File size: 7,896 Bytes
68ffcc1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 121,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 44.33979039021003,
      "learning_rate": 3.846153846153846e-08,
      "logits/chosen": -3.751237154006958,
      "logits/rejected": -3.652125358581543,
      "logps/chosen": -995.5263671875,
      "logps/rejected": -1318.9669189453125,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.08,
      "grad_norm": 35.835351014179686,
      "learning_rate": 3.8461538461538463e-07,
      "logits/chosen": -3.71555233001709,
      "logits/rejected": -3.6416714191436768,
      "logps/chosen": -873.4622192382812,
      "logps/rejected": -1458.0814208984375,
      "loss": 0.6871,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.000241430607275106,
      "rewards/margins": 0.014367452822625637,
      "rewards/rejected": -0.014608883298933506,
      "step": 10
    },
    {
      "epoch": 0.17,
      "grad_norm": 29.834055642040035,
      "learning_rate": 4.948351554413879e-07,
      "logits/chosen": -3.8361048698425293,
      "logits/rejected": -3.7403998374938965,
      "logps/chosen": -946.3114013671875,
      "logps/rejected": -1385.838134765625,
      "loss": 0.5664,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -0.045135729014873505,
      "rewards/margins": 0.27529585361480713,
      "rewards/rejected": -0.32043159008026123,
      "step": 20
    },
    {
      "epoch": 0.25,
      "grad_norm": 19.851574913053625,
      "learning_rate": 4.700503477950277e-07,
      "logits/chosen": -4.062998294830322,
      "logits/rejected": -4.014255046844482,
      "logps/chosen": -949.5398559570312,
      "logps/rejected": -1611.4945068359375,
      "loss": 0.3144,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.3165217936038971,
      "rewards/margins": 1.7194368839263916,
      "rewards/rejected": -2.035958766937256,
      "step": 30
    },
    {
      "epoch": 0.33,
      "grad_norm": 20.09319706827758,
      "learning_rate": 4.2677669529663686e-07,
      "logits/chosen": -4.247502326965332,
      "logits/rejected": -4.206511974334717,
      "logps/chosen": -1011.1369018554688,
      "logps/rejected": -1838.808349609375,
      "loss": 0.2629,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -0.8711185455322266,
      "rewards/margins": 3.5932796001434326,
      "rewards/rejected": -4.464398384094238,
      "step": 40
    },
    {
      "epoch": 0.41,
      "grad_norm": 17.14618338383071,
      "learning_rate": 3.686500924369101e-07,
      "logits/chosen": -4.1785712242126465,
      "logits/rejected": -4.129928112030029,
      "logps/chosen": -1003.81005859375,
      "logps/rejected": -1773.5804443359375,
      "loss": 0.2915,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.6944042444229126,
      "rewards/margins": 3.5044853687286377,
      "rewards/rejected": -4.19888973236084,
      "step": 50
    },
    {
      "epoch": 0.5,
      "grad_norm": 22.001296504737557,
      "learning_rate": 3.005543930830095e-07,
      "logits/chosen": -4.102808475494385,
      "logits/rejected": -4.0734052658081055,
      "logps/chosen": -1009.1546020507812,
      "logps/rejected": -1869.724365234375,
      "loss": 0.1679,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.5237128138542175,
      "rewards/margins": 4.181941509246826,
      "rewards/rejected": -4.705655097961426,
      "step": 60
    },
    {
      "epoch": 0.58,
      "grad_norm": 22.911683114341198,
      "learning_rate": 2.2821106431308543e-07,
      "logits/chosen": -4.085400104522705,
      "logits/rejected": -4.0569257736206055,
      "logps/chosen": -1070.66650390625,
      "logps/rejected": -1899.2978515625,
      "loss": 0.1352,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.6565460562705994,
      "rewards/margins": 4.237053871154785,
      "rewards/rejected": -4.893599987030029,
      "step": 70
    },
    {
      "epoch": 0.66,
      "grad_norm": 13.574682459891369,
      "learning_rate": 1.5769846317182892e-07,
      "logits/chosen": -4.115840435028076,
      "logits/rejected": -4.082175254821777,
      "logps/chosen": -976.5784912109375,
      "logps/rejected": -2003.0341796875,
      "loss": 0.119,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.6749869585037231,
      "rewards/margins": 5.019448757171631,
      "rewards/rejected": -5.694436073303223,
      "step": 80
    },
    {
      "epoch": 0.74,
      "grad_norm": 17.137536022407392,
      "learning_rate": 9.494112718293502e-08,
      "logits/chosen": -4.116685390472412,
      "logits/rejected": -4.105366230010986,
      "logps/chosen": -969.3150634765625,
      "logps/rejected": -2016.6741943359375,
      "loss": 0.105,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.6770291328430176,
      "rewards/margins": 5.554581642150879,
      "rewards/rejected": -6.231610298156738,
      "step": 90
    },
    {
      "epoch": 0.83,
      "grad_norm": 19.248054573741623,
      "learning_rate": 4.521198892775202e-08,
      "logits/chosen": -4.130964279174805,
      "logits/rejected": -4.107068061828613,
      "logps/chosen": -997.1163940429688,
      "logps/rejected": -2062.62109375,
      "loss": 0.1232,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.8503810167312622,
      "rewards/margins": 5.699716091156006,
      "rewards/rejected": -6.550097465515137,
      "step": 100
    },
    {
      "epoch": 0.83,
      "eval_logits/chosen": -4.124736309051514,
      "eval_logits/rejected": -4.084261894226074,
      "eval_logps/chosen": -456.256591796875,
      "eval_logps/rejected": -691.86181640625,
      "eval_loss": 0.46714693307876587,
      "eval_rewards/accuracies": 0.75,
      "eval_rewards/chosen": -0.502022922039032,
      "eval_rewards/margins": 0.8244528770446777,
      "eval_rewards/rejected": -1.326475739479065,
      "eval_runtime": 14.3746,
      "eval_samples_per_second": 5.287,
      "eval_steps_per_second": 0.209,
      "step": 100
    },
    {
      "epoch": 0.91,
      "grad_norm": 12.936607152276345,
      "learning_rate": 1.2689339106741526e-08,
      "logits/chosen": -4.11704158782959,
      "logits/rejected": -4.1085028648376465,
      "logps/chosen": -983.2184448242188,
      "logps/rejected": -2123.88623046875,
      "loss": 0.1405,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.7897412776947021,
      "rewards/margins": 5.98792028427124,
      "rewards/rejected": -6.7776618003845215,
      "step": 110
    },
    {
      "epoch": 0.99,
      "grad_norm": 18.334577187135583,
      "learning_rate": 1.0576247944985018e-10,
      "logits/chosen": -4.1415205001831055,
      "logits/rejected": -4.129425048828125,
      "logps/chosen": -978.7972412109375,
      "logps/rejected": -2053.748046875,
      "loss": 0.1121,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.8511762619018555,
      "rewards/margins": 5.528683185577393,
      "rewards/rejected": -6.37985897064209,
      "step": 120
    },
    {
      "epoch": 1.0,
      "step": 121,
      "total_flos": 0.0,
      "train_loss": 0.021612109477854958,
      "train_runtime": 343.2656,
      "train_samples_per_second": 22.536,
      "train_steps_per_second": 0.352
    }
  ],
  "logging_steps": 10,
  "max_steps": 121,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}