File size: 5,749 Bytes
3af00bc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.994535519125683,
  "eval_steps": 500,
  "global_step": 91,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 19.759845167026548,
      "learning_rate": 5e-08,
      "logits/chosen": -2.483218193054199,
      "logits/rejected": -2.51887583732605,
      "logps/chosen": -173.969482421875,
      "logps/rejected": -466.1477966308594,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.11,
      "grad_norm": 20.030240595618945,
      "learning_rate": 5e-07,
      "logits/chosen": -2.575467586517334,
      "logits/rejected": -2.533165454864502,
      "logps/chosen": -198.4555206298828,
      "logps/rejected": -388.2120361328125,
      "loss": 0.6923,
      "rewards/accuracies": 0.5416666865348816,
      "rewards/chosen": 0.0029368316754698753,
      "rewards/margins": 0.0030300142243504524,
      "rewards/rejected": -9.318282536696643e-05,
      "step": 10
    },
    {
      "epoch": 0.22,
      "grad_norm": 18.975065752388222,
      "learning_rate": 4.814309923172227e-07,
      "logits/chosen": -2.5513834953308105,
      "logits/rejected": -2.62607479095459,
      "logps/chosen": -228.58560180664062,
      "logps/rejected": -401.50775146484375,
      "loss": 0.6717,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.040665917098522186,
      "rewards/margins": 0.04159024730324745,
      "rewards/rejected": -0.0009243293898180127,
      "step": 20
    },
    {
      "epoch": 0.33,
      "grad_norm": 20.69855762605237,
      "learning_rate": 4.284824336394748e-07,
      "logits/chosen": -2.7989377975463867,
      "logits/rejected": -2.578190326690674,
      "logps/chosen": -222.75927734375,
      "logps/rejected": -369.4903869628906,
      "loss": 0.6245,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 0.1455826759338379,
      "rewards/margins": 0.1740579754114151,
      "rewards/rejected": -0.028475318104028702,
      "step": 30
    },
    {
      "epoch": 0.44,
      "grad_norm": 22.28073899245071,
      "learning_rate": 3.490199415097892e-07,
      "logits/chosen": -3.0074660778045654,
      "logits/rejected": -3.020289421081543,
      "logps/chosen": -209.62252807617188,
      "logps/rejected": -435.1111755371094,
      "loss": 0.5559,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 0.13761594891548157,
      "rewards/margins": 0.311123788356781,
      "rewards/rejected": -0.17350785434246063,
      "step": 40
    },
    {
      "epoch": 0.55,
      "grad_norm": 22.260338017683775,
      "learning_rate": 2.548478329429561e-07,
      "logits/chosen": -3.5367908477783203,
      "logits/rejected": -3.447880983352661,
      "logps/chosen": -207.8536834716797,
      "logps/rejected": -427.22650146484375,
      "loss": 0.4648,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 0.021256055682897568,
      "rewards/margins": 0.612757682800293,
      "rewards/rejected": -0.5915015935897827,
      "step": 50
    },
    {
      "epoch": 0.66,
      "grad_norm": 26.04547260085489,
      "learning_rate": 1.5995556879882243e-07,
      "logits/chosen": -3.8727428913116455,
      "logits/rejected": -3.773128032684326,
      "logps/chosen": -249.71963500976562,
      "logps/rejected": -478.6151428222656,
      "loss": 0.411,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.19216546416282654,
      "rewards/margins": 0.928192138671875,
      "rewards/rejected": -1.1203575134277344,
      "step": 60
    },
    {
      "epoch": 0.77,
      "grad_norm": 18.92789031237145,
      "learning_rate": 7.843959053281663e-08,
      "logits/chosen": -3.900287628173828,
      "logits/rejected": -3.9814674854278564,
      "logps/chosen": -228.2233123779297,
      "logps/rejected": -514.8861694335938,
      "loss": 0.3719,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -0.07359044253826141,
      "rewards/margins": 1.2602514028549194,
      "rewards/rejected": -1.3338419198989868,
      "step": 70
    },
    {
      "epoch": 0.87,
      "grad_norm": 19.01382974055132,
      "learning_rate": 2.2409264758463358e-08,
      "logits/chosen": -3.850555419921875,
      "logits/rejected": -3.8778367042541504,
      "logps/chosen": -222.34033203125,
      "logps/rejected": -520.5603637695312,
      "loss": 0.3401,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -0.008037117309868336,
      "rewards/margins": 1.370126724243164,
      "rewards/rejected": -1.3781636953353882,
      "step": 80
    },
    {
      "epoch": 0.98,
      "grad_norm": 21.287217551030523,
      "learning_rate": 1.8801187394248964e-10,
      "logits/chosen": -3.9702095985412598,
      "logits/rejected": -3.810866117477417,
      "logps/chosen": -205.1525421142578,
      "logps/rejected": -497.5603942871094,
      "loss": 0.3408,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.03864166885614395,
      "rewards/margins": 1.3442068099975586,
      "rewards/rejected": -1.3055652379989624,
      "step": 90
    },
    {
      "epoch": 0.99,
      "step": 91,
      "total_flos": 0.0,
      "train_loss": 0.49557680382833375,
      "train_runtime": 1348.5444,
      "train_samples_per_second": 4.321,
      "train_steps_per_second": 0.067
    }
  ],
  "logging_steps": 10,
  "max_steps": 91,
  "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
}