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/Jun22_03-41-55_n136-100-194/events.out.tfevents.1718998971.n136-100-194.1421217.0 +2 -2
- train_results.json +3 -3
- trainer_state.json +389 -389
- 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": 50000,
|
6 |
-
"train_samples_per_second": 8.
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 0.39275855483152927,
|
4 |
+
"train_runtime": 5850.2913,
|
5 |
"train_samples": 50000,
|
6 |
+
"train_samples_per_second": 8.547,
|
7 |
"train_steps_per_second": 0.067
|
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 sha256:1c0dc0c292d489e5548e3b75999fe6cd030919ab96386b961a7432dab2fccb25
|
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 sha256:78055b666c3b09cd419d126d46142b7e4cbc261fd5db79d249138904dfc1a57d
|
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 sha256:04307748a9391cbe41e028618c312d319041f81f61a38ff22b6d44f0630d8343
|
3 |
size 4540516344
|
runs/Jun22_03-41-55_n136-100-194/events.out.tfevents.1718998971.n136-100-194.1421217.0
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 sha256:529e639fdd42d7dbf95d6bf6be4135fb5040103f9df18e1b40660966896a5200
|
3 |
+
size 32644
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
-
"train_loss": 0.
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 50000,
|
6 |
-
"train_samples_per_second": 8.
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 1.0,
|
3 |
+
"train_loss": 0.39275855483152927,
|
4 |
+
"train_runtime": 5850.2913,
|
5 |
"train_samples": 50000,
|
6 |
+
"train_samples_per_second": 8.547,
|
7 |
"train_steps_per_second": 0.067
|
8 |
}
|
trainer_state.json
CHANGED
@@ -2,7 +2,7 @@
|
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
"epoch": 0.9980806142034548,
|
5 |
-
"eval_steps":
|
6 |
"global_step": 390,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
@@ -10,13 +10,13 @@
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0,
|
13 |
-
"grad_norm":
|
14 |
"learning_rate": 1.282051282051282e-08,
|
15 |
"logits/chosen": -2.5583817958831787,
|
16 |
"logits/rejected": -2.4487552642822266,
|
17 |
"logps/chosen": -258.1644592285156,
|
18 |
"logps/rejected": -216.25729370117188,
|
19 |
-
"loss": 0.
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
@@ -25,596 +25,596 @@
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.03,
|
28 |
-
"grad_norm":
|
29 |
"learning_rate": 1.2820512820512818e-07,
|
30 |
-
"logits/chosen": -2.
|
31 |
-
"logits/rejected": -2.
|
32 |
-
"logps/chosen": -267.
|
33 |
-
"logps/rejected": -217.
|
34 |
-
"loss": 0.
|
35 |
-
"rewards/accuracies": 0.
|
36 |
-
"rewards/chosen":
|
37 |
-
"rewards/margins":
|
38 |
-
"rewards/rejected":
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
-
"grad_norm":
|
44 |
"learning_rate": 2.5641025641025636e-07,
|
45 |
-
"logits/chosen": -2.
|
46 |
-
"logits/rejected": -2.
|
47 |
-
"logps/chosen": -260.
|
48 |
-
"logps/rejected": -207.
|
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": 0.08,
|
58 |
-
"grad_norm":
|
59 |
"learning_rate": 3.8461538461538463e-07,
|
60 |
-
"logits/chosen": -2.
|
61 |
-
"logits/rejected": -2.
|
62 |
-
"logps/chosen": -250.
|
63 |
-
"logps/rejected": -198.
|
64 |
-
"loss": 0.
|
65 |
-
"rewards/accuracies": 0.
|
66 |
-
"rewards/chosen":
|
67 |
-
"rewards/margins": 2.
|
68 |
-
"rewards/rejected": 0.
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.1,
|
73 |
-
"grad_norm":
|
74 |
"learning_rate": 4.99989986344963e-07,
|
75 |
-
"logits/chosen": -2.
|
76 |
-
"logits/rejected": -2.
|
77 |
-
"logps/chosen": -243.
|
78 |
-
"logps/rejected": -
|
79 |
-
"loss": 0.
|
80 |
"rewards/accuracies": 0.824999988079071,
|
81 |
-
"rewards/chosen":
|
82 |
-
"rewards/margins": 4.
|
83 |
-
"rewards/rejected": 0.
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.13,
|
88 |
-
"grad_norm":
|
89 |
"learning_rate": 4.987893180827479e-07,
|
90 |
-
"logits/chosen": -2.
|
91 |
-
"logits/rejected": -2.
|
92 |
-
"logps/chosen": -257.
|
93 |
-
"logps/rejected": -203.
|
94 |
-
"loss": 0.
|
95 |
-
"rewards/accuracies": 0.
|
96 |
-
"rewards/chosen": 7.
|
97 |
-
"rewards/margins": 6.
|
98 |
-
"rewards/rejected": 1.
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.15,
|
103 |
-
"grad_norm":
|
104 |
"learning_rate": 4.955969343539162e-07,
|
105 |
-
"logits/chosen": -2.
|
106 |
-
"logits/rejected": -2.
|
107 |
-
"logps/chosen": -261.
|
108 |
-
"logps/rejected": -209.
|
109 |
-
"loss": 0.
|
110 |
-
"rewards/accuracies": 0.
|
111 |
-
"rewards/chosen": 4.
|
112 |
-
"rewards/margins": 5.
|
113 |
-
"rewards/rejected": -0.
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.18,
|
118 |
-
"grad_norm":
|
119 |
"learning_rate": 4.90438392204474e-07,
|
120 |
-
"logits/chosen": -2.
|
121 |
-
"logits/rejected": -2.
|
122 |
-
"logps/chosen": -291.
|
123 |
-
"logps/rejected": -
|
124 |
-
"loss": 0.
|
125 |
-
"rewards/accuracies": 0.
|
126 |
-
"rewards/chosen": 4.
|
127 |
-
"rewards/margins": 6.
|
128 |
-
"rewards/rejected": -
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.2,
|
133 |
-
"grad_norm":
|
134 |
"learning_rate": 4.83354989019146e-07,
|
135 |
-
"logits/chosen": -2.
|
136 |
-
"logits/rejected": -2.
|
137 |
-
"logps/chosen": -259.
|
138 |
-
"logps/rejected": -204.
|
139 |
-
"loss": 0.
|
140 |
"rewards/accuracies": 0.887499988079071,
|
141 |
-
"rewards/chosen": 5.
|
142 |
-
"rewards/margins": 6.
|
143 |
-
"rewards/rejected": -1.
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.23,
|
148 |
-
"grad_norm":
|
149 |
"learning_rate": 4.7440343190975353e-07,
|
150 |
-
"logits/chosen": -2.
|
151 |
-
"logits/rejected": -2.
|
152 |
-
"logps/chosen": -257.
|
153 |
-
"logps/rejected": -217.
|
154 |
-
"loss": 0.
|
155 |
-
"rewards/accuracies": 0.
|
156 |
-
"rewards/chosen":
|
157 |
-
"rewards/margins": 5.
|
158 |
-
"rewards/rejected": -2.
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.26,
|
163 |
-
"grad_norm":
|
164 |
"learning_rate": 4.6365538373900506e-07,
|
165 |
-
"logits/chosen": -2.
|
166 |
-
"logits/rejected": -2.
|
167 |
-
"logps/chosen": -236.
|
168 |
-
"logps/rejected": -201.
|
169 |
-
"loss": 0.
|
170 |
-
"rewards/accuracies": 0.
|
171 |
-
"rewards/chosen": 4.
|
172 |
-
"rewards/margins": 6.
|
173 |
-
"rewards/rejected": -
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.28,
|
178 |
-
"grad_norm":
|
179 |
"learning_rate": 4.5119688941406386e-07,
|
180 |
-
"logits/chosen": -2.
|
181 |
-
"logits/rejected": -2.
|
182 |
-
"logps/chosen": -256.
|
183 |
-
"logps/rejected": -209.
|
184 |
-
"loss": 0.
|
185 |
-
"rewards/accuracies": 0.
|
186 |
-
"rewards/chosen": 6.
|
187 |
-
"rewards/margins": 6.
|
188 |
-
"rewards/rejected": -0.
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.31,
|
193 |
-
"grad_norm":
|
194 |
"learning_rate": 4.3712768704277524e-07,
|
195 |
-
"logits/chosen": -2.
|
196 |
-
"logits/rejected": -2.
|
197 |
-
"logps/chosen": -261.
|
198 |
-
"logps/rejected": -208.
|
199 |
-
"loss": 0.
|
200 |
"rewards/accuracies": 0.862500011920929,
|
201 |
-
"rewards/chosen": 5.
|
202 |
-
"rewards/margins": 6.
|
203 |
-
"rewards/rejected": -1.
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.33,
|
208 |
-
"grad_norm":
|
209 |
"learning_rate": 4.2156040946718343e-07,
|
210 |
-
"logits/chosen": -2.
|
211 |
-
"logits/rejected": -2.
|
212 |
-
"logps/chosen": -
|
213 |
-
"logps/rejected": -
|
214 |
-
"loss": 0.
|
215 |
-
"rewards/accuracies": 0.
|
216 |
-
"rewards/chosen":
|
217 |
-
"rewards/margins": 6.
|
218 |
-
"rewards/rejected": -2.
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.36,
|
223 |
-
"grad_norm":
|
224 |
"learning_rate": 4.046196825665637e-07,
|
225 |
-
"logits/chosen": -2.
|
226 |
-
"logits/rejected": -2.
|
227 |
-
"logps/chosen": -
|
228 |
-
"logps/rejected": -
|
229 |
-
"loss": 0.
|
230 |
-
"rewards/accuracies": 0.
|
231 |
-
"rewards/chosen":
|
232 |
-
"rewards/margins": 6.
|
233 |
-
"rewards/rejected": -
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.38,
|
238 |
-
"grad_norm":
|
239 |
"learning_rate": 3.864411275486261e-07,
|
240 |
-
"logits/chosen": -2.
|
241 |
-
"logits/rejected": -2.
|
242 |
-
"logps/chosen": -
|
243 |
-
"logps/rejected": -
|
244 |
-
"loss": 0.
|
245 |
-
"rewards/accuracies": 0.
|
246 |
-
"rewards/chosen": 5.
|
247 |
-
"rewards/margins": 7.
|
248 |
-
"rewards/rejected": -1.
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.41,
|
253 |
-
"grad_norm":
|
254 |
"learning_rate": 3.671702752161759e-07,
|
255 |
-
"logits/chosen": -2.
|
256 |
-
"logits/rejected": -2.
|
257 |
-
"logps/chosen": -
|
258 |
-
"logps/rejected": -
|
259 |
-
"loss": 0.
|
260 |
-
"rewards/accuracies": 0.
|
261 |
-
"rewards/chosen":
|
262 |
-
"rewards/margins": 7.
|
263 |
-
"rewards/rejected": -
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.44,
|
268 |
-
"grad_norm":
|
269 |
"learning_rate": 3.4696140090121375e-07,
|
270 |
-
"logits/chosen": -2.
|
271 |
-
"logits/rejected": -2.
|
272 |
-
"logps/chosen": -
|
273 |
-
"logps/rejected": -
|
274 |
-
"loss": 0.
|
275 |
-
"rewards/accuracies": 0.
|
276 |
-
"rewards/chosen":
|
277 |
-
"rewards/margins": 7.
|
278 |
-
"rewards/rejected": -
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.46,
|
283 |
-
"grad_norm":
|
284 |
"learning_rate": 3.259762893935617e-07,
|
285 |
-
"logits/chosen": -2.
|
286 |
-
"logits/rejected": -2.
|
287 |
-
"logps/chosen": -
|
288 |
-
"logps/rejected": -
|
289 |
-
"loss": 0.
|
290 |
-
"rewards/accuracies": 0.
|
291 |
-
"rewards/chosen":
|
292 |
-
"rewards/margins": 6.
|
293 |
-
"rewards/rejected": -
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.49,
|
298 |
-
"grad_norm":
|
299 |
"learning_rate": 3.0438293975154184e-07,
|
300 |
-
"logits/chosen": -2.
|
301 |
-
"logits/rejected": -2.
|
302 |
-
"logps/chosen": -
|
303 |
-
"logps/rejected": -
|
304 |
-
"loss": 0.
|
305 |
"rewards/accuracies": 0.8812500238418579,
|
306 |
-
"rewards/chosen":
|
307 |
-
"rewards/margins": 7.
|
308 |
-
"rewards/rejected": -
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.51,
|
313 |
-
"grad_norm":
|
314 |
"learning_rate": 2.823542203635138e-07,
|
315 |
-
"logits/chosen": -2.
|
316 |
-
"logits/rejected": -2.
|
317 |
-
"logps/chosen": -
|
318 |
-
"logps/rejected": -
|
319 |
-
"loss": 0.
|
320 |
-
"rewards/accuracies": 0.
|
321 |
-
"rewards/chosen":
|
322 |
-
"rewards/margins": 8.
|
323 |
-
"rewards/rejected": -
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.54,
|
328 |
-
"grad_norm":
|
329 |
"learning_rate": 2.600664850273538e-07,
|
330 |
-
"logits/chosen": -2.
|
331 |
-
"logits/rejected": -2.
|
332 |
-
"logps/chosen": -
|
333 |
-
"logps/rejected": -
|
334 |
-
"loss": 0.
|
335 |
"rewards/accuracies": 0.856249988079071,
|
336 |
-
"rewards/chosen":
|
337 |
-
"rewards/margins": 6.
|
338 |
-
"rewards/rejected": -
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.56,
|
343 |
-
"grad_norm":
|
344 |
"learning_rate": 2.3769816112703045e-07,
|
345 |
-
"logits/chosen": -2.
|
346 |
-
"logits/rejected": -2.
|
347 |
-
"logps/chosen": -
|
348 |
-
"logps/rejected": -
|
349 |
-
"loss": 0.
|
350 |
-
"rewards/accuracies": 0.
|
351 |
-
"rewards/chosen":
|
352 |
-
"rewards/margins": 6.
|
353 |
-
"rewards/rejected": -
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.59,
|
358 |
-
"grad_norm":
|
359 |
"learning_rate": 2.1542832120881677e-07,
|
360 |
-
"logits/chosen": -2.
|
361 |
-
"logits/rejected": -2.
|
362 |
-
"logps/chosen": -
|
363 |
-
"logps/rejected": -
|
364 |
-
"loss": 0.
|
365 |
-
"rewards/accuracies": 0.
|
366 |
-
"rewards/chosen":
|
367 |
-
"rewards/margins": 7.
|
368 |
-
"rewards/rejected": -
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.61,
|
373 |
-
"grad_norm":
|
374 |
"learning_rate": 1.934352493925695e-07,
|
375 |
-
"logits/chosen": -2.
|
376 |
-
"logits/rejected": -2.
|
377 |
-
"logps/chosen": -
|
378 |
-
"logps/rejected": -
|
379 |
-
"loss": 0.
|
380 |
"rewards/accuracies": 0.893750011920929,
|
381 |
-
"rewards/chosen":
|
382 |
-
"rewards/margins": 8.
|
383 |
-
"rewards/rejected": -
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.64,
|
388 |
-
"grad_norm":
|
389 |
"learning_rate": 1.7189501409486059e-07,
|
390 |
-
"logits/chosen": -2.
|
391 |
-
"logits/rejected": -2.
|
392 |
-
"logps/chosen": -
|
393 |
-
"logps/rejected": -
|
394 |
-
"loss": 0.
|
395 |
-
"rewards/accuracies": 0.
|
396 |
-
"rewards/chosen":
|
397 |
-
"rewards/margins": 7.
|
398 |
-
"rewards/rejected": -
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.67,
|
403 |
-
"grad_norm":
|
404 |
"learning_rate": 1.5098005849021078e-07,
|
405 |
-
"logits/chosen": -2.
|
406 |
-
"logits/rejected": -2.
|
407 |
-
"logps/chosen": -
|
408 |
-
"logps/rejected": -
|
409 |
-
"loss": 0.
|
410 |
-
"rewards/accuracies": 0.
|
411 |
-
"rewards/chosen":
|
412 |
-
"rewards/margins": 6.
|
413 |
-
"rewards/rejected": -
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.69,
|
418 |
-
"grad_norm":
|
419 |
"learning_rate": 1.30857819994673e-07,
|
420 |
-
"logits/chosen": -2.
|
421 |
-
"logits/rejected": -2.
|
422 |
-
"logps/chosen": -
|
423 |
-
"logps/rejected": -
|
424 |
-
"loss": 0.
|
425 |
-
"rewards/accuracies": 0.
|
426 |
-
"rewards/chosen":
|
427 |
-
"rewards/margins":
|
428 |
-
"rewards/rejected": -
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.72,
|
433 |
-
"grad_norm":
|
434 |
"learning_rate": 1.116893898236716e-07,
|
435 |
-
"logits/chosen": -2.
|
436 |
-
"logits/rejected": -2.
|
437 |
-
"logps/chosen": -
|
438 |
-
"logps/rejected": -
|
439 |
-
"loss": 0.
|
440 |
-
"rewards/accuracies": 0.
|
441 |
-
"rewards/chosen":
|
442 |
-
"rewards/margins": 7.
|
443 |
-
"rewards/rejected": -
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.74,
|
448 |
-
"grad_norm":
|
449 |
"learning_rate": 9.362822335518062e-08,
|
450 |
-
"logits/chosen": -2.
|
451 |
-
"logits/rejected": -2.
|
452 |
-
"logps/chosen": -
|
453 |
-
"logps/rejected": -
|
454 |
-
"loss": 0.
|
455 |
-
"rewards/accuracies": 0.
|
456 |
-
"rewards/chosen":
|
457 |
-
"rewards/margins": 7.
|
458 |
-
"rewards/rejected": -
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.77,
|
463 |
-
"grad_norm":
|
464 |
"learning_rate": 7.681891162260015e-08,
|
465 |
-
"logits/chosen": -2.
|
466 |
-
"logits/rejected": -2.
|
467 |
-
"logps/chosen": -
|
468 |
-
"logps/rejected": -
|
469 |
-
"loss": 0.
|
470 |
-
"rewards/accuracies": 0.
|
471 |
-
"rewards/chosen":
|
472 |
-
"rewards/margins": 7.
|
473 |
-
"rewards/rejected": -
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.79,
|
478 |
-
"grad_norm":
|
479 |
"learning_rate": 6.139602377230247e-08,
|
480 |
-
"logits/chosen": -2.
|
481 |
-
"logits/rejected": -2.
|
482 |
-
"logps/chosen": -
|
483 |
-
"logps/rejected": -
|
484 |
-
"loss": 0.
|
485 |
-
"rewards/accuracies": 0.
|
486 |
-
"rewards/chosen":
|
487 |
-
"rewards/margins": 7.
|
488 |
-
"rewards/rejected": -
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.82,
|
493 |
-
"grad_norm":
|
494 |
"learning_rate": 4.748302975270837e-08,
|
495 |
-
"logits/chosen": -2.
|
496 |
-
"logits/rejected": -2.
|
497 |
-
"logps/chosen": -
|
498 |
-
"logps/rejected": -
|
499 |
-
"loss": 0.
|
500 |
"rewards/accuracies": 0.8812500238418579,
|
501 |
-
"rewards/chosen":
|
502 |
-
"rewards/margins": 6.
|
503 |
-
"rewards/rejected": -
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.84,
|
508 |
-
"grad_norm":
|
509 |
"learning_rate": 3.5191311859445795e-08,
|
510 |
-
"logits/chosen": -2.
|
511 |
-
"logits/rejected": -2.
|
512 |
-
"logps/chosen": -
|
513 |
-
"logps/rejected": -
|
514 |
-
"loss": 0.
|
515 |
-
"rewards/accuracies": 0.
|
516 |
-
"rewards/chosen":
|
517 |
-
"rewards/margins": 7.
|
518 |
-
"rewards/rejected": -
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.87,
|
523 |
-
"grad_norm":
|
524 |
"learning_rate": 2.4619273049795996e-08,
|
525 |
-
"logits/chosen": -2.
|
526 |
-
"logits/rejected": -2.
|
527 |
-
"logps/chosen": -
|
528 |
-
"logps/rejected": -
|
529 |
-
"loss": 0.
|
530 |
-
"rewards/accuracies": 0.
|
531 |
-
"rewards/chosen":
|
532 |
-
"rewards/margins": 7.
|
533 |
-
"rewards/rejected": -
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.9,
|
538 |
-
"grad_norm":
|
539 |
"learning_rate": 1.5851549164932115e-08,
|
540 |
-
"logits/chosen": -2.
|
541 |
-
"logits/rejected": -2.
|
542 |
-
"logps/chosen": -
|
543 |
-
"logps/rejected": -
|
544 |
-
"loss": 0.
|
545 |
-
"rewards/accuracies": 0.
|
546 |
-
"rewards/chosen":
|
547 |
-
"rewards/margins": 7.
|
548 |
-
"rewards/rejected": -
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.92,
|
553 |
-
"grad_norm":
|
554 |
"learning_rate": 8.958331366609423e-09,
|
555 |
-
"logits/chosen": -2.
|
556 |
-
"logits/rejected": -2.
|
557 |
-
"logps/chosen": -
|
558 |
-
"logps/rejected": -
|
559 |
-
"loss": 0.
|
560 |
-
"rewards/accuracies": 0.
|
561 |
-
"rewards/chosen":
|
562 |
-
"rewards/margins": 7.
|
563 |
-
"rewards/rejected": -
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.95,
|
568 |
-
"grad_norm":
|
569 |
"learning_rate": 3.994804212627461e-09,
|
570 |
-
"logits/chosen": -2.
|
571 |
-
"logits/rejected": -2.
|
572 |
-
"logps/chosen": -
|
573 |
-
"logps/rejected": -
|
574 |
-
"loss": 0.
|
575 |
-
"rewards/accuracies": 0.
|
576 |
-
"rewards/chosen":
|
577 |
-
"rewards/margins": 7.
|
578 |
-
"rewards/rejected": -
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.97,
|
583 |
-
"grad_norm":
|
584 |
"learning_rate": 1.0007038696262516e-09,
|
585 |
-
"logits/chosen": -2.
|
586 |
-
"logits/rejected": -2.
|
587 |
-
"logps/chosen": -
|
588 |
-
"logps/rejected": -
|
589 |
-
"loss": 0.
|
590 |
-
"rewards/accuracies": 0.
|
591 |
-
"rewards/chosen":
|
592 |
-
"rewards/margins": 7.
|
593 |
-
"rewards/rejected": -
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 1.0,
|
598 |
-
"grad_norm":
|
599 |
"learning_rate": 0.0,
|
600 |
-
"logits/chosen": -2.
|
601 |
-
"logits/rejected": -2.
|
602 |
-
"logps/chosen": -
|
603 |
-
"logps/rejected": -
|
604 |
-
"loss": 0.
|
605 |
-
"rewards/accuracies": 0.
|
606 |
-
"rewards/chosen":
|
607 |
-
"rewards/margins":
|
608 |
-
"rewards/rejected": -
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 1.0,
|
613 |
"step": 390,
|
614 |
"total_flos": 0.0,
|
615 |
-
"train_loss": 0.
|
616 |
-
"train_runtime":
|
617 |
-
"train_samples_per_second": 8.
|
618 |
"train_steps_per_second": 0.067
|
619 |
}
|
620 |
],
|
|
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
"epoch": 0.9980806142034548,
|
5 |
+
"eval_steps": 1000000,
|
6 |
"global_step": 390,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0,
|
13 |
+
"grad_norm": 992.8142742850837,
|
14 |
"learning_rate": 1.282051282051282e-08,
|
15 |
"logits/chosen": -2.5583817958831787,
|
16 |
"logits/rejected": -2.4487552642822266,
|
17 |
"logps/chosen": -258.1644592285156,
|
18 |
"logps/rejected": -216.25729370117188,
|
19 |
+
"loss": 0.6965,
|
20 |
"rewards/accuracies": 0.0,
|
21 |
"rewards/chosen": 0.0,
|
22 |
"rewards/margins": 0.0,
|
|
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.03,
|
28 |
+
"grad_norm": 972.5242632724868,
|
29 |
"learning_rate": 1.2820512820512818e-07,
|
30 |
+
"logits/chosen": -2.606093168258667,
|
31 |
+
"logits/rejected": -2.5532331466674805,
|
32 |
+
"logps/chosen": -267.5519714355469,
|
33 |
+
"logps/rejected": -217.63278198242188,
|
34 |
+
"loss": 0.7101,
|
35 |
+
"rewards/accuracies": 0.4375,
|
36 |
+
"rewards/chosen": 0.008765195496380329,
|
37 |
+
"rewards/margins": 0.004809141159057617,
|
38 |
+
"rewards/rejected": 0.0039560552686452866,
|
39 |
"step": 10
|
40 |
},
|
41 |
{
|
42 |
"epoch": 0.05,
|
43 |
+
"grad_norm": 659.3539113664132,
|
44 |
"learning_rate": 2.5641025641025636e-07,
|
45 |
+
"logits/chosen": -2.6304450035095215,
|
46 |
+
"logits/rejected": -2.5677618980407715,
|
47 |
+
"logps/chosen": -260.57421875,
|
48 |
+
"logps/rejected": -207.02157592773438,
|
49 |
+
"loss": 0.5261,
|
50 |
+
"rewards/accuracies": 0.7250000238418579,
|
51 |
+
"rewards/chosen": 0.5109978318214417,
|
52 |
+
"rewards/margins": 0.48820367455482483,
|
53 |
+
"rewards/rejected": 0.022794198244810104,
|
54 |
"step": 20
|
55 |
},
|
56 |
{
|
57 |
"epoch": 0.08,
|
58 |
+
"grad_norm": 779.0248633950893,
|
59 |
"learning_rate": 3.8461538461538463e-07,
|
60 |
+
"logits/chosen": -2.648696184158325,
|
61 |
+
"logits/rejected": -2.573786973953247,
|
62 |
+
"logps/chosen": -250.76904296875,
|
63 |
+
"logps/rejected": -198.43540954589844,
|
64 |
+
"loss": 0.3369,
|
65 |
+
"rewards/accuracies": 0.856249988079071,
|
66 |
+
"rewards/chosen": 3.0647289752960205,
|
67 |
+
"rewards/margins": 2.7615580558776855,
|
68 |
+
"rewards/rejected": 0.3031708300113678,
|
69 |
"step": 30
|
70 |
},
|
71 |
{
|
72 |
"epoch": 0.1,
|
73 |
+
"grad_norm": 384.3988310437092,
|
74 |
"learning_rate": 4.99989986344963e-07,
|
75 |
+
"logits/chosen": -2.6418910026550293,
|
76 |
+
"logits/rejected": -2.5625271797180176,
|
77 |
+
"logps/chosen": -243.13668823242188,
|
78 |
+
"logps/rejected": -192.8317108154297,
|
79 |
+
"loss": 0.3186,
|
80 |
"rewards/accuracies": 0.824999988079071,
|
81 |
+
"rewards/chosen": 5.048842430114746,
|
82 |
+
"rewards/margins": 4.450641632080078,
|
83 |
+
"rewards/rejected": 0.5982006192207336,
|
84 |
"step": 40
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.13,
|
88 |
+
"grad_norm": 557.6508821014772,
|
89 |
"learning_rate": 4.987893180827479e-07,
|
90 |
+
"logits/chosen": -2.651369333267212,
|
91 |
+
"logits/rejected": -2.5784316062927246,
|
92 |
+
"logps/chosen": -257.6221923828125,
|
93 |
+
"logps/rejected": -203.21920776367188,
|
94 |
+
"loss": 0.3491,
|
95 |
+
"rewards/accuracies": 0.862500011920929,
|
96 |
+
"rewards/chosen": 7.43048620223999,
|
97 |
+
"rewards/margins": 6.079649448394775,
|
98 |
+
"rewards/rejected": 1.3508365154266357,
|
99 |
"step": 50
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.15,
|
103 |
+
"grad_norm": 595.0405769986861,
|
104 |
"learning_rate": 4.955969343539162e-07,
|
105 |
+
"logits/chosen": -2.606095552444458,
|
106 |
+
"logits/rejected": -2.5313313007354736,
|
107 |
+
"logps/chosen": -261.7010192871094,
|
108 |
+
"logps/rejected": -209.05029296875,
|
109 |
+
"loss": 0.3678,
|
110 |
+
"rewards/accuracies": 0.8687499761581421,
|
111 |
+
"rewards/chosen": 4.988667011260986,
|
112 |
+
"rewards/margins": 5.682937145233154,
|
113 |
+
"rewards/rejected": -0.6942700147628784,
|
114 |
"step": 60
|
115 |
},
|
116 |
{
|
117 |
"epoch": 0.18,
|
118 |
+
"grad_norm": 406.24830016448993,
|
119 |
"learning_rate": 4.90438392204474e-07,
|
120 |
+
"logits/chosen": -2.57716703414917,
|
121 |
+
"logits/rejected": -2.5019829273223877,
|
122 |
+
"logps/chosen": -291.1893615722656,
|
123 |
+
"logps/rejected": -227.8859100341797,
|
124 |
+
"loss": 0.3366,
|
125 |
+
"rewards/accuracies": 0.9125000238418579,
|
126 |
+
"rewards/chosen": 4.704905033111572,
|
127 |
+
"rewards/margins": 6.591601371765137,
|
128 |
+
"rewards/rejected": -1.8866965770721436,
|
129 |
"step": 70
|
130 |
},
|
131 |
{
|
132 |
"epoch": 0.2,
|
133 |
+
"grad_norm": 771.1936928093826,
|
134 |
"learning_rate": 4.83354989019146e-07,
|
135 |
+
"logits/chosen": -2.534698247909546,
|
136 |
+
"logits/rejected": -2.4588425159454346,
|
137 |
+
"logps/chosen": -259.4831848144531,
|
138 |
+
"logps/rejected": -204.6412353515625,
|
139 |
+
"loss": 0.3162,
|
140 |
"rewards/accuracies": 0.887499988079071,
|
141 |
+
"rewards/chosen": 5.652478218078613,
|
142 |
+
"rewards/margins": 6.747942924499512,
|
143 |
+
"rewards/rejected": -1.0954645872116089,
|
144 |
"step": 80
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.23,
|
148 |
+
"grad_norm": 746.9475535161084,
|
149 |
"learning_rate": 4.7440343190975353e-07,
|
150 |
+
"logits/chosen": -2.5645816326141357,
|
151 |
+
"logits/rejected": -2.5068752765655518,
|
152 |
+
"logps/chosen": -257.1772155761719,
|
153 |
+
"logps/rejected": -217.83139038085938,
|
154 |
+
"loss": 0.31,
|
155 |
+
"rewards/accuracies": 0.84375,
|
156 |
+
"rewards/chosen": 3.197655200958252,
|
157 |
+
"rewards/margins": 5.532810211181641,
|
158 |
+
"rewards/rejected": -2.3351550102233887,
|
159 |
"step": 90
|
160 |
},
|
161 |
{
|
162 |
"epoch": 0.26,
|
163 |
+
"grad_norm": 823.7911810864297,
|
164 |
"learning_rate": 4.6365538373900506e-07,
|
165 |
+
"logits/chosen": -2.6224541664123535,
|
166 |
+
"logits/rejected": -2.5483615398406982,
|
167 |
+
"logps/chosen": -236.20339965820312,
|
168 |
+
"logps/rejected": -201.47940063476562,
|
169 |
+
"loss": 0.5348,
|
170 |
+
"rewards/accuracies": 0.8500000238418579,
|
171 |
+
"rewards/chosen": 4.17586088180542,
|
172 |
+
"rewards/margins": 6.1418962478637695,
|
173 |
+
"rewards/rejected": -1.9660362005233765,
|
174 |
"step": 100
|
175 |
},
|
176 |
{
|
177 |
"epoch": 0.28,
|
178 |
+
"grad_norm": 579.991739437595,
|
179 |
"learning_rate": 4.5119688941406386e-07,
|
180 |
+
"logits/chosen": -2.6264939308166504,
|
181 |
+
"logits/rejected": -2.5470776557922363,
|
182 |
+
"logps/chosen": -256.7718200683594,
|
183 |
+
"logps/rejected": -209.46231079101562,
|
184 |
+
"loss": 0.4091,
|
185 |
+
"rewards/accuracies": 0.856249988079071,
|
186 |
+
"rewards/chosen": 6.23184061050415,
|
187 |
+
"rewards/margins": 6.958665370941162,
|
188 |
+
"rewards/rejected": -0.7268255352973938,
|
189 |
"step": 110
|
190 |
},
|
191 |
{
|
192 |
"epoch": 0.31,
|
193 |
+
"grad_norm": 690.1539082640518,
|
194 |
"learning_rate": 4.3712768704277524e-07,
|
195 |
+
"logits/chosen": -2.6043858528137207,
|
196 |
+
"logits/rejected": -2.5360493659973145,
|
197 |
+
"logps/chosen": -261.69561767578125,
|
198 |
+
"logps/rejected": -208.3325653076172,
|
199 |
+
"loss": 0.4048,
|
200 |
"rewards/accuracies": 0.862500011920929,
|
201 |
+
"rewards/chosen": 5.436749458312988,
|
202 |
+
"rewards/margins": 6.8130621910095215,
|
203 |
+
"rewards/rejected": -1.3763134479522705,
|
204 |
"step": 120
|
205 |
},
|
206 |
{
|
207 |
"epoch": 0.33,
|
208 |
+
"grad_norm": 451.1743521072461,
|
209 |
"learning_rate": 4.2156040946718343e-07,
|
210 |
+
"logits/chosen": -2.578007698059082,
|
211 |
+
"logits/rejected": -2.5135886669158936,
|
212 |
+
"logps/chosen": -250.7688446044922,
|
213 |
+
"logps/rejected": -196.73779296875,
|
214 |
+
"loss": 0.3698,
|
215 |
+
"rewards/accuracies": 0.8374999761581421,
|
216 |
+
"rewards/chosen": 4.574068069458008,
|
217 |
+
"rewards/margins": 6.677270412445068,
|
218 |
+
"rewards/rejected": -2.1032023429870605,
|
219 |
"step": 130
|
220 |
},
|
221 |
{
|
222 |
"epoch": 0.36,
|
223 |
+
"grad_norm": 465.41383196085104,
|
224 |
"learning_rate": 4.046196825665637e-07,
|
225 |
+
"logits/chosen": -2.600147247314453,
|
226 |
+
"logits/rejected": -2.534235715866089,
|
227 |
+
"logps/chosen": -269.448974609375,
|
228 |
+
"logps/rejected": -216.6135711669922,
|
229 |
+
"loss": 0.4171,
|
230 |
+
"rewards/accuracies": 0.862500011920929,
|
231 |
+
"rewards/chosen": 4.009121417999268,
|
232 |
+
"rewards/margins": 6.148682594299316,
|
233 |
+
"rewards/rejected": -2.139561176300049,
|
234 |
"step": 140
|
235 |
},
|
236 |
{
|
237 |
"epoch": 0.38,
|
238 |
+
"grad_norm": 485.4989695160128,
|
239 |
"learning_rate": 3.864411275486261e-07,
|
240 |
+
"logits/chosen": -2.583752155303955,
|
241 |
+
"logits/rejected": -2.51660418510437,
|
242 |
+
"logps/chosen": -262.35479736328125,
|
243 |
+
"logps/rejected": -212.38021850585938,
|
244 |
+
"loss": 0.429,
|
245 |
+
"rewards/accuracies": 0.856249988079071,
|
246 |
+
"rewards/chosen": 5.765866756439209,
|
247 |
+
"rewards/margins": 7.155278205871582,
|
248 |
+
"rewards/rejected": -1.3894122838974,
|
249 |
"step": 150
|
250 |
},
|
251 |
{
|
252 |
"epoch": 0.41,
|
253 |
+
"grad_norm": 546.3479101912284,
|
254 |
"learning_rate": 3.671702752161759e-07,
|
255 |
+
"logits/chosen": -2.5821540355682373,
|
256 |
+
"logits/rejected": -2.514200210571289,
|
257 |
+
"logps/chosen": -242.9073944091797,
|
258 |
+
"logps/rejected": -197.12026977539062,
|
259 |
+
"loss": 0.4562,
|
260 |
+
"rewards/accuracies": 0.8812500238418579,
|
261 |
+
"rewards/chosen": 4.221369743347168,
|
262 |
+
"rewards/margins": 7.090517520904541,
|
263 |
+
"rewards/rejected": -2.869147777557373,
|
264 |
"step": 160
|
265 |
},
|
266 |
{
|
267 |
"epoch": 0.44,
|
268 |
+
"grad_norm": 655.017465367729,
|
269 |
"learning_rate": 3.4696140090121375e-07,
|
270 |
+
"logits/chosen": -2.582387685775757,
|
271 |
+
"logits/rejected": -2.517115831375122,
|
272 |
+
"logps/chosen": -264.36541748046875,
|
273 |
+
"logps/rejected": -210.4468231201172,
|
274 |
+
"loss": 0.3517,
|
275 |
+
"rewards/accuracies": 0.875,
|
276 |
+
"rewards/chosen": 4.428440570831299,
|
277 |
+
"rewards/margins": 7.276960849761963,
|
278 |
+
"rewards/rejected": -2.8485207557678223,
|
279 |
"step": 170
|
280 |
},
|
281 |
{
|
282 |
"epoch": 0.46,
|
283 |
+
"grad_norm": 492.2183630043871,
|
284 |
"learning_rate": 3.259762893935617e-07,
|
285 |
+
"logits/chosen": -2.6398727893829346,
|
286 |
+
"logits/rejected": -2.5528438091278076,
|
287 |
+
"logps/chosen": -235.0208282470703,
|
288 |
+
"logps/rejected": -185.29156494140625,
|
289 |
+
"loss": 0.4065,
|
290 |
+
"rewards/accuracies": 0.856249988079071,
|
291 |
+
"rewards/chosen": 4.908794403076172,
|
292 |
+
"rewards/margins": 6.33758544921875,
|
293 |
+
"rewards/rejected": -1.4287910461425781,
|
294 |
"step": 180
|
295 |
},
|
296 |
{
|
297 |
"epoch": 0.49,
|
298 |
+
"grad_norm": 502.0957454157663,
|
299 |
"learning_rate": 3.0438293975154184e-07,
|
300 |
+
"logits/chosen": -2.5951266288757324,
|
301 |
+
"logits/rejected": -2.519169330596924,
|
302 |
+
"logps/chosen": -258.04022216796875,
|
303 |
+
"logps/rejected": -202.89727783203125,
|
304 |
+
"loss": 0.3426,
|
305 |
"rewards/accuracies": 0.8812500238418579,
|
306 |
+
"rewards/chosen": 4.746270179748535,
|
307 |
+
"rewards/margins": 7.24340295791626,
|
308 |
+
"rewards/rejected": -2.4971330165863037,
|
309 |
"step": 190
|
310 |
},
|
311 |
{
|
312 |
"epoch": 0.51,
|
313 |
+
"grad_norm": 536.1992163678008,
|
314 |
"learning_rate": 2.823542203635138e-07,
|
315 |
+
"logits/chosen": -2.6266205310821533,
|
316 |
+
"logits/rejected": -2.5365912914276123,
|
317 |
+
"logps/chosen": -273.0335388183594,
|
318 |
+
"logps/rejected": -217.910888671875,
|
319 |
+
"loss": 0.3772,
|
320 |
+
"rewards/accuracies": 0.875,
|
321 |
+
"rewards/chosen": 5.242685794830322,
|
322 |
+
"rewards/margins": 8.247489929199219,
|
323 |
+
"rewards/rejected": -3.004803419113159,
|
324 |
"step": 200
|
325 |
},
|
326 |
{
|
327 |
"epoch": 0.54,
|
328 |
+
"grad_norm": 550.4099704660905,
|
329 |
"learning_rate": 2.600664850273538e-07,
|
330 |
+
"logits/chosen": -2.6228280067443848,
|
331 |
+
"logits/rejected": -2.550497531890869,
|
332 |
+
"logps/chosen": -265.55755615234375,
|
333 |
+
"logps/rejected": -210.61660766601562,
|
334 |
+
"loss": 0.5254,
|
335 |
"rewards/accuracies": 0.856249988079071,
|
336 |
+
"rewards/chosen": 3.847407102584839,
|
337 |
+
"rewards/margins": 6.826489448547363,
|
338 |
+
"rewards/rejected": -2.979081869125366,
|
339 |
"step": 210
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.56,
|
343 |
+
"grad_norm": 581.5523533407168,
|
344 |
"learning_rate": 2.3769816112703045e-07,
|
345 |
+
"logits/chosen": -2.6491611003875732,
|
346 |
+
"logits/rejected": -2.5865907669067383,
|
347 |
+
"logps/chosen": -255.453857421875,
|
348 |
+
"logps/rejected": -212.5009765625,
|
349 |
+
"loss": 0.4229,
|
350 |
+
"rewards/accuracies": 0.8687499761581421,
|
351 |
+
"rewards/chosen": 4.356105327606201,
|
352 |
+
"rewards/margins": 6.26999568939209,
|
353 |
+
"rewards/rejected": -1.9138901233673096,
|
354 |
"step": 220
|
355 |
},
|
356 |
{
|
357 |
"epoch": 0.59,
|
358 |
+
"grad_norm": 486.64295333011063,
|
359 |
"learning_rate": 2.1542832120881677e-07,
|
360 |
+
"logits/chosen": -2.6853079795837402,
|
361 |
+
"logits/rejected": -2.600752830505371,
|
362 |
+
"logps/chosen": -264.80615234375,
|
363 |
+
"logps/rejected": -214.556396484375,
|
364 |
+
"loss": 0.3669,
|
365 |
+
"rewards/accuracies": 0.862500011920929,
|
366 |
+
"rewards/chosen": 6.017881393432617,
|
367 |
+
"rewards/margins": 7.167391300201416,
|
368 |
+
"rewards/rejected": -1.1495091915130615,
|
369 |
"step": 230
|
370 |
},
|
371 |
{
|
372 |
"epoch": 0.61,
|
373 |
+
"grad_norm": 831.6224733555287,
|
374 |
"learning_rate": 1.934352493925695e-07,
|
375 |
+
"logits/chosen": -2.667158603668213,
|
376 |
+
"logits/rejected": -2.620408058166504,
|
377 |
+
"logps/chosen": -261.4100036621094,
|
378 |
+
"logps/rejected": -219.61032104492188,
|
379 |
+
"loss": 0.3738,
|
380 |
"rewards/accuracies": 0.893750011920929,
|
381 |
+
"rewards/chosen": 4.424870014190674,
|
382 |
+
"rewards/margins": 8.652185440063477,
|
383 |
+
"rewards/rejected": -4.227315425872803,
|
384 |
"step": 240
|
385 |
},
|
386 |
{
|
387 |
"epoch": 0.64,
|
388 |
+
"grad_norm": 563.6526543418182,
|
389 |
"learning_rate": 1.7189501409486059e-07,
|
390 |
+
"logits/chosen": -2.674989938735962,
|
391 |
+
"logits/rejected": -2.606071949005127,
|
392 |
+
"logps/chosen": -266.0426940917969,
|
393 |
+
"logps/rejected": -221.0687713623047,
|
394 |
+
"loss": 0.367,
|
395 |
+
"rewards/accuracies": 0.84375,
|
396 |
+
"rewards/chosen": 4.015201091766357,
|
397 |
+
"rewards/margins": 7.296302795410156,
|
398 |
+
"rewards/rejected": -3.281102418899536,
|
399 |
"step": 250
|
400 |
},
|
401 |
{
|
402 |
"epoch": 0.67,
|
403 |
+
"grad_norm": 457.0558678943042,
|
404 |
"learning_rate": 1.5098005849021078e-07,
|
405 |
+
"logits/chosen": -2.6647772789001465,
|
406 |
+
"logits/rejected": -2.6081786155700684,
|
407 |
+
"logps/chosen": -260.12384033203125,
|
408 |
+
"logps/rejected": -207.2174835205078,
|
409 |
+
"loss": 0.3387,
|
410 |
+
"rewards/accuracies": 0.875,
|
411 |
+
"rewards/chosen": 4.0153021812438965,
|
412 |
+
"rewards/margins": 6.823422908782959,
|
413 |
+
"rewards/rejected": -2.8081207275390625,
|
414 |
"step": 260
|
415 |
},
|
416 |
{
|
417 |
"epoch": 0.69,
|
418 |
+
"grad_norm": 478.6275517216598,
|
419 |
"learning_rate": 1.30857819994673e-07,
|
420 |
+
"logits/chosen": -2.641366481781006,
|
421 |
+
"logits/rejected": -2.558147668838501,
|
422 |
+
"logps/chosen": -272.80377197265625,
|
423 |
+
"logps/rejected": -228.5403594970703,
|
424 |
+
"loss": 0.4885,
|
425 |
+
"rewards/accuracies": 0.90625,
|
426 |
+
"rewards/chosen": 2.646123170852661,
|
427 |
+
"rewards/margins": 8.172276496887207,
|
428 |
+
"rewards/rejected": -5.526152610778809,
|
429 |
"step": 270
|
430 |
},
|
431 |
{
|
432 |
"epoch": 0.72,
|
433 |
+
"grad_norm": 369.20521183996,
|
434 |
"learning_rate": 1.116893898236716e-07,
|
435 |
+
"logits/chosen": -2.6526267528533936,
|
436 |
+
"logits/rejected": -2.597562789916992,
|
437 |
+
"logps/chosen": -268.1092224121094,
|
438 |
+
"logps/rejected": -217.9169464111328,
|
439 |
+
"loss": 0.3441,
|
440 |
+
"rewards/accuracies": 0.875,
|
441 |
+
"rewards/chosen": 3.9179558753967285,
|
442 |
+
"rewards/margins": 7.3985185623168945,
|
443 |
+
"rewards/rejected": -3.480562686920166,
|
444 |
"step": 280
|
445 |
},
|
446 |
{
|
447 |
"epoch": 0.74,
|
448 |
+
"grad_norm": 498.11211487860106,
|
449 |
"learning_rate": 9.362822335518062e-08,
|
450 |
+
"logits/chosen": -2.6078855991363525,
|
451 |
+
"logits/rejected": -2.5603675842285156,
|
452 |
+
"logps/chosen": -265.8270263671875,
|
453 |
+
"logps/rejected": -215.421142578125,
|
454 |
+
"loss": 0.3216,
|
455 |
+
"rewards/accuracies": 0.8687499761581421,
|
456 |
+
"rewards/chosen": 4.283304691314697,
|
457 |
+
"rewards/margins": 7.295420169830322,
|
458 |
+
"rewards/rejected": -3.012115478515625,
|
459 |
"step": 290
|
460 |
},
|
461 |
{
|
462 |
"epoch": 0.77,
|
463 |
+
"grad_norm": 433.49223304725587,
|
464 |
"learning_rate": 7.681891162260015e-08,
|
465 |
+
"logits/chosen": -2.6258552074432373,
|
466 |
+
"logits/rejected": -2.5720603466033936,
|
467 |
+
"logps/chosen": -272.7791442871094,
|
468 |
+
"logps/rejected": -219.44140625,
|
469 |
+
"loss": 0.3717,
|
470 |
+
"rewards/accuracies": 0.862500011920929,
|
471 |
+
"rewards/chosen": 4.056910514831543,
|
472 |
+
"rewards/margins": 7.150809288024902,
|
473 |
+
"rewards/rejected": -3.093899726867676,
|
474 |
"step": 300
|
475 |
},
|
476 |
{
|
477 |
"epoch": 0.79,
|
478 |
+
"grad_norm": 369.6165987845038,
|
479 |
"learning_rate": 6.139602377230247e-08,
|
480 |
+
"logits/chosen": -2.5929980278015137,
|
481 |
+
"logits/rejected": -2.524878978729248,
|
482 |
+
"logps/chosen": -276.72418212890625,
|
483 |
+
"logps/rejected": -214.73757934570312,
|
484 |
+
"loss": 0.3974,
|
485 |
+
"rewards/accuracies": 0.887499988079071,
|
486 |
+
"rewards/chosen": 3.884862184524536,
|
487 |
+
"rewards/margins": 7.488539695739746,
|
488 |
+
"rewards/rejected": -3.6036782264709473,
|
489 |
"step": 310
|
490 |
},
|
491 |
{
|
492 |
"epoch": 0.82,
|
493 |
+
"grad_norm": 542.2853213894015,
|
494 |
"learning_rate": 4.748302975270837e-08,
|
495 |
+
"logits/chosen": -2.618070125579834,
|
496 |
+
"logits/rejected": -2.5723397731781006,
|
497 |
+
"logps/chosen": -259.38433837890625,
|
498 |
+
"logps/rejected": -202.79501342773438,
|
499 |
+
"loss": 0.3593,
|
500 |
"rewards/accuracies": 0.8812500238418579,
|
501 |
+
"rewards/chosen": 3.869190216064453,
|
502 |
+
"rewards/margins": 6.644041538238525,
|
503 |
+
"rewards/rejected": -2.7748515605926514,
|
504 |
"step": 320
|
505 |
},
|
506 |
{
|
507 |
"epoch": 0.84,
|
508 |
+
"grad_norm": 1179.920803140471,
|
509 |
"learning_rate": 3.5191311859445795e-08,
|
510 |
+
"logits/chosen": -2.639923572540283,
|
511 |
+
"logits/rejected": -2.582639455795288,
|
512 |
+
"logps/chosen": -262.33123779296875,
|
513 |
+
"logps/rejected": -215.505615234375,
|
514 |
+
"loss": 0.3667,
|
515 |
+
"rewards/accuracies": 0.8999999761581421,
|
516 |
+
"rewards/chosen": 5.157271385192871,
|
517 |
+
"rewards/margins": 7.100647926330566,
|
518 |
+
"rewards/rejected": -1.9433765411376953,
|
519 |
"step": 330
|
520 |
},
|
521 |
{
|
522 |
"epoch": 0.87,
|
523 |
+
"grad_norm": 467.7658799716397,
|
524 |
"learning_rate": 2.4619273049795996e-08,
|
525 |
+
"logits/chosen": -2.6265780925750732,
|
526 |
+
"logits/rejected": -2.5702362060546875,
|
527 |
+
"logps/chosen": -257.75689697265625,
|
528 |
+
"logps/rejected": -209.0670623779297,
|
529 |
+
"loss": 0.324,
|
530 |
+
"rewards/accuracies": 0.9125000238418579,
|
531 |
+
"rewards/chosen": 4.93801212310791,
|
532 |
+
"rewards/margins": 7.8753981590271,
|
533 |
+
"rewards/rejected": -2.9373855590820312,
|
534 |
"step": 340
|
535 |
},
|
536 |
{
|
537 |
"epoch": 0.9,
|
538 |
+
"grad_norm": 345.87107109384476,
|
539 |
"learning_rate": 1.5851549164932115e-08,
|
540 |
+
"logits/chosen": -2.637718677520752,
|
541 |
+
"logits/rejected": -2.589808940887451,
|
542 |
+
"logps/chosen": -267.7754821777344,
|
543 |
+
"logps/rejected": -225.0698699951172,
|
544 |
+
"loss": 0.3399,
|
545 |
+
"rewards/accuracies": 0.893750011920929,
|
546 |
+
"rewards/chosen": 4.532751560211182,
|
547 |
+
"rewards/margins": 7.0415191650390625,
|
548 |
+
"rewards/rejected": -2.508768320083618,
|
549 |
"step": 350
|
550 |
},
|
551 |
{
|
552 |
"epoch": 0.92,
|
553 |
+
"grad_norm": 491.63656420568554,
|
554 |
"learning_rate": 8.958331366609423e-09,
|
555 |
+
"logits/chosen": -2.641171932220459,
|
556 |
+
"logits/rejected": -2.574852705001831,
|
557 |
+
"logps/chosen": -272.898193359375,
|
558 |
+
"logps/rejected": -217.81997680664062,
|
559 |
+
"loss": 0.3962,
|
560 |
+
"rewards/accuracies": 0.893750011920929,
|
561 |
+
"rewards/chosen": 4.423092842102051,
|
562 |
+
"rewards/margins": 7.286849021911621,
|
563 |
+
"rewards/rejected": -2.863755464553833,
|
564 |
"step": 360
|
565 |
},
|
566 |
{
|
567 |
"epoch": 0.95,
|
568 |
+
"grad_norm": 863.5117187901297,
|
569 |
"learning_rate": 3.994804212627461e-09,
|
570 |
+
"logits/chosen": -2.600586175918579,
|
571 |
+
"logits/rejected": -2.565150737762451,
|
572 |
+
"logps/chosen": -271.3792419433594,
|
573 |
+
"logps/rejected": -227.0088348388672,
|
574 |
+
"loss": 0.4043,
|
575 |
+
"rewards/accuracies": 0.862500011920929,
|
576 |
+
"rewards/chosen": 5.524130821228027,
|
577 |
+
"rewards/margins": 7.229277610778809,
|
578 |
+
"rewards/rejected": -1.7051467895507812,
|
579 |
"step": 370
|
580 |
},
|
581 |
{
|
582 |
"epoch": 0.97,
|
583 |
+
"grad_norm": 386.7539067890418,
|
584 |
"learning_rate": 1.0007038696262516e-09,
|
585 |
+
"logits/chosen": -2.649324893951416,
|
586 |
+
"logits/rejected": -2.608961820602417,
|
587 |
+
"logps/chosen": -261.44415283203125,
|
588 |
+
"logps/rejected": -229.10617065429688,
|
589 |
+
"loss": 0.3514,
|
590 |
+
"rewards/accuracies": 0.862500011920929,
|
591 |
+
"rewards/chosen": 4.858848571777344,
|
592 |
+
"rewards/margins": 7.17071008682251,
|
593 |
+
"rewards/rejected": -2.311861276626587,
|
594 |
"step": 380
|
595 |
},
|
596 |
{
|
597 |
"epoch": 1.0,
|
598 |
+
"grad_norm": 658.7609084129675,
|
599 |
"learning_rate": 0.0,
|
600 |
+
"logits/chosen": -2.657970428466797,
|
601 |
+
"logits/rejected": -2.5974597930908203,
|
602 |
+
"logps/chosen": -249.08193969726562,
|
603 |
+
"logps/rejected": -208.86935424804688,
|
604 |
+
"loss": 0.389,
|
605 |
+
"rewards/accuracies": 0.8687499761581421,
|
606 |
+
"rewards/chosen": 4.138771057128906,
|
607 |
+
"rewards/margins": 6.9808807373046875,
|
608 |
+
"rewards/rejected": -2.8421101570129395,
|
609 |
"step": 390
|
610 |
},
|
611 |
{
|
612 |
"epoch": 1.0,
|
613 |
"step": 390,
|
614 |
"total_flos": 0.0,
|
615 |
+
"train_loss": 0.39275855483152927,
|
616 |
+
"train_runtime": 5850.2913,
|
617 |
+
"train_samples_per_second": 8.547,
|
618 |
"train_steps_per_second": 0.067
|
619 |
}
|
620 |
],
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6264
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:69755ef6c720cdfd7d14b231fbf0ad5a8fc29c369a6e770d1d3031286dc30b37
|
3 |
size 6264
|