Model save
Browse files- all_results.json +4 -4
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- runs/Jul03_01-53-04_n136-100-194/events.out.tfevents.1719942806.n136-100-194.1930716.0 +2 -2
- train_results.json +4 -4
- trainer_state.json +434 -434
- training_args.bin +1 -1
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
-
"train_loss": 0.
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 56236,
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 0.5807303016441016,
|
4 |
+
"train_runtime": 6860.7908,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.197,
|
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 sha256:82d958e991ee8e8b81e29d04fcc8041bd5590fa21a2b4b16d78ac83600e78932
|
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:ded0292a572c378c561e70cb2c28e9ce4fd32ea16fbfbec8c2b36978e671b88f
|
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 sha256:23081b1de882d06c5f50e034e626fcca1f28e623615e921b1870ad9b5ff57021
|
3 |
size 4540516344
|
runs/Jul03_01-53-04_n136-100-194/events.out.tfevents.1719942806.n136-100-194.1930716.0
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 sha256:29dfb50f1ac3f975f3605a500c6d2377cafa56d1723892e0483456ad85387b5a
|
3 |
+
size 35415
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
-
"train_loss": 0.
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 56236,
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 0.5807303016441016,
|
4 |
+
"train_runtime": 6860.7908,
|
5 |
"train_samples": 56236,
|
6 |
+
"train_samples_per_second": 8.197,
|
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": 0.
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
@@ -25,657 +25,657 @@
|
|
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": 0.
|
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": 0.
|
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": 0.
|
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": 0.
|
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": 0.
|
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": 0.
|
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": 0.
|
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": 0.
|
140 |
-
"rewards/accuracies": 0.
|
141 |
-
"rewards/chosen": -0.
|
142 |
-
"rewards/margins": 0.
|
143 |
-
"rewards/rejected": -0.
|
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": 0.
|
155 |
-
"rewards/accuracies": 0.
|
156 |
-
"rewards/chosen": -0.
|
157 |
-
"rewards/margins": 0.
|
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": 0.
|
170 |
"rewards/accuracies": 0.6937500238418579,
|
171 |
-
"rewards/chosen": -0.
|
172 |
-
"rewards/margins": 0.
|
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": 0.
|
185 |
-
"rewards/accuracies": 0.
|
186 |
-
"rewards/chosen": -0.
|
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": -1.
|
199 |
-
"loss": 0.
|
200 |
-
"rewards/accuracies": 0.
|
201 |
-
"rewards/chosen": -0.
|
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": -1.
|
213 |
-
"logps/rejected": -1.
|
214 |
-
"loss": 0.
|
215 |
-
"rewards/accuracies": 0.
|
216 |
-
"rewards/chosen": -1.
|
217 |
-
"rewards/margins": 0.
|
218 |
-
"rewards/rejected": -1.
|
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": 0.
|
230 |
"rewards/accuracies": 0.737500011920929,
|
231 |
-
"rewards/chosen": -1.
|
232 |
-
"rewards/margins": 0.
|
233 |
-
"rewards/rejected": -1.
|
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": -
|
244 |
-
"loss": 0.
|
245 |
-
"rewards/accuracies": 0.
|
246 |
-
"rewards/chosen": -1.
|
247 |
-
"rewards/margins": 0.
|
248 |
-
"rewards/rejected": -1.
|
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": -
|
259 |
-
"loss": 0.
|
260 |
-
"rewards/accuracies": 0.
|
261 |
-
"rewards/chosen": -1.
|
262 |
-
"rewards/margins": 0.
|
263 |
-
"rewards/rejected": -1.
|
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": -
|
274 |
-
"loss": 0.
|
275 |
-
"rewards/accuracies": 0.
|
276 |
-
"rewards/chosen": -1.
|
277 |
-
"rewards/margins": 0.
|
278 |
-
"rewards/rejected": -1.
|
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": -
|
289 |
-
"loss": 0.
|
290 |
-
"rewards/accuracies": 0.
|
291 |
-
"rewards/chosen": -1.
|
292 |
-
"rewards/margins": 0.
|
293 |
-
"rewards/rejected": -1.
|
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": 0.
|
305 |
-
"rewards/accuracies": 0.
|
306 |
-
"rewards/chosen": -1.
|
307 |
-
"rewards/margins": 0.
|
308 |
-
"rewards/rejected": -2.
|
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": -2.
|
319 |
-
"loss": 0.
|
320 |
-
"rewards/accuracies": 0.
|
321 |
-
"rewards/chosen": -1.
|
322 |
-
"rewards/margins": 0.
|
323 |
-
"rewards/rejected": -2.
|
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": -2.
|
334 |
-
"loss": 0.
|
335 |
-
"rewards/accuracies": 0.
|
336 |
-
"rewards/chosen": -
|
337 |
-
"rewards/margins": 0.
|
338 |
-
"rewards/rejected": -2.
|
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": -2.
|
349 |
-
"loss": 0.
|
350 |
-
"rewards/accuracies": 0.
|
351 |
-
"rewards/chosen": -
|
352 |
-
"rewards/margins": 0.
|
353 |
-
"rewards/rejected": -2.
|
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": 0.
|
365 |
-
"rewards/accuracies": 0.
|
366 |
-
"rewards/chosen": -1.
|
367 |
-
"rewards/margins": 0.
|
368 |
-
"rewards/rejected": -2.
|
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": -
|
378 |
-
"logps/rejected": -
|
379 |
-
"loss": 0.
|
380 |
-
"rewards/accuracies": 0.
|
381 |
-
"rewards/chosen": -
|
382 |
-
"rewards/margins": 0.
|
383 |
-
"rewards/rejected": -2.
|
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": 0.
|
395 |
-
"rewards/accuracies": 0.
|
396 |
-
"rewards/chosen": -
|
397 |
-
"rewards/margins": 0.
|
398 |
-
"rewards/rejected": -2.
|
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": -2.
|
409 |
-
"loss": 0.
|
410 |
-
"rewards/accuracies": 0.
|
411 |
-
"rewards/chosen": -
|
412 |
-
"rewards/margins": 0.
|
413 |
-
"rewards/rejected": -2.
|
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": -2.
|
424 |
-
"loss": 0.
|
425 |
-
"rewards/accuracies": 0.
|
426 |
-
"rewards/chosen": -
|
427 |
-
"rewards/margins": 0.
|
428 |
-
"rewards/rejected": -2.
|
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": -2.
|
439 |
-
"loss": 0.
|
440 |
-
"rewards/accuracies": 0.
|
441 |
-
"rewards/chosen": -
|
442 |
-
"rewards/margins": 0.
|
443 |
-
"rewards/rejected": -2.
|
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": -2.
|
454 |
-
"loss": 0.
|
455 |
-
"rewards/accuracies": 0.
|
456 |
-
"rewards/chosen": -
|
457 |
-
"rewards/margins": 0.
|
458 |
-
"rewards/rejected": -2.
|
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": -2.
|
469 |
-
"loss": 0.
|
470 |
-
"rewards/accuracies": 0.
|
471 |
-
"rewards/chosen": -2.
|
472 |
-
"rewards/margins": 0.
|
473 |
-
"rewards/rejected": -2.
|
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": -
|
483 |
-
"logps/rejected": -2.
|
484 |
-
"loss": 0.
|
485 |
-
"rewards/accuracies": 0.
|
486 |
-
"rewards/chosen": -
|
487 |
-
"rewards/margins": 0.
|
488 |
-
"rewards/rejected": -2.
|
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": -2.
|
499 |
-
"loss": 0.
|
500 |
-
"rewards/accuracies": 0.
|
501 |
-
"rewards/chosen": -
|
502 |
-
"rewards/margins": 0.
|
503 |
-
"rewards/rejected": -2.
|
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": -2.
|
514 |
-
"loss": 0.
|
515 |
-
"rewards/accuracies": 0.
|
516 |
-
"rewards/chosen": -
|
517 |
-
"rewards/margins": 0.
|
518 |
-
"rewards/rejected": -2.
|
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": -2.
|
529 |
-
"loss": 0.
|
530 |
-
"rewards/accuracies": 0.
|
531 |
-
"rewards/chosen": -
|
532 |
-
"rewards/margins": 0.
|
533 |
-
"rewards/rejected": -2.
|
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": -2.
|
544 |
-
"loss": 0.
|
545 |
-
"rewards/accuracies": 0.
|
546 |
-
"rewards/chosen": -2.
|
547 |
-
"rewards/margins": 0.
|
548 |
-
"rewards/rejected": -2.
|
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": -2.
|
559 |
-
"loss": 0.
|
560 |
-
"rewards/accuracies": 0.
|
561 |
-
"rewards/chosen": -
|
562 |
-
"rewards/margins": 0.
|
563 |
-
"rewards/rejected": -2.
|
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": -2.
|
574 |
-
"loss": 0.
|
575 |
-
"rewards/accuracies": 0.
|
576 |
-
"rewards/chosen": -
|
577 |
-
"rewards/margins": 0.
|
578 |
-
"rewards/rejected": -2.
|
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": -2.
|
589 |
-
"loss": 0.
|
590 |
-
"rewards/accuracies": 0.
|
591 |
-
"rewards/chosen": -
|
592 |
-
"rewards/margins": 0.
|
593 |
-
"rewards/rejected": -2.
|
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": 0.
|
605 |
-
"rewards/accuracies": 0.
|
606 |
-
"rewards/chosen": -
|
607 |
-
"rewards/margins": 0.
|
608 |
-
"rewards/rejected": -2.
|
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": -2.
|
619 |
-
"loss": 0.
|
620 |
-
"rewards/accuracies": 0.
|
621 |
-
"rewards/chosen": -2.
|
622 |
-
"rewards/margins": 0.
|
623 |
-
"rewards/rejected": -2.
|
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": -2.
|
634 |
-
"loss": 0.
|
635 |
-
"rewards/accuracies": 0.
|
636 |
-
"rewards/chosen": -
|
637 |
-
"rewards/margins": 0.
|
638 |
-
"rewards/rejected": -2.
|
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": -2.
|
649 |
-
"loss": 0.
|
650 |
-
"rewards/accuracies": 0.
|
651 |
-
"rewards/chosen": -2.
|
652 |
-
"rewards/margins": 0.
|
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": -2.
|
664 |
-
"loss": 0.
|
665 |
-
"rewards/accuracies": 0.
|
666 |
-
"rewards/chosen": -2.
|
667 |
-
"rewards/margins": 0.
|
668 |
-
"rewards/rejected": -
|
669 |
"step": 430
|
670 |
},
|
671 |
{
|
672 |
"epoch": 1.0,
|
673 |
"step": 439,
|
674 |
"total_flos": 0.0,
|
675 |
-
"train_loss": 0.
|
676 |
-
"train_runtime":
|
677 |
-
"train_samples_per_second":
|
678 |
-
"train_steps_per_second": 0.
|
679 |
}
|
680 |
],
|
681 |
"logging_steps": 10,
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0,
|
13 |
+
"grad_norm": 34.16262092771155,
|
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": 0.6932,
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.02,
|
28 |
+
"grad_norm": 38.39373462126025,
|
29 |
"learning_rate": 2.2727272727272725e-08,
|
30 |
+
"logits/chosen": -1.703444004058838,
|
31 |
+
"logits/rejected": -1.6687698364257812,
|
32 |
+
"logps/chosen": -1.2133899927139282,
|
33 |
+
"logps/rejected": -1.2203023433685303,
|
34 |
+
"loss": 0.6932,
|
35 |
+
"rewards/accuracies": 0.4236111044883728,
|
36 |
+
"rewards/chosen": -0.00025349482893943787,
|
37 |
+
"rewards/margins": -0.0015861267456784844,
|
38 |
+
"rewards/rejected": 0.0013326319167390466,
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
+
"grad_norm": 39.25857800494746,
|
44 |
"learning_rate": 4.545454545454545e-08,
|
45 |
+
"logits/chosen": -1.7795432806015015,
|
46 |
+
"logits/rejected": -1.734358549118042,
|
47 |
+
"logps/chosen": -1.1447466611862183,
|
48 |
+
"logps/rejected": -1.185644507408142,
|
49 |
+
"loss": 0.692,
|
50 |
+
"rewards/accuracies": 0.53125,
|
51 |
+
"rewards/chosen": -0.004572628065943718,
|
52 |
+
"rewards/margins": 0.001982099376618862,
|
53 |
+
"rewards/rejected": -0.00655472744256258,
|
54 |
"step": 20
|
55 |
},
|
56 |
{
|
57 |
"epoch": 0.07,
|
58 |
+
"grad_norm": 42.49760342477406,
|
59 |
"learning_rate": 6.818181818181817e-08,
|
60 |
+
"logits/chosen": -1.7439743280410767,
|
61 |
+
"logits/rejected": -1.6747764348983765,
|
62 |
+
"logps/chosen": -1.195011019706726,
|
63 |
+
"logps/rejected": -1.247870922088623,
|
64 |
+
"loss": 0.6863,
|
65 |
+
"rewards/accuracies": 0.612500011920929,
|
66 |
+
"rewards/chosen": -0.03632856160402298,
|
67 |
+
"rewards/margins": 0.01689038798213005,
|
68 |
+
"rewards/rejected": -0.05321894958615303,
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.09,
|
73 |
+
"grad_norm": 28.94059188031956,
|
74 |
"learning_rate": 9.09090909090909e-08,
|
75 |
+
"logits/chosen": -1.7316467761993408,
|
76 |
+
"logits/rejected": -1.665168046951294,
|
77 |
+
"logps/chosen": -1.2489526271820068,
|
78 |
+
"logps/rejected": -1.330916166305542,
|
79 |
+
"loss": 0.6749,
|
80 |
+
"rewards/accuracies": 0.637499988079071,
|
81 |
+
"rewards/chosen": -0.12601229548454285,
|
82 |
+
"rewards/margins": 0.08891994506120682,
|
83 |
+
"rewards/rejected": -0.21493223309516907,
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.11,
|
88 |
+
"grad_norm": 34.449211690195526,
|
89 |
"learning_rate": 9.994307990108962e-08,
|
90 |
+
"logits/chosen": -1.693706750869751,
|
91 |
+
"logits/rejected": -1.628753662109375,
|
92 |
+
"logps/chosen": -1.2911241054534912,
|
93 |
+
"logps/rejected": -1.3476231098175049,
|
94 |
+
"loss": 0.6547,
|
95 |
+
"rewards/accuracies": 0.71875,
|
96 |
+
"rewards/chosen": -0.2491108626127243,
|
97 |
+
"rewards/margins": 0.10921992361545563,
|
98 |
+
"rewards/rejected": -0.35833072662353516,
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.14,
|
103 |
+
"grad_norm": 45.18662841636004,
|
104 |
"learning_rate": 9.959570405988094e-08,
|
105 |
+
"logits/chosen": -1.7251135110855103,
|
106 |
+
"logits/rejected": -1.644749402999878,
|
107 |
+
"logps/chosen": -1.2596534490585327,
|
108 |
+
"logps/rejected": -1.3467203378677368,
|
109 |
+
"loss": 0.6529,
|
110 |
+
"rewards/accuracies": 0.643750011920929,
|
111 |
+
"rewards/chosen": -0.4516659379005432,
|
112 |
+
"rewards/margins": 0.10541460663080215,
|
113 |
+
"rewards/rejected": -0.5570805668830872,
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.16,
|
118 |
+
"grad_norm": 31.773982345964285,
|
119 |
"learning_rate": 9.893476820924666e-08,
|
120 |
+
"logits/chosen": -1.8094285726547241,
|
121 |
+
"logits/rejected": -1.723013162612915,
|
122 |
+
"logps/chosen": -1.4289662837982178,
|
123 |
+
"logps/rejected": -1.5425760746002197,
|
124 |
+
"loss": 0.6391,
|
125 |
+
"rewards/accuracies": 0.675000011920929,
|
126 |
+
"rewards/chosen": -0.648373007774353,
|
127 |
+
"rewards/margins": 0.20083490014076233,
|
128 |
+
"rewards/rejected": -0.849207878112793,
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.18,
|
133 |
+
"grad_norm": 34.373704859871594,
|
134 |
"learning_rate": 9.796445099843647e-08,
|
135 |
+
"logits/chosen": -1.8001670837402344,
|
136 |
+
"logits/rejected": -1.7168426513671875,
|
137 |
+
"logps/chosen": -1.458465337753296,
|
138 |
+
"logps/rejected": -1.5951015949249268,
|
139 |
+
"loss": 0.6452,
|
140 |
+
"rewards/accuracies": 0.637499988079071,
|
141 |
+
"rewards/chosen": -0.7544907331466675,
|
142 |
+
"rewards/margins": 0.22831876575946808,
|
143 |
+
"rewards/rejected": -0.9828095436096191,
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2,
|
148 |
+
"grad_norm": 33.91159834305824,
|
149 |
"learning_rate": 9.669088708527066e-08,
|
150 |
+
"logits/chosen": -1.7510133981704712,
|
151 |
+
"logits/rejected": -1.6854064464569092,
|
152 |
+
"logps/chosen": -1.5039077997207642,
|
153 |
+
"logps/rejected": -1.5803452730178833,
|
154 |
+
"loss": 0.6218,
|
155 |
+
"rewards/accuracies": 0.65625,
|
156 |
+
"rewards/chosen": -0.8046310544013977,
|
157 |
+
"rewards/margins": 0.20817911624908447,
|
158 |
+
"rewards/rejected": -1.0128101110458374,
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.23,
|
163 |
+
"grad_norm": 37.79881627068083,
|
164 |
"learning_rate": 9.512212835085849e-08,
|
165 |
+
"logits/chosen": -1.7936757802963257,
|
166 |
+
"logits/rejected": -1.7088886499404907,
|
167 |
+
"logps/chosen": -1.4624369144439697,
|
168 |
+
"logps/rejected": -1.5909606218338013,
|
169 |
+
"loss": 0.6102,
|
170 |
"rewards/accuracies": 0.6937500238418579,
|
171 |
+
"rewards/chosen": -0.881945013999939,
|
172 |
+
"rewards/margins": 0.25744158029556274,
|
173 |
+
"rewards/rejected": -1.139386534690857,
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.25,
|
178 |
+
"grad_norm": 38.73339240736025,
|
179 |
"learning_rate": 9.326809299301306e-08,
|
180 |
+
"logits/chosen": -1.7911834716796875,
|
181 |
+
"logits/rejected": -1.6914081573486328,
|
182 |
+
"logps/chosen": -1.4954617023468018,
|
183 |
+
"logps/rejected": -1.65456223487854,
|
184 |
+
"loss": 0.6056,
|
185 |
+
"rewards/accuracies": 0.706250011920929,
|
186 |
+
"rewards/chosen": -0.8844586610794067,
|
187 |
+
"rewards/margins": 0.3224923014640808,
|
188 |
+
"rewards/rejected": -1.2069510221481323,
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.27,
|
193 |
+
"grad_norm": 35.446910248333666,
|
194 |
"learning_rate": 9.114050282021158e-08,
|
195 |
+
"logits/chosen": -1.786220908164978,
|
196 |
+
"logits/rejected": -1.7271583080291748,
|
197 |
+
"logps/chosen": -1.4743096828460693,
|
198 |
+
"logps/rejected": -1.6279948949813843,
|
199 |
+
"loss": 0.6015,
|
200 |
+
"rewards/accuracies": 0.7562500238418579,
|
201 |
+
"rewards/chosen": -0.9802410006523132,
|
202 |
+
"rewards/margins": 0.3188786804676056,
|
203 |
+
"rewards/rejected": -1.2991197109222412,
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.3,
|
208 |
+
"grad_norm": 34.33852298554088,
|
209 |
"learning_rate": 8.875280914254802e-08,
|
210 |
+
"logits/chosen": -1.7842994928359985,
|
211 |
+
"logits/rejected": -1.6957006454467773,
|
212 |
+
"logps/chosen": -1.6290218830108643,
|
213 |
+
"logps/rejected": -1.7964956760406494,
|
214 |
+
"loss": 0.5858,
|
215 |
+
"rewards/accuracies": 0.699999988079071,
|
216 |
+
"rewards/chosen": -1.308574914932251,
|
217 |
+
"rewards/margins": 0.41714730858802795,
|
218 |
+
"rewards/rejected": -1.725722312927246,
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.32,
|
223 |
+
"grad_norm": 31.7434623187325,
|
224 |
"learning_rate": 8.612010772821971e-08,
|
225 |
+
"logits/chosen": -1.8148624897003174,
|
226 |
+
"logits/rejected": -1.7723850011825562,
|
227 |
+
"logps/chosen": -1.629787802696228,
|
228 |
+
"logps/rejected": -1.7449651956558228,
|
229 |
+
"loss": 0.5865,
|
230 |
"rewards/accuracies": 0.737500011920929,
|
231 |
+
"rewards/chosen": -1.2477266788482666,
|
232 |
+
"rewards/margins": 0.4202614724636078,
|
233 |
+
"rewards/rejected": -1.6679881811141968,
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.34,
|
238 |
+
"grad_norm": 30.04410305195209,
|
239 |
"learning_rate": 8.325904336322055e-08,
|
240 |
+
"logits/chosen": -1.7936302423477173,
|
241 |
+
"logits/rejected": -1.740534782409668,
|
242 |
+
"logps/chosen": -1.6605161428451538,
|
243 |
+
"logps/rejected": -1.848489761352539,
|
244 |
+
"loss": 0.5948,
|
245 |
+
"rewards/accuracies": 0.699999988079071,
|
246 |
+
"rewards/chosen": -1.4794538021087646,
|
247 |
+
"rewards/margins": 0.4223996102809906,
|
248 |
+
"rewards/rejected": -1.901853322982788,
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.36,
|
253 |
+
"grad_norm": 31.6961444390199,
|
254 |
"learning_rate": 8.01877046176447e-08,
|
255 |
+
"logits/chosen": -1.7392256259918213,
|
256 |
+
"logits/rejected": -1.6678926944732666,
|
257 |
+
"logps/chosen": -1.707990288734436,
|
258 |
+
"logps/rejected": -1.8706386089324951,
|
259 |
+
"loss": 0.5828,
|
260 |
+
"rewards/accuracies": 0.668749988079071,
|
261 |
+
"rewards/chosen": -1.5890707969665527,
|
262 |
+
"rewards/margins": 0.33378690481185913,
|
263 |
+
"rewards/rejected": -1.922857642173767,
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.39,
|
268 |
+
"grad_norm": 29.03821659608711,
|
269 |
"learning_rate": 7.692550948392249e-08,
|
270 |
+
"logits/chosen": -1.7871320247650146,
|
271 |
+
"logits/rejected": -1.7301868200302124,
|
272 |
+
"logps/chosen": -1.6618130207061768,
|
273 |
+
"logps/rejected": -1.81256902217865,
|
274 |
+
"loss": 0.584,
|
275 |
+
"rewards/accuracies": 0.6937500238418579,
|
276 |
+
"rewards/chosen": -1.3546721935272217,
|
277 |
+
"rewards/margins": 0.42093437910079956,
|
278 |
+
"rewards/rejected": -1.7756065130233765,
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.41,
|
283 |
+
"grad_norm": 36.01706358849117,
|
284 |
"learning_rate": 7.349308261002021e-08,
|
285 |
+
"logits/chosen": -1.743697166442871,
|
286 |
+
"logits/rejected": -1.6886732578277588,
|
287 |
+
"logps/chosen": -1.658620834350586,
|
288 |
+
"logps/rejected": -1.8341089487075806,
|
289 |
+
"loss": 0.5784,
|
290 |
+
"rewards/accuracies": 0.75,
|
291 |
+
"rewards/chosen": -1.3846042156219482,
|
292 |
+
"rewards/margins": 0.3775123953819275,
|
293 |
+
"rewards/rejected": -1.7621164321899414,
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.43,
|
298 |
+
"grad_norm": 31.031052349259785,
|
299 |
"learning_rate": 6.991212490377531e-08,
|
300 |
+
"logits/chosen": -1.797126054763794,
|
301 |
+
"logits/rejected": -1.7482073307037354,
|
302 |
+
"logps/chosen": -1.731537103652954,
|
303 |
+
"logps/rejected": -1.9256778955459595,
|
304 |
+
"loss": 0.5544,
|
305 |
+
"rewards/accuracies": 0.7124999761581421,
|
306 |
+
"rewards/chosen": -1.504301905632019,
|
307 |
+
"rewards/margins": 0.517795979976654,
|
308 |
+
"rewards/rejected": -2.0220980644226074,
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.46,
|
313 |
+
"grad_norm": 41.499506266425385,
|
314 |
"learning_rate": 6.620527633276978e-08,
|
315 |
+
"logits/chosen": -1.7355127334594727,
|
316 |
+
"logits/rejected": -1.6644805669784546,
|
317 |
+
"logps/chosen": -1.7263151407241821,
|
318 |
+
"logps/rejected": -2.0103392601013184,
|
319 |
+
"loss": 0.5689,
|
320 |
+
"rewards/accuracies": 0.737500011920929,
|
321 |
+
"rewards/chosen": -1.5955860614776611,
|
322 |
+
"rewards/margins": 0.591705858707428,
|
323 |
+
"rewards/rejected": -2.1872916221618652,
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.48,
|
328 |
+
"grad_norm": 34.74514581193027,
|
329 |
"learning_rate": 6.239597278716581e-08,
|
330 |
+
"logits/chosen": -1.7984533309936523,
|
331 |
+
"logits/rejected": -1.7380235195159912,
|
332 |
+
"logps/chosen": -1.856172800064087,
|
333 |
+
"logps/rejected": -2.031567096710205,
|
334 |
+
"loss": 0.5577,
|
335 |
+
"rewards/accuracies": 0.75,
|
336 |
+
"rewards/chosen": -1.8325283527374268,
|
337 |
+
"rewards/margins": 0.5881733894348145,
|
338 |
+
"rewards/rejected": -2.420701742172241,
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.5,
|
343 |
+
"grad_norm": 30.0311431190145,
|
344 |
"learning_rate": 5.8508297910462456e-08,
|
345 |
+
"logits/chosen": -1.7461389303207397,
|
346 |
+
"logits/rejected": -1.665144681930542,
|
347 |
+
"logps/chosen": -1.78778874874115,
|
348 |
+
"logps/rejected": -2.0407111644744873,
|
349 |
+
"loss": 0.5611,
|
350 |
+
"rewards/accuracies": 0.731249988079071,
|
351 |
+
"rewards/chosen": -1.872820258140564,
|
352 |
+
"rewards/margins": 0.5854478478431702,
|
353 |
+
"rewards/rejected": -2.4582679271698,
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.52,
|
358 |
+
"grad_norm": 33.05734006337458,
|
359 |
"learning_rate": 5.456683083494731e-08,
|
360 |
+
"logits/chosen": -1.7369384765625,
|
361 |
+
"logits/rejected": -1.69870924949646,
|
362 |
+
"logps/chosen": -1.7765703201293945,
|
363 |
+
"logps/rejected": -1.9211543798446655,
|
364 |
+
"loss": 0.5821,
|
365 |
+
"rewards/accuracies": 0.643750011920929,
|
366 |
+
"rewards/chosen": -1.7482845783233643,
|
367 |
+
"rewards/margins": 0.3625262379646301,
|
368 |
+
"rewards/rejected": -2.1108105182647705,
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.55,
|
373 |
+
"grad_norm": 42.868772014524744,
|
374 |
"learning_rate": 5.059649078450834e-08,
|
375 |
+
"logits/chosen": -1.7265160083770752,
|
376 |
+
"logits/rejected": -1.6816940307617188,
|
377 |
+
"logps/chosen": -1.7579532861709595,
|
378 |
+
"logps/rejected": -1.9656188488006592,
|
379 |
+
"loss": 0.5499,
|
380 |
+
"rewards/accuracies": 0.65625,
|
381 |
+
"rewards/chosen": -1.7687995433807373,
|
382 |
+
"rewards/margins": 0.43688035011291504,
|
383 |
+
"rewards/rejected": -2.2056798934936523,
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.57,
|
388 |
+
"grad_norm": 32.81853514392731,
|
389 |
"learning_rate": 4.6622379527277186e-08,
|
390 |
+
"logits/chosen": -1.7367656230926514,
|
391 |
+
"logits/rejected": -1.685956358909607,
|
392 |
+
"logps/chosen": -1.7992569208145142,
|
393 |
+
"logps/rejected": -1.9660634994506836,
|
394 |
+
"loss": 0.5585,
|
395 |
+
"rewards/accuracies": 0.6625000238418579,
|
396 |
+
"rewards/chosen": -1.981104850769043,
|
397 |
+
"rewards/margins": 0.4011806845664978,
|
398 |
+
"rewards/rejected": -2.3822855949401855,
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.59,
|
403 |
+
"grad_norm": 33.27003433225899,
|
404 |
"learning_rate": 4.26696226741691e-08,
|
405 |
+
"logits/chosen": -1.7512362003326416,
|
406 |
+
"logits/rejected": -1.6821905374526978,
|
407 |
+
"logps/chosen": -1.897579550743103,
|
408 |
+
"logps/rejected": -2.0977420806884766,
|
409 |
+
"loss": 0.5601,
|
410 |
+
"rewards/accuracies": 0.6499999761581421,
|
411 |
+
"rewards/chosen": -1.9942684173583984,
|
412 |
+
"rewards/margins": 0.5405869483947754,
|
413 |
+
"rewards/rejected": -2.534855365753174,
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.61,
|
418 |
+
"grad_norm": 28.394830302289282,
|
419 |
"learning_rate": 3.876321082668098e-08,
|
420 |
+
"logits/chosen": -1.8098827600479126,
|
421 |
+
"logits/rejected": -1.7416326999664307,
|
422 |
+
"logps/chosen": -1.8279718160629272,
|
423 |
+
"logps/rejected": -2.0366756916046143,
|
424 |
+
"loss": 0.5516,
|
425 |
+
"rewards/accuracies": 0.7562500238418579,
|
426 |
+
"rewards/chosen": -1.7684434652328491,
|
427 |
+
"rewards/margins": 0.5981845855712891,
|
428 |
+
"rewards/rejected": -2.3666281700134277,
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.64,
|
433 |
+
"grad_norm": 33.9712584613191,
|
434 |
"learning_rate": 3.492784157826244e-08,
|
435 |
+
"logits/chosen": -1.7408695220947266,
|
436 |
+
"logits/rejected": -1.6451225280761719,
|
437 |
+
"logps/chosen": -1.844485878944397,
|
438 |
+
"logps/rejected": -2.0920097827911377,
|
439 |
+
"loss": 0.5537,
|
440 |
+
"rewards/accuracies": 0.71875,
|
441 |
+
"rewards/chosen": -1.864828109741211,
|
442 |
+
"rewards/margins": 0.6364026665687561,
|
443 |
+
"rewards/rejected": -2.5012307167053223,
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.66,
|
448 |
+
"grad_norm": 26.97002465180278,
|
449 |
"learning_rate": 3.118776336817812e-08,
|
450 |
+
"logits/chosen": -1.782135009765625,
|
451 |
+
"logits/rejected": -1.72174870967865,
|
452 |
+
"logps/chosen": -1.8031295537948608,
|
453 |
+
"logps/rejected": -2.0307796001434326,
|
454 |
+
"loss": 0.5443,
|
455 |
+
"rewards/accuracies": 0.71875,
|
456 |
+
"rewards/chosen": -1.7996422052383423,
|
457 |
+
"rewards/margins": 0.6613016128540039,
|
458 |
+
"rewards/rejected": -2.4609439373016357,
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.68,
|
463 |
+
"grad_norm": 31.17322900966105,
|
464 |
"learning_rate": 2.7566622175067443e-08,
|
465 |
+
"logits/chosen": -1.7659027576446533,
|
466 |
+
"logits/rejected": -1.7027981281280518,
|
467 |
+
"logps/chosen": -1.8818680047988892,
|
468 |
+
"logps/rejected": -2.163147449493408,
|
469 |
+
"loss": 0.5502,
|
470 |
+
"rewards/accuracies": 0.706250011920929,
|
471 |
+
"rewards/chosen": -2.0198538303375244,
|
472 |
+
"rewards/margins": 0.6098783612251282,
|
473 |
+
"rewards/rejected": -2.629732131958008,
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.71,
|
478 |
+
"grad_norm": 36.262422147908055,
|
479 |
"learning_rate": 2.408731201945432e-08,
|
480 |
+
"logits/chosen": -1.761371374130249,
|
481 |
+
"logits/rejected": -1.7142196893692017,
|
482 |
+
"logps/chosen": -1.8247020244598389,
|
483 |
+
"logps/rejected": -2.0095391273498535,
|
484 |
+
"loss": 0.5485,
|
485 |
+
"rewards/accuracies": 0.6875,
|
486 |
+
"rewards/chosen": -1.8924195766448975,
|
487 |
+
"rewards/margins": 0.49404746294021606,
|
488 |
+
"rewards/rejected": -2.386467218399048,
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.73,
|
493 |
+
"grad_norm": 32.89905332615021,
|
494 |
"learning_rate": 2.0771830220378112e-08,
|
495 |
+
"logits/chosen": -1.7170169353485107,
|
496 |
+
"logits/rejected": -1.6577152013778687,
|
497 |
+
"logps/chosen": -1.8545100688934326,
|
498 |
+
"logps/rejected": -2.0157198905944824,
|
499 |
+
"loss": 0.5515,
|
500 |
+
"rewards/accuracies": 0.6625000238418579,
|
501 |
+
"rewards/chosen": -1.8948768377304077,
|
502 |
+
"rewards/margins": 0.5071803331375122,
|
503 |
+
"rewards/rejected": -2.40205717086792,
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.75,
|
508 |
+
"grad_norm": 29.752828533537873,
|
509 |
"learning_rate": 1.7641138321260257e-08,
|
510 |
+
"logits/chosen": -1.7509727478027344,
|
511 |
+
"logits/rejected": -1.6813408136367798,
|
512 |
+
"logps/chosen": -1.8006902933120728,
|
513 |
+
"logps/rejected": -2.068765878677368,
|
514 |
+
"loss": 0.5463,
|
515 |
+
"rewards/accuracies": 0.8125,
|
516 |
+
"rewards/chosen": -1.8505061864852905,
|
517 |
+
"rewards/margins": 0.7216218113899231,
|
518 |
+
"rewards/rejected": -2.5721280574798584,
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.77,
|
523 |
+
"grad_norm": 29.865199582942665,
|
524 |
"learning_rate": 1.4715029564277793e-08,
|
525 |
+
"logits/chosen": -1.8179073333740234,
|
526 |
+
"logits/rejected": -1.771498680114746,
|
527 |
+
"logps/chosen": -1.7910722494125366,
|
528 |
+
"logps/rejected": -2.0462353229522705,
|
529 |
+
"loss": 0.5539,
|
530 |
+
"rewards/accuracies": 0.7437499761581421,
|
531 |
+
"rewards/chosen": -1.759235143661499,
|
532 |
+
"rewards/margins": 0.6542983651161194,
|
533 |
+
"rewards/rejected": -2.4135334491729736,
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.8,
|
538 |
+
"grad_norm": 33.27484680109776,
|
539 |
"learning_rate": 1.2012003751113343e-08,
|
540 |
+
"logits/chosen": -1.806565284729004,
|
541 |
+
"logits/rejected": -1.745749831199646,
|
542 |
+
"logps/chosen": -1.8502833843231201,
|
543 |
+
"logps/rejected": -2.1222567558288574,
|
544 |
+
"loss": 0.5334,
|
545 |
+
"rewards/accuracies": 0.7124999761581421,
|
546 |
+
"rewards/chosen": -2.0643410682678223,
|
547 |
+
"rewards/margins": 0.6653475761413574,
|
548 |
+
"rewards/rejected": -2.729688882827759,
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.82,
|
553 |
+
"grad_norm": 35.608600933807864,
|
554 |
"learning_rate": 9.549150281252633e-09,
|
555 |
+
"logits/chosen": -1.762521505355835,
|
556 |
+
"logits/rejected": -1.7128925323486328,
|
557 |
+
"logps/chosen": -1.8086223602294922,
|
558 |
+
"logps/rejected": -2.0623652935028076,
|
559 |
+
"loss": 0.5356,
|
560 |
+
"rewards/accuracies": 0.699999988079071,
|
561 |
+
"rewards/chosen": -1.9094616174697876,
|
562 |
+
"rewards/margins": 0.6060369610786438,
|
563 |
+
"rewards/rejected": -2.515498399734497,
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.84,
|
568 |
+
"grad_norm": 31.739203329110374,
|
569 |
"learning_rate": 7.3420401072985306e-09,
|
570 |
+
"logits/chosen": -1.7991466522216797,
|
571 |
+
"logits/rejected": -1.746799111366272,
|
572 |
+
"logps/chosen": -1.8659775257110596,
|
573 |
+
"logps/rejected": -2.1334667205810547,
|
574 |
+
"loss": 0.5356,
|
575 |
+
"rewards/accuracies": 0.7124999761581421,
|
576 |
+
"rewards/chosen": -1.9540916681289673,
|
577 |
+
"rewards/margins": 0.6300768256187439,
|
578 |
+
"rewards/rejected": -2.5841686725616455,
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.86,
|
583 |
+
"grad_norm": 37.231267463423904,
|
584 |
"learning_rate": 5.404627290395369e-09,
|
585 |
+
"logits/chosen": -1.7623279094696045,
|
586 |
+
"logits/rejected": -1.6976515054702759,
|
587 |
+
"logps/chosen": -1.7954782247543335,
|
588 |
+
"logps/rejected": -2.0490074157714844,
|
589 |
+
"loss": 0.5377,
|
590 |
+
"rewards/accuracies": 0.7562500238418579,
|
591 |
+
"rewards/chosen": -1.8176171779632568,
|
592 |
+
"rewards/margins": 0.6405991315841675,
|
593 |
+
"rewards/rejected": -2.4582161903381348,
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 0.89,
|
598 |
+
"grad_norm": 48.31129307964873,
|
599 |
"learning_rate": 3.74916077816162e-09,
|
600 |
+
"logits/chosen": -1.7699306011199951,
|
601 |
+
"logits/rejected": -1.7098052501678467,
|
602 |
+
"logps/chosen": -1.7985498905181885,
|
603 |
+
"logps/rejected": -1.9994351863861084,
|
604 |
+
"loss": 0.5502,
|
605 |
+
"rewards/accuracies": 0.699999988079071,
|
606 |
+
"rewards/chosen": -1.946265459060669,
|
607 |
+
"rewards/margins": 0.5542693734169006,
|
608 |
+
"rewards/rejected": -2.5005345344543457,
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 0.91,
|
613 |
+
"grad_norm": 36.09144092538641,
|
614 |
"learning_rate": 2.386106962899165e-09,
|
615 |
+
"logits/chosen": -1.6945453882217407,
|
616 |
+
"logits/rejected": -1.62551748752594,
|
617 |
+
"logps/chosen": -1.9125568866729736,
|
618 |
+
"logps/rejected": -2.1548168659210205,
|
619 |
+
"loss": 0.5447,
|
620 |
+
"rewards/accuracies": 0.7124999761581421,
|
621 |
+
"rewards/chosen": -2.05840802192688,
|
622 |
+
"rewards/margins": 0.5577607154846191,
|
623 |
+
"rewards/rejected": -2.61616849899292,
|
624 |
"step": 400
|
625 |
},
|
626 |
{
|
627 |
"epoch": 0.93,
|
628 |
+
"grad_norm": 35.37496893135356,
|
629 |
"learning_rate": 1.3240835096913706e-09,
|
630 |
+
"logits/chosen": -1.7273473739624023,
|
631 |
+
"logits/rejected": -1.6288312673568726,
|
632 |
+
"logps/chosen": -1.787335753440857,
|
633 |
+
"logps/rejected": -2.1021180152893066,
|
634 |
+
"loss": 0.5546,
|
635 |
+
"rewards/accuracies": 0.731249988079071,
|
636 |
+
"rewards/chosen": -1.8473552465438843,
|
637 |
+
"rewards/margins": 0.7361037731170654,
|
638 |
+
"rewards/rejected": -2.5834591388702393,
|
639 |
"step": 410
|
640 |
},
|
641 |
{
|
642 |
"epoch": 0.96,
|
643 |
+
"grad_norm": 32.802494867000924,
|
644 |
"learning_rate": 5.698048727497462e-10,
|
645 |
+
"logits/chosen": -1.7570117712020874,
|
646 |
+
"logits/rejected": -1.6864467859268188,
|
647 |
+
"logps/chosen": -1.8803859949111938,
|
648 |
+
"logps/rejected": -2.1749351024627686,
|
649 |
+
"loss": 0.5408,
|
650 |
+
"rewards/accuracies": 0.737500011920929,
|
651 |
+
"rewards/chosen": -2.0672643184661865,
|
652 |
+
"rewards/margins": 0.6827760338783264,
|
653 |
+
"rewards/rejected": -2.7500405311584473,
|
654 |
"step": 420
|
655 |
},
|
656 |
{
|
657 |
"epoch": 0.98,
|
658 |
+
"grad_norm": 30.44073541133352,
|
659 |
"learning_rate": 1.2803984447259387e-10,
|
660 |
+
"logits/chosen": -1.7648996114730835,
|
661 |
+
"logits/rejected": -1.713091254234314,
|
662 |
+
"logps/chosen": -1.8600318431854248,
|
663 |
+
"logps/rejected": -2.127103805541992,
|
664 |
+
"loss": 0.5345,
|
665 |
+
"rewards/accuracies": 0.71875,
|
666 |
+
"rewards/chosen": -2.0330777168273926,
|
667 |
+
"rewards/margins": 0.6582567095756531,
|
668 |
+
"rewards/rejected": -2.6913342475891113,
|
669 |
"step": 430
|
670 |
},
|
671 |
{
|
672 |
"epoch": 1.0,
|
673 |
"step": 439,
|
674 |
"total_flos": 0.0,
|
675 |
+
"train_loss": 0.5807303016441016,
|
676 |
+
"train_runtime": 6860.7908,
|
677 |
+
"train_samples_per_second": 8.197,
|
678 |
+
"train_steps_per_second": 0.064
|
679 |
}
|
680 |
],
|
681 |
"logging_steps": 10,
|
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 sha256:52fd958028d2a8e77c3db64f0cef52c04c727dbfa54965417b4be2b66b01aefd
|
3 |
size 6328
|