": -1.
|
274 |
-
"loss":
|
275 |
-
"rewards/accuracies": 0.
|
276 |
-
"rewards/chosen": -
|
277 |
-
"rewards/margins":
|
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": -1.
|
288 |
-
"logps/rejected": -1.
|
289 |
-
"loss":
|
290 |
"rewards/accuracies": 0.75,
|
291 |
-
"rewards/chosen": -
|
292 |
-
"rewards/margins":
|
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": -1.
|
303 |
-
"logps/rejected": -1.
|
304 |
-
"loss":
|
305 |
-
"rewards/accuracies": 0.
|
306 |
-
"rewards/chosen": -
|
307 |
-
"rewards/margins":
|
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": -1.
|
318 |
-
"logps/rejected": -1.
|
319 |
-
"loss":
|
320 |
-
"rewards/accuracies": 0.
|
321 |
-
"rewards/chosen": -
|
322 |
-
"rewards/margins":
|
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": -1.
|
333 |
-
"logps/rejected": -1.
|
334 |
-
"loss":
|
335 |
-
"rewards/accuracies": 0.
|
336 |
-
"rewards/chosen": -
|
337 |
-
"rewards/margins":
|
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": -1.
|
348 |
-
"logps/rejected": -1.
|
349 |
-
"loss":
|
350 |
-
"rewards/accuracies": 0.
|
351 |
-
"rewards/chosen": -
|
352 |
-
"rewards/margins":
|
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": -1.
|
363 |
-
"logps/rejected": -1.
|
364 |
-
"loss":
|
365 |
-
"rewards/accuracies": 0.
|
366 |
-
"rewards/chosen": -
|
367 |
-
"rewards/margins":
|
368 |
-
"rewards/rejected": -
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.55,
|
373 |
-
"grad_norm":
|
374 |
"learning_rate": 5.059649078450834e-08,
|
375 |
-
"logits/chosen": -1.
|
376 |
-
"logits/rejected": -1.
|
377 |
-
"logps/chosen": -1.
|
378 |
-
"logps/rejected": -1.
|
379 |
-
"loss":
|
380 |
-
"rewards/accuracies": 0.
|
381 |
-
"rewards/chosen": -
|
382 |
-
"rewards/margins":
|
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": -1.
|
393 |
-
"logps/rejected": -1.
|
394 |
-
"loss":
|
395 |
-
"rewards/accuracies": 0.
|
396 |
-
"rewards/chosen": -
|
397 |
-
"rewards/margins":
|
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": -1.
|
408 |
-
"logps/rejected": -1.
|
409 |
-
"loss":
|
410 |
-
"rewards/accuracies": 0.
|
411 |
-
"rewards/chosen": -
|
412 |
-
"rewards/margins":
|
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": -1.
|
423 |
-
"logps/rejected": -1.
|
424 |
-
"loss":
|
425 |
-
"rewards/accuracies": 0.
|
426 |
-
"rewards/chosen": -
|
427 |
-
"rewards/margins":
|
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": -1.
|
438 |
-
"logps/rejected": -1.
|
439 |
-
"loss":
|
440 |
-
"rewards/accuracies": 0.
|
441 |
-
"rewards/chosen": -
|
442 |
-
"rewards/margins":
|
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": -1.
|
453 |
-
"logps/rejected": -1.
|
454 |
-
"loss":
|
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": -1.
|
468 |
-
"logps/rejected": -1.
|
469 |
-
"loss":
|
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":
|
479 |
"learning_rate": 2.408731201945432e-08,
|
480 |
-
"logits/chosen": -1.
|
481 |
-
"logits/rejected": -1.
|
482 |
-
"logps/chosen": -1.
|
483 |
-
"logps/rejected": -1.
|
484 |
-
"loss":
|
485 |
-
"rewards/accuracies": 0.
|
486 |
-
"rewards/chosen": -
|
487 |
-
"rewards/margins":
|
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": -1.
|
498 |
-
"logps/rejected": -1.
|
499 |
-
"loss":
|
500 |
-
"rewards/accuracies": 0.
|
501 |
-
"rewards/chosen": -
|
502 |
-
"rewards/margins":
|
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": -1.
|
513 |
-
"logps/rejected": -1.
|
514 |
-
"loss":
|
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": -1.
|
528 |
-
"logps/rejected": -1.
|
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": -1.
|
543 |
-
"logps/rejected": -1.
|
544 |
-
"loss":
|
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": -1.
|
558 |
-
"logps/rejected": -1.
|
559 |
-
"loss":
|
560 |
-
"rewards/accuracies": 0.
|
561 |
-
"rewards/chosen": -
|
562 |
-
"rewards/margins":
|
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": -1.
|
573 |
-
"logps/rejected": -1.
|
574 |
-
"loss":
|
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": -1.
|
588 |
-
"logps/rejected": -1.
|
589 |
-
"loss":
|
590 |
-
"rewards/accuracies": 0.
|
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": -1.
|
603 |
-
"logps/rejected": -1.
|
604 |
-
"loss":
|
605 |
-
"rewards/accuracies": 0.
|
606 |
-
"rewards/chosen": -
|
607 |
-
"rewards/margins":
|
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": -1.
|
618 |
-
"logps/rejected": -1.
|
619 |
-
"loss":
|
620 |
-
"rewards/accuracies": 0.
|
621 |
-
"rewards/chosen": -
|
622 |
-
"rewards/margins":
|
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": -1.
|
633 |
-
"logps/rejected": -1.
|
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": -1.
|
648 |
-
"logps/rejected": -1.
|
649 |
-
"loss":
|
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": -1.
|
663 |
-
"logps/rejected": -1.
|
664 |
-
"loss":
|
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": 1666.4801111075699,
|
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": 1.3182,
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.02,
|
28 |
+
"grad_norm": 1931.392535785282,
|
29 |
"learning_rate": 2.2727272727272725e-08,
|
30 |
+
"logits/chosen": -1.7030127048492432,
|
31 |
+
"logits/rejected": -1.6681498289108276,
|
32 |
+
"logps/chosen": -1.2134774923324585,
|
33 |
+
"logps/rejected": -1.2203598022460938,
|
34 |
+
"loss": 1.3237,
|
35 |
+
"rewards/accuracies": 0.3958333432674408,
|
36 |
+
"rewards/chosen": -0.017192421481013298,
|
37 |
+
"rewards/margins": -0.05587056651711464,
|
38 |
+
"rewards/rejected": 0.03867815062403679,
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
+
"grad_norm": 1830.5113687291991,
|
44 |
"learning_rate": 4.545454545454545e-08,
|
45 |
+
"logits/chosen": -1.780092477798462,
|
46 |
+
"logits/rejected": -1.7350375652313232,
|
47 |
+
"logps/chosen": -1.1447092294692993,
|
48 |
+
"logps/rejected": -1.1850192546844482,
|
49 |
+
"loss": 1.2938,
|
50 |
+
"rewards/accuracies": 0.512499988079071,
|
51 |
+
"rewards/chosen": -0.148686021566391,
|
52 |
+
"rewards/margins": 0.007283961866050959,
|
53 |
+
"rewards/rejected": -0.1559699922800064,
|
54 |
"step": 20
|
55 |
},
|
56 |
{
|
57 |
"epoch": 0.07,
|
58 |
+
"grad_norm": 1936.12548502703,
|
59 |
"learning_rate": 6.818181818181817e-08,
|
60 |
+
"logits/chosen": -1.7440731525421143,
|
61 |
+
"logits/rejected": -1.674883246421814,
|
62 |
+
"logps/chosen": -1.1901237964630127,
|
63 |
+
"logps/rejected": -1.2408424615859985,
|
64 |
+
"loss": 1.2335,
|
65 |
+
"rewards/accuracies": 0.637499988079071,
|
66 |
+
"rewards/chosen": -0.7222227454185486,
|
67 |
+
"rewards/margins": 0.348898708820343,
|
68 |
+
"rewards/rejected": -1.0711214542388916,
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.09,
|
73 |
+
"grad_norm": 1274.5224242026777,
|
74 |
"learning_rate": 9.09090909090909e-08,
|
75 |
+
"logits/chosen": -1.730063796043396,
|
76 |
+
"logits/rejected": -1.6635162830352783,
|
77 |
+
"logps/chosen": -1.2168413400650024,
|
78 |
+
"logps/rejected": -1.274864912033081,
|
79 |
+
"loss": 1.2163,
|
80 |
"rewards/accuracies": 0.643750011920929,
|
81 |
+
"rewards/chosen": -0.9892686009407043,
|
82 |
+
"rewards/margins": 0.5700188875198364,
|
83 |
+
"rewards/rejected": -1.559287428855896,
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.11,
|
88 |
+
"grad_norm": 1227.2878182263496,
|
89 |
"learning_rate": 9.994307990108962e-08,
|
90 |
+
"logits/chosen": -1.6924018859863281,
|
91 |
+
"logits/rejected": -1.6282413005828857,
|
92 |
+
"logps/chosen": -1.2177562713623047,
|
93 |
+
"logps/rejected": -1.2450090646743774,
|
94 |
+
"loss": 1.1002,
|
95 |
"rewards/accuracies": 0.6937500238418579,
|
96 |
+
"rewards/chosen": -0.9669270515441895,
|
97 |
+
"rewards/margins": 0.7160126566886902,
|
98 |
+
"rewards/rejected": -1.6829397678375244,
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.14,
|
103 |
+
"grad_norm": 1544.774039586819,
|
104 |
"learning_rate": 9.959570405988094e-08,
|
105 |
+
"logits/chosen": -1.721835732460022,
|
106 |
+
"logits/rejected": -1.6408640146255493,
|
107 |
+
"logps/chosen": -1.1233644485473633,
|
108 |
+
"logps/rejected": -1.1840593814849854,
|
109 |
+
"loss": 1.0709,
|
110 |
"rewards/accuracies": 0.7124999761581421,
|
111 |
+
"rewards/chosen": -1.4266197681427002,
|
112 |
+
"rewards/margins": 0.8766496777534485,
|
113 |
+
"rewards/rejected": -2.303269386291504,
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.16,
|
118 |
+
"grad_norm": 1005.439321360097,
|
119 |
"learning_rate": 9.893476820924666e-08,
|
120 |
+
"logits/chosen": -1.8021934032440186,
|
121 |
+
"logits/rejected": -1.7167574167251587,
|
122 |
+
"logps/chosen": -1.230804443359375,
|
123 |
+
"logps/rejected": -1.2889682054519653,
|
124 |
+
"loss": 1.0139,
|
125 |
+
"rewards/accuracies": 0.699999988079071,
|
126 |
+
"rewards/chosen": -1.796240210533142,
|
127 |
+
"rewards/margins": 1.1498826742172241,
|
128 |
+
"rewards/rejected": -2.946122884750366,
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.18,
|
133 |
+
"grad_norm": 1396.6389927727296,
|
134 |
"learning_rate": 9.796445099843647e-08,
|
135 |
+
"logits/chosen": -1.7873592376708984,
|
136 |
+
"logits/rejected": -1.6995294094085693,
|
137 |
+
"logps/chosen": -1.2222901582717896,
|
138 |
+
"logps/rejected": -1.2922863960266113,
|
139 |
+
"loss": 1.1257,
|
140 |
+
"rewards/accuracies": 0.7250000238418579,
|
141 |
+
"rewards/chosen": -1.5321768522262573,
|
142 |
+
"rewards/margins": 0.9466292262077332,
|
143 |
+
"rewards/rejected": -2.4788060188293457,
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2,
|
148 |
+
"grad_norm": 1030.6978947603225,
|
149 |
"learning_rate": 9.669088708527066e-08,
|
150 |
+
"logits/chosen": -1.734185814857483,
|
151 |
+
"logits/rejected": -1.6654260158538818,
|
152 |
+
"logps/chosen": -1.2461029291152954,
|
153 |
+
"logps/rejected": -1.265190839767456,
|
154 |
+
"loss": 0.9922,
|
155 |
+
"rewards/accuracies": 0.75,
|
156 |
+
"rewards/chosen": -1.0405571460723877,
|
157 |
+
"rewards/margins": 1.204342246055603,
|
158 |
+
"rewards/rejected": -2.244899272918701,
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.23,
|
163 |
+
"grad_norm": 1586.6455210629026,
|
164 |
"learning_rate": 9.512212835085849e-08,
|
165 |
+
"logits/chosen": -1.765062689781189,
|
166 |
+
"logits/rejected": -1.6767990589141846,
|
167 |
+
"logps/chosen": -1.1784775257110596,
|
168 |
+
"logps/rejected": -1.233091115951538,
|
169 |
+
"loss": 1.016,
|
170 |
+
"rewards/accuracies": 0.737500011920929,
|
171 |
+
"rewards/chosen": -1.0022172927856445,
|
172 |
+
"rewards/margins": 1.1903674602508545,
|
173 |
+
"rewards/rejected": -2.192584753036499,
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.25,
|
178 |
+
"grad_norm": 1144.8635215765032,
|
179 |
"learning_rate": 9.326809299301306e-08,
|
180 |
+
"logits/chosen": -1.7722240686416626,
|
181 |
+
"logits/rejected": -1.6646277904510498,
|
182 |
+
"logps/chosen": -1.2064539194107056,
|
183 |
+
"logps/rejected": -1.2708475589752197,
|
184 |
+
"loss": 0.9709,
|
185 |
+
"rewards/accuracies": 0.7562500238418579,
|
186 |
+
"rewards/chosen": -0.5811842679977417,
|
187 |
+
"rewards/margins": 1.279047966003418,
|
188 |
+
"rewards/rejected": -1.8602321147918701,
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.27,
|
193 |
+
"grad_norm": 1417.2025435553687,
|
194 |
"learning_rate": 9.114050282021158e-08,
|
195 |
+
"logits/chosen": -1.762213110923767,
|
196 |
+
"logits/rejected": -1.6987812519073486,
|
197 |
+
"logps/chosen": -1.1565983295440674,
|
198 |
+
"logps/rejected": -1.2166643142700195,
|
199 |
+
"loss": 0.9277,
|
200 |
+
"rewards/accuracies": 0.71875,
|
201 |
+
"rewards/chosen": -0.9035682678222656,
|
202 |
+
"rewards/margins": 1.2673728466033936,
|
203 |
+
"rewards/rejected": -2.1709413528442383,
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.3,
|
208 |
+
"grad_norm": 1150.053731245993,
|
209 |
"learning_rate": 8.875280914254802e-08,
|
210 |
+
"logits/chosen": -1.7586562633514404,
|
211 |
+
"logits/rejected": -1.6647939682006836,
|
212 |
+
"logps/chosen": -1.20124089717865,
|
213 |
+
"logps/rejected": -1.2465784549713135,
|
214 |
+
"loss": 0.9452,
|
215 |
+
"rewards/accuracies": 0.768750011920929,
|
216 |
+
"rewards/chosen": -0.841080367565155,
|
217 |
+
"rewards/margins": 1.6912786960601807,
|
218 |
+
"rewards/rejected": -2.5323591232299805,
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.32,
|
223 |
+
"grad_norm": 1256.1179361929435,
|
224 |
"learning_rate": 8.612010772821971e-08,
|
225 |
+
"logits/chosen": -1.7858541011810303,
|
226 |
+
"logits/rejected": -1.7347157001495361,
|
227 |
+
"logps/chosen": -1.2216544151306152,
|
228 |
+
"logps/rejected": -1.2101521492004395,
|
229 |
+
"loss": 0.9734,
|
230 |
+
"rewards/accuracies": 0.78125,
|
231 |
+
"rewards/chosen": -0.7775429487228394,
|
232 |
+
"rewards/margins": 1.340767502784729,
|
233 |
+
"rewards/rejected": -2.1183104515075684,
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.34,
|
238 |
+
"grad_norm": 1137.7394214504618,
|
239 |
"learning_rate": 8.325904336322055e-08,
|
240 |
+
"logits/chosen": -1.7536169290542603,
|
241 |
+
"logits/rejected": -1.691907286643982,
|
242 |
+
"logps/chosen": -1.1751422882080078,
|
243 |
+
"logps/rejected": -1.2404186725616455,
|
244 |
+
"loss": 0.9503,
|
245 |
+
"rewards/accuracies": 0.71875,
|
246 |
+
"rewards/chosen": -0.7777413725852966,
|
247 |
+
"rewards/margins": 1.8102855682373047,
|
248 |
+
"rewards/rejected": -2.588027000427246,
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.36,
|
253 |
+
"grad_norm": 1056.887388548025,
|
254 |
"learning_rate": 8.01877046176447e-08,
|
255 |
+
"logits/chosen": -1.6971683502197266,
|
256 |
+
"logits/rejected": -1.6201127767562866,
|
257 |
+
"logps/chosen": -1.192384958267212,
|
258 |
+
"logps/rejected": -1.2586729526519775,
|
259 |
+
"loss": 0.9572,
|
260 |
+
"rewards/accuracies": 0.75,
|
261 |
+
"rewards/chosen": -1.4085007905960083,
|
262 |
+
"rewards/margins": 1.490207314491272,
|
263 |
+
"rewards/rejected": -2.898707866668701,
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.39,
|
268 |
+
"grad_norm": 991.2349708721418,
|
269 |
"learning_rate": 7.692550948392249e-08,
|
270 |
+
"logits/chosen": -1.759788155555725,
|
271 |
+
"logits/rejected": -1.6842018365859985,
|
272 |
+
"logps/chosen": -1.2190964221954346,
|
273 |
+
"logps/rejected": -1.244114637374878,
|
274 |
+
"loss": 0.9699,
|
275 |
+
"rewards/accuracies": 0.7250000238418579,
|
276 |
+
"rewards/chosen": -0.884096622467041,
|
277 |
+
"rewards/margins": 1.4573404788970947,
|
278 |
+
"rewards/rejected": -2.3414371013641357,
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.41,
|
283 |
+
"grad_norm": 1143.0994123075566,
|
284 |
"learning_rate": 7.349308261002021e-08,
|
285 |
+
"logits/chosen": -1.7149896621704102,
|
286 |
+
"logits/rejected": -1.649047613143921,
|
287 |
+
"logps/chosen": -1.2045981884002686,
|
288 |
+
"logps/rejected": -1.267183780670166,
|
289 |
+
"loss": 0.9333,
|
290 |
"rewards/accuracies": 0.75,
|
291 |
+
"rewards/chosen": -0.7512127161026001,
|
292 |
+
"rewards/margins": 1.2935186624526978,
|
293 |
+
"rewards/rejected": -2.044731378555298,
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.43,
|
298 |
+
"grad_norm": 1220.476164637326,
|
299 |
"learning_rate": 6.991212490377531e-08,
|
300 |
+
"logits/chosen": -1.7686491012573242,
|
301 |
+
"logits/rejected": -1.7093932628631592,
|
302 |
+
"logps/chosen": -1.2307132482528687,
|
303 |
+
"logps/rejected": -1.2679650783538818,
|
304 |
+
"loss": 0.9238,
|
305 |
+
"rewards/accuracies": 0.7749999761581421,
|
306 |
+
"rewards/chosen": -0.06102483719587326,
|
307 |
+
"rewards/margins": 1.5709805488586426,
|
308 |
+
"rewards/rejected": -1.6320053339004517,
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.46,
|
313 |
+
"grad_norm": 1207.0013256647494,
|
314 |
"learning_rate": 6.620527633276978e-08,
|
315 |
+
"logits/chosen": -1.7065632343292236,
|
316 |
+
"logits/rejected": -1.6239337921142578,
|
317 |
+
"logps/chosen": -1.197272777557373,
|
318 |
+
"logps/rejected": -1.302229642868042,
|
319 |
+
"loss": 0.912,
|
320 |
+
"rewards/accuracies": 0.7875000238418579,
|
321 |
+
"rewards/chosen": -0.28195858001708984,
|
322 |
+
"rewards/margins": 1.8168141841888428,
|
323 |
+
"rewards/rejected": -2.0987727642059326,
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.48,
|
328 |
+
"grad_norm": 1547.2477497808447,
|
329 |
"learning_rate": 6.239597278716581e-08,
|
330 |
+
"logits/chosen": -1.776951551437378,
|
331 |
+
"logits/rejected": -1.708361029624939,
|
332 |
+
"logps/chosen": -1.24917471408844,
|
333 |
+
"logps/rejected": -1.246996283531189,
|
334 |
+
"loss": 0.9177,
|
335 |
+
"rewards/accuracies": 0.7875000238418579,
|
336 |
+
"rewards/chosen": -0.38446491956710815,
|
337 |
+
"rewards/margins": 1.8485082387924194,
|
338 |
+
"rewards/rejected": -2.232973098754883,
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.5,
|
343 |
+
"grad_norm": 986.8221787714062,
|
344 |
"learning_rate": 5.8508297910462456e-08,
|
345 |
+
"logits/chosen": -1.727354645729065,
|
346 |
+
"logits/rejected": -1.6401188373565674,
|
347 |
+
"logps/chosen": -1.1687428951263428,
|
348 |
+
"logps/rejected": -1.2474868297576904,
|
349 |
+
"loss": 0.8855,
|
350 |
+
"rewards/accuracies": 0.8062499761581421,
|
351 |
+
"rewards/chosen": -0.5227571725845337,
|
352 |
+
"rewards/margins": 2.097059726715088,
|
353 |
+
"rewards/rejected": -2.619816780090332,
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.52,
|
358 |
+
"grad_norm": 1466.4304397519809,
|
359 |
"learning_rate": 5.456683083494731e-08,
|
360 |
+
"logits/chosen": -1.7255643606185913,
|
361 |
+
"logits/rejected": -1.6826807260513306,
|
362 |
+
"logps/chosen": -1.2014678716659546,
|
363 |
+
"logps/rejected": -1.2398700714111328,
|
364 |
+
"loss": 0.9518,
|
365 |
+
"rewards/accuracies": 0.7437499761581421,
|
366 |
+
"rewards/chosen": -0.7659143209457397,
|
367 |
+
"rewards/margins": 1.4660184383392334,
|
368 |
+
"rewards/rejected": -2.231933116912842,
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.55,
|
373 |
+
"grad_norm": 1385.4891697791213,
|
374 |
"learning_rate": 5.059649078450834e-08,
|
375 |
+
"logits/chosen": -1.7211616039276123,
|
376 |
+
"logits/rejected": -1.6692482233047485,
|
377 |
+
"logps/chosen": -1.1779934167861938,
|
378 |
+
"logps/rejected": -1.2568769454956055,
|
379 |
+
"loss": 0.8955,
|
380 |
+
"rewards/accuracies": 0.7749999761581421,
|
381 |
+
"rewards/chosen": -0.9639942049980164,
|
382 |
+
"rewards/margins": 1.6844803094863892,
|
383 |
+
"rewards/rejected": -2.6484744548797607,
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.57,
|
388 |
+
"grad_norm": 1256.1363696264702,
|
389 |
"learning_rate": 4.6622379527277186e-08,
|
390 |
+
"logits/chosen": -1.7240123748779297,
|
391 |
+
"logits/rejected": -1.6668732166290283,
|
392 |
+
"logps/chosen": -1.1512458324432373,
|
393 |
+
"logps/rejected": -1.2001845836639404,
|
394 |
+
"loss": 0.8877,
|
395 |
+
"rewards/accuracies": 0.737500011920929,
|
396 |
+
"rewards/chosen": -1.2357298135757446,
|
397 |
+
"rewards/margins": 1.5858875513076782,
|
398 |
+
"rewards/rejected": -2.821617364883423,
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.59,
|
403 |
+
"grad_norm": 882.5182339158957,
|
404 |
"learning_rate": 4.26696226741691e-08,
|
405 |
+
"logits/chosen": -1.736823320388794,
|
406 |
+
"logits/rejected": -1.657954216003418,
|
407 |
+
"logps/chosen": -1.244571328163147,
|
408 |
+
"logps/rejected": -1.2810423374176025,
|
409 |
+
"loss": 0.8757,
|
410 |
+
"rewards/accuracies": 0.7437499761581421,
|
411 |
+
"rewards/chosen": -1.174777626991272,
|
412 |
+
"rewards/margins": 1.6504262685775757,
|
413 |
+
"rewards/rejected": -2.8252038955688477,
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.61,
|
418 |
+
"grad_norm": 927.5282198039703,
|
419 |
"learning_rate": 3.876321082668098e-08,
|
420 |
+
"logits/chosen": -1.8001619577407837,
|
421 |
+
"logits/rejected": -1.7206510305404663,
|
422 |
+
"logps/chosen": -1.2467237710952759,
|
423 |
+
"logps/rejected": -1.272040843963623,
|
424 |
+
"loss": 0.8981,
|
425 |
+
"rewards/accuracies": 0.7749999761581421,
|
426 |
+
"rewards/chosen": -0.8233057856559753,
|
427 |
+
"rewards/margins": 1.6008100509643555,
|
428 |
+
"rewards/rejected": -2.4241158962249756,
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.64,
|
433 |
+
"grad_norm": 1414.5359647496766,
|
434 |
"learning_rate": 3.492784157826244e-08,
|
435 |
+
"logits/chosen": -1.7251765727996826,
|
436 |
+
"logits/rejected": -1.6225860118865967,
|
437 |
+
"logps/chosen": -1.2284091711044312,
|
438 |
+
"logps/rejected": -1.2839155197143555,
|
439 |
+
"loss": 0.8929,
|
440 |
+
"rewards/accuracies": 0.7749999761581421,
|
441 |
+
"rewards/chosen": -0.5533002018928528,
|
442 |
+
"rewards/margins": 2.0116381645202637,
|
443 |
+
"rewards/rejected": -2.5649383068084717,
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.66,
|
448 |
+
"grad_norm": 1016.8878069934569,
|
449 |
"learning_rate": 3.118776336817812e-08,
|
450 |
+
"logits/chosen": -1.777353048324585,
|
451 |
+
"logits/rejected": -1.704052209854126,
|
452 |
+
"logps/chosen": -1.2131690979003906,
|
453 |
+
"logps/rejected": -1.2390577793121338,
|
454 |
+
"loss": 0.8692,
|
455 |
+
"rewards/accuracies": 0.8187500238418579,
|
456 |
+
"rewards/chosen": -0.992035984992981,
|
457 |
+
"rewards/margins": 1.8672508001327515,
|
458 |
+
"rewards/rejected": -2.8592867851257324,
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.68,
|
463 |
+
"grad_norm": 1172.1973055026576,
|
464 |
"learning_rate": 2.7566622175067443e-08,
|
465 |
+
"logits/chosen": -1.75467050075531,
|
466 |
+
"logits/rejected": -1.6821622848510742,
|
467 |
+
"logps/chosen": -1.2167221307754517,
|
468 |
+
"logps/rejected": -1.312439203262329,
|
469 |
+
"loss": 0.9238,
|
470 |
+
"rewards/accuracies": 0.78125,
|
471 |
+
"rewards/chosen": -0.8138607144355774,
|
472 |
+
"rewards/margins": 1.7730525732040405,
|
473 |
+
"rewards/rejected": -2.5869133472442627,
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.71,
|
478 |
+
"grad_norm": 1325.0572604746271,
|
479 |
"learning_rate": 2.408731201945432e-08,
|
480 |
+
"logits/chosen": -1.7551631927490234,
|
481 |
+
"logits/rejected": -1.7048372030258179,
|
482 |
+
"logps/chosen": -1.1999702453613281,
|
483 |
+
"logps/rejected": -1.2347406148910522,
|
484 |
+
"loss": 0.9048,
|
485 |
+
"rewards/accuracies": 0.78125,
|
486 |
+
"rewards/chosen": -0.6074885725975037,
|
487 |
+
"rewards/margins": 1.4615652561187744,
|
488 |
+
"rewards/rejected": -2.0690536499023438,
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.73,
|
493 |
+
"grad_norm": 1307.0038735408154,
|
494 |
"learning_rate": 2.0771830220378112e-08,
|
495 |
+
"logits/chosen": -1.7086585760116577,
|
496 |
+
"logits/rejected": -1.6484758853912354,
|
497 |
+
"logps/chosen": -1.2315673828125,
|
498 |
+
"logps/rejected": -1.2379032373428345,
|
499 |
+
"loss": 0.8568,
|
500 |
+
"rewards/accuracies": 0.768750011920929,
|
501 |
+
"rewards/chosen": -0.8682917356491089,
|
502 |
+
"rewards/margins": 1.418613314628601,
|
503 |
+
"rewards/rejected": -2.28690505027771,
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.75,
|
508 |
+
"grad_norm": 1081.044775563757,
|
509 |
"learning_rate": 1.7641138321260257e-08,
|
510 |
+
"logits/chosen": -1.744683861732483,
|
511 |
+
"logits/rejected": -1.6652017831802368,
|
512 |
+
"logps/chosen": -1.1902309656143188,
|
513 |
+
"logps/rejected": -1.2386529445648193,
|
514 |
+
"loss": 0.8653,
|
515 |
+
"rewards/accuracies": 0.824999988079071,
|
516 |
+
"rewards/chosen": -0.6376137733459473,
|
517 |
+
"rewards/margins": 2.0886900424957275,
|
518 |
+
"rewards/rejected": -2.726303815841675,
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.77,
|
523 |
+
"grad_norm": 981.6058112333443,
|
524 |
"learning_rate": 1.4715029564277793e-08,
|
525 |
+
"logits/chosen": -1.8191375732421875,
|
526 |
+
"logits/rejected": -1.7604516744613647,
|
527 |
+
"logps/chosen": -1.2077248096466064,
|
528 |
+
"logps/rejected": -1.262688398361206,
|
529 |
+
"loss": 0.9084,
|
530 |
+
"rewards/accuracies": 0.7749999761581421,
|
531 |
+
"rewards/chosen": -0.30642908811569214,
|
532 |
+
"rewards/margins": 1.7900102138519287,
|
533 |
+
"rewards/rejected": -2.0964391231536865,
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.8,
|
538 |
+
"grad_norm": 1453.9865513146688,
|
539 |
"learning_rate": 1.2012003751113343e-08,
|
540 |
+
"logits/chosen": -1.7971642017364502,
|
541 |
+
"logits/rejected": -1.7289268970489502,
|
542 |
+
"logps/chosen": -1.171277642250061,
|
543 |
+
"logps/rejected": -1.2399911880493164,
|
544 |
+
"loss": 0.8408,
|
545 |
+
"rewards/accuracies": 0.78125,
|
546 |
+
"rewards/chosen": -0.9107818603515625,
|
547 |
+
"rewards/margins": 1.8522905111312866,
|
548 |
+
"rewards/rejected": -2.7630724906921387,
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.82,
|
553 |
+
"grad_norm": 1215.5568858398267,
|
554 |
"learning_rate": 9.549150281252633e-09,
|
555 |
+
"logits/chosen": -1.7647504806518555,
|
556 |
+
"logits/rejected": -1.7033681869506836,
|
557 |
+
"logps/chosen": -1.1781424283981323,
|
558 |
+
"logps/rejected": -1.246336579322815,
|
559 |
+
"loss": 0.8691,
|
560 |
+
"rewards/accuracies": 0.7437499761581421,
|
561 |
+
"rewards/chosen": -0.6007339358329773,
|
562 |
+
"rewards/margins": 1.6463441848754883,
|
563 |
+
"rewards/rejected": -2.2470781803131104,
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.84,
|
568 |
+
"grad_norm": 989.7203519251033,
|
569 |
"learning_rate": 7.3420401072985306e-09,
|
570 |
+
"logits/chosen": -1.7938206195831299,
|
571 |
+
"logits/rejected": -1.7368061542510986,
|
572 |
+
"logps/chosen": -1.2212886810302734,
|
573 |
+
"logps/rejected": -1.2962526082992554,
|
574 |
+
"loss": 0.8304,
|
575 |
+
"rewards/accuracies": 0.7250000238418579,
|
576 |
+
"rewards/chosen": -0.6675101518630981,
|
577 |
+
"rewards/margins": 1.7500286102294922,
|
578 |
+
"rewards/rejected": -2.41753888130188,
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.86,
|
583 |
+
"grad_norm": 1103.1154426715473,
|
584 |
"learning_rate": 5.404627290395369e-09,
|
585 |
+
"logits/chosen": -1.7590833902359009,
|
586 |
+
"logits/rejected": -1.6879241466522217,
|
587 |
+
"logps/chosen": -1.1950905323028564,
|
588 |
+
"logps/rejected": -1.2529264688491821,
|
589 |
+
"loss": 0.8354,
|
590 |
+
"rewards/accuracies": 0.78125,
|
591 |
+
"rewards/chosen": -0.5484554171562195,
|
592 |
+
"rewards/margins": 1.7839889526367188,
|
593 |
+
"rewards/rejected": -2.332444429397583,
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 0.89,
|
598 |
+
"grad_norm": 1735.4553729147015,
|
599 |
"learning_rate": 3.74916077816162e-09,
|
600 |
+
"logits/chosen": -1.7692234516143799,
|
601 |
+
"logits/rejected": -1.7004655599594116,
|
602 |
+
"logps/chosen": -1.1599231958389282,
|
603 |
+
"logps/rejected": -1.1907106637954712,
|
604 |
+
"loss": 0.8962,
|
605 |
+
"rewards/accuracies": 0.7124999761581421,
|
606 |
+
"rewards/chosen": -1.0128358602523804,
|
607 |
+
"rewards/margins": 1.4658567905426025,
|
608 |
+
"rewards/rejected": -2.4786927700042725,
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 0.91,
|
613 |
+
"grad_norm": 1092.0703383397774,
|
614 |
"learning_rate": 2.386106962899165e-09,
|
615 |
+
"logits/chosen": -1.6911523342132568,
|
616 |
+
"logits/rejected": -1.6125024557113647,
|
617 |
+
"logps/chosen": -1.233060359954834,
|
618 |
+
"logps/rejected": -1.3061407804489136,
|
619 |
+
"loss": 0.8745,
|
620 |
+
"rewards/accuracies": 0.7437499761581421,
|
621 |
+
"rewards/chosen": -0.6639580130577087,
|
622 |
+
"rewards/margins": 1.6740585565567017,
|
623 |
+
"rewards/rejected": -2.3380167484283447,
|
624 |
"step": 400
|
625 |
},
|
626 |
{
|
627 |
"epoch": 0.93,
|
628 |
+
"grad_norm": 1112.8116007670076,
|
629 |
"learning_rate": 1.3240835096913706e-09,
|
630 |
+
"logits/chosen": -1.7251644134521484,
|
631 |
+
"logits/rejected": -1.613480567932129,
|
632 |
+
"logps/chosen": -1.1751197576522827,
|
633 |
+
"logps/rejected": -1.2636477947235107,
|
634 |
+
"loss": 0.9325,
|
635 |
+
"rewards/accuracies": 0.7562500238418579,
|
636 |
+
"rewards/chosen": -0.35689860582351685,
|
637 |
+
"rewards/margins": 1.9113714694976807,
|
638 |
+
"rewards/rejected": -2.2682700157165527,
|
639 |
"step": 410
|
640 |
},
|
641 |
{
|
642 |
"epoch": 0.96,
|
643 |
+
"grad_norm": 1072.5150365604325,
|
644 |
"learning_rate": 5.698048727497462e-10,
|
645 |
+
"logits/chosen": -1.7486751079559326,
|
646 |
+
"logits/rejected": -1.6716489791870117,
|
647 |
+
"logps/chosen": -1.198484182357788,
|
648 |
+
"logps/rejected": -1.2844183444976807,
|
649 |
+
"loss": 0.8652,
|
650 |
+
"rewards/accuracies": 0.800000011920929,
|
651 |
+
"rewards/chosen": -0.7186421155929565,
|
652 |
+
"rewards/margins": 1.897702932357788,
|
653 |
+
"rewards/rejected": -2.616345167160034,
|
654 |
"step": 420
|
655 |
},
|
656 |
{
|
657 |
"epoch": 0.98,
|
658 |
+
"grad_norm": 1007.4274803164712,
|
659 |
"learning_rate": 1.2803984447259387e-10,
|
660 |
+
"logits/chosen": -1.756765365600586,
|
661 |
+
"logits/rejected": -1.6918004751205444,
|
662 |
+
"logps/chosen": -1.189513921737671,
|
663 |
+
"logps/rejected": -1.2563507556915283,
|
664 |
+
"loss": 0.8416,
|
665 |
+
"rewards/accuracies": 0.78125,
|
666 |
+
"rewards/chosen": -0.7174621820449829,
|
667 |
+
"rewards/margins": 1.9183887243270874,
|
668 |
+
"rewards/rejected": -2.635850667953491,
|
669 |
"step": 430
|
670 |
},
|
671 |
{
|
672 |
"epoch": 1.0,
|
673 |
"step": 439,
|
674 |
"total_flos": 0.0,
|
675 |
+
"train_loss": 0.9518976428785737,
|
676 |
+
"train_runtime": 6868.0311,
|
677 |
+
"train_samples_per_second": 8.188,
|
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 sha256:0dbff6071c55671d44cac95124a8b52fa521ae57b58e89f00a53c9ffd395eeac
|
3 |
size 6328
|