File size: 10,932 Bytes
c336a3b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9921671018276762,
  "eval_steps": 500,
  "global_step": 95,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 6.531942491552821,
      "learning_rate": 5e-08,
      "logits/chosen": -2.851747512817383,
      "logits/rejected": -2.833996534347534,
      "logps/chosen": -165.70089721679688,
      "logps/rejected": -198.857666015625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.05,
      "grad_norm": 5.930803989300868,
      "learning_rate": 2.5e-07,
      "logits/chosen": -2.770416259765625,
      "logits/rejected": -2.7731680870056152,
      "logps/chosen": -171.3281707763672,
      "logps/rejected": -172.58348083496094,
      "loss": 0.693,
      "rewards/accuracies": 0.375,
      "rewards/chosen": 0.00034834028338082135,
      "rewards/margins": 4.263037408236414e-05,
      "rewards/rejected": 0.0003057100111618638,
      "step": 5
    },
    {
      "epoch": 0.1,
      "grad_norm": 7.205939520530408,
      "learning_rate": 5e-07,
      "logits/chosen": -2.785672664642334,
      "logits/rejected": -2.7945070266723633,
      "logps/chosen": -189.79400634765625,
      "logps/rejected": -194.38011169433594,
      "loss": 0.6923,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.005779535509645939,
      "rewards/margins": 0.0015561816981062293,
      "rewards/rejected": 0.004223353695124388,
      "step": 10
    },
    {
      "epoch": 0.16,
      "grad_norm": 7.119689881451758,
      "learning_rate": 4.957432749209755e-07,
      "logits/chosen": -2.841862678527832,
      "logits/rejected": -2.8522396087646484,
      "logps/chosen": -196.4453582763672,
      "logps/rejected": -186.3593292236328,
      "loss": 0.6891,
      "rewards/accuracies": 0.581250011920929,
      "rewards/chosen": 0.0351785309612751,
      "rewards/margins": 0.009719189256429672,
      "rewards/rejected": 0.025459343567490578,
      "step": 15
    },
    {
      "epoch": 0.21,
      "grad_norm": 7.025742204681022,
      "learning_rate": 4.83118057351089e-07,
      "logits/chosen": -2.8577423095703125,
      "logits/rejected": -2.8679168224334717,
      "logps/chosen": -163.30587768554688,
      "logps/rejected": -176.16122436523438,
      "loss": 0.6822,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.054369617253541946,
      "rewards/margins": 0.01927168108522892,
      "rewards/rejected": 0.035097938030958176,
      "step": 20
    },
    {
      "epoch": 0.26,
      "grad_norm": 7.047833772227819,
      "learning_rate": 4.6255428393240354e-07,
      "logits/chosen": -2.8176944255828857,
      "logits/rejected": -2.8154852390289307,
      "logps/chosen": -127.52900695800781,
      "logps/rejected": -149.99598693847656,
      "loss": 0.6734,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.039430197328329086,
      "rewards/margins": 0.044888969510793686,
      "rewards/rejected": -0.005458767991513014,
      "step": 25
    },
    {
      "epoch": 0.31,
      "grad_norm": 7.4255717276037405,
      "learning_rate": 4.3475222930516473e-07,
      "logits/chosen": -2.781858444213867,
      "logits/rejected": -2.7814831733703613,
      "logps/chosen": -161.177734375,
      "logps/rejected": -173.82421875,
      "loss": 0.665,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.01698228344321251,
      "rewards/margins": 0.0684308260679245,
      "rewards/rejected": -0.08541311323642731,
      "step": 30
    },
    {
      "epoch": 0.37,
      "grad_norm": 7.891881929971765,
      "learning_rate": 4.006586590948141e-07,
      "logits/chosen": -2.848252296447754,
      "logits/rejected": -2.8431050777435303,
      "logps/chosen": -192.15963745117188,
      "logps/rejected": -209.07540893554688,
      "loss": 0.6544,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.008883295580744743,
      "rewards/margins": 0.0794510543346405,
      "rewards/rejected": -0.0883343443274498,
      "step": 35
    },
    {
      "epoch": 0.42,
      "grad_norm": 9.209834953181781,
      "learning_rate": 3.614345889441346e-07,
      "logits/chosen": -2.7681477069854736,
      "logits/rejected": -2.78022837638855,
      "logps/chosen": -135.9792022705078,
      "logps/rejected": -164.3667449951172,
      "loss": 0.6465,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.031201759353280067,
      "rewards/margins": 0.13447019457817078,
      "rewards/rejected": -0.16567197442054749,
      "step": 40
    },
    {
      "epoch": 0.47,
      "grad_norm": 13.835886251568184,
      "learning_rate": 3.184157475180207e-07,
      "logits/chosen": -2.7284975051879883,
      "logits/rejected": -2.7436182498931885,
      "logps/chosen": -205.8560028076172,
      "logps/rejected": -212.56710815429688,
      "loss": 0.6496,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.17208269238471985,
      "rewards/margins": 0.13685402274131775,
      "rewards/rejected": -0.30893674492836,
      "step": 45
    },
    {
      "epoch": 0.52,
      "grad_norm": 10.309186722273289,
      "learning_rate": 2.730670898658255e-07,
      "logits/chosen": -2.7203848361968994,
      "logits/rejected": -2.7220139503479004,
      "logps/chosen": -183.94479370117188,
      "logps/rejected": -218.1922149658203,
      "loss": 0.6248,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.09618374705314636,
      "rewards/margins": 0.18668127059936523,
      "rewards/rejected": -0.2828650176525116,
      "step": 50
    },
    {
      "epoch": 0.57,
      "grad_norm": 12.04836501966109,
      "learning_rate": 2.2693291013417452e-07,
      "logits/chosen": -2.6110920906066895,
      "logits/rejected": -2.6190452575683594,
      "logps/chosen": -151.388916015625,
      "logps/rejected": -174.0006561279297,
      "loss": 0.6317,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.032760851085186005,
      "rewards/margins": 0.17945989966392517,
      "rewards/rejected": -0.2122207134962082,
      "step": 55
    },
    {
      "epoch": 0.63,
      "grad_norm": 14.818475765214615,
      "learning_rate": 1.8158425248197928e-07,
      "logits/chosen": -2.69221568107605,
      "logits/rejected": -2.689034938812256,
      "logps/chosen": -181.30128479003906,
      "logps/rejected": -231.5193634033203,
      "loss": 0.6122,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.09343220293521881,
      "rewards/margins": 0.2633873522281647,
      "rewards/rejected": -0.3568195104598999,
      "step": 60
    },
    {
      "epoch": 0.68,
      "grad_norm": 15.265728023102268,
      "learning_rate": 1.3856541105586545e-07,
      "logits/chosen": -2.7168681621551514,
      "logits/rejected": -2.7309060096740723,
      "logps/chosen": -185.16700744628906,
      "logps/rejected": -220.42764282226562,
      "loss": 0.6045,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.23089858889579773,
      "rewards/margins": 0.27521029114723206,
      "rewards/rejected": -0.5061088800430298,
      "step": 65
    },
    {
      "epoch": 0.73,
      "grad_norm": 13.52741638941588,
      "learning_rate": 9.934134090518592e-08,
      "logits/chosen": -2.6834919452667236,
      "logits/rejected": -2.6923632621765137,
      "logps/chosen": -200.9665069580078,
      "logps/rejected": -217.9497528076172,
      "loss": 0.6094,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.23829719424247742,
      "rewards/margins": 0.2035256326198578,
      "rewards/rejected": -0.4418228268623352,
      "step": 70
    },
    {
      "epoch": 0.78,
      "grad_norm": 14.871873879280589,
      "learning_rate": 6.524777069483525e-08,
      "logits/chosen": -2.6725871562957764,
      "logits/rejected": -2.6699538230895996,
      "logps/chosen": -185.2981719970703,
      "logps/rejected": -229.42092895507812,
      "loss": 0.5985,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.13438589870929718,
      "rewards/margins": 0.32371044158935547,
      "rewards/rejected": -0.45809632539749146,
      "step": 75
    },
    {
      "epoch": 0.84,
      "grad_norm": 12.313204564006284,
      "learning_rate": 3.74457160675965e-08,
      "logits/chosen": -2.6488523483276367,
      "logits/rejected": -2.6512537002563477,
      "logps/chosen": -177.8891143798828,
      "logps/rejected": -211.4371795654297,
      "loss": 0.6019,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.09899892657995224,
      "rewards/margins": 0.26206719875335693,
      "rewards/rejected": -0.36106616258621216,
      "step": 80
    },
    {
      "epoch": 0.89,
      "grad_norm": 17.242389025181602,
      "learning_rate": 1.6881942648911074e-08,
      "logits/chosen": -2.6852784156799316,
      "logits/rejected": -2.6899216175079346,
      "logps/chosen": -171.39414978027344,
      "logps/rejected": -207.66738891601562,
      "loss": 0.6214,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.11210503429174423,
      "rewards/margins": 0.2755950093269348,
      "rewards/rejected": -0.38770005106925964,
      "step": 85
    },
    {
      "epoch": 0.94,
      "grad_norm": 13.932688124952723,
      "learning_rate": 4.256725079024553e-09,
      "logits/chosen": -2.6324477195739746,
      "logits/rejected": -2.6469483375549316,
      "logps/chosen": -181.08218383789062,
      "logps/rejected": -215.79953002929688,
      "loss": 0.6081,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.12142710387706757,
      "rewards/margins": 0.2807455062866211,
      "rewards/rejected": -0.40217262506484985,
      "step": 90
    },
    {
      "epoch": 0.99,
      "grad_norm": 15.247505163019246,
      "learning_rate": 0.0,
      "logits/chosen": -2.682211399078369,
      "logits/rejected": -2.697298765182495,
      "logps/chosen": -198.21182250976562,
      "logps/rejected": -223.2611541748047,
      "loss": 0.6054,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.12603162229061127,
      "rewards/margins": 0.2787989377975464,
      "rewards/rejected": -0.40483060479164124,
      "step": 95
    },
    {
      "epoch": 0.99,
      "step": 95,
      "total_flos": 0.0,
      "train_loss": 0.6401761331056294,
      "train_runtime": 2555.4095,
      "train_samples_per_second": 4.785,
      "train_steps_per_second": 0.037
    }
  ],
  "logging_steps": 5,
  "max_steps": 95,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}