File size: 10,457 Bytes
7df5122
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 95.85,
      "completions/clipped_ratio": 0.125,
      "completions/max_length": 256.0,
      "completions/max_terminated_length": 195.6,
      "completions/mean_length": 95.85,
      "completions/mean_terminated_length": 72.8880989074707,
      "completions/min_length": 36.6,
      "completions/min_terminated_length": 36.6,
      "epoch": 0.5,
      "frac_reward_zero_std": 0.2,
      "grad_norm": 0.3041034936904907,
      "kl": 0.00021898818758927518,
      "learning_rate": 4e-05,
      "loss": 0.0329527884721756,
      "num_tokens": 21444.0,
      "reward": 0.7767857074737549,
      "reward_std": 0.7150911569595337,
      "rewards/reward_fn/mean": 0.7767857074737549,
      "rewards/reward_fn/std": 0.9959433197975158,
      "step": 5
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 95.1,
      "completions/clipped_ratio": 0.0875,
      "completions/max_length": 225.2,
      "completions/max_terminated_length": 159.2,
      "completions/mean_length": 95.1,
      "completions/mean_terminated_length": 80.34466400146485,
      "completions/min_length": 34.6,
      "completions/min_terminated_length": 34.6,
      "epoch": 1.0,
      "frac_reward_zero_std": 0.4,
      "grad_norm": 0.2708108425140381,
      "kl": 0.006061523332027719,
      "learning_rate": 9e-05,
      "loss": -0.016053691506385803,
      "num_tokens": 42220.0,
      "reward": 0.6875000119209289,
      "reward_std": 0.5175672590732574,
      "rewards/reward_fn/mean": 0.6875,
      "rewards/reward_fn/std": 0.8034753799438477,
      "step": 10
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 146.275,
      "completions/clipped_ratio": 0.325,
      "completions/max_length": 256.0,
      "completions/max_terminated_length": 225.4,
      "completions/mean_length": 146.275,
      "completions/mean_terminated_length": 92.8663101196289,
      "completions/min_length": 32.6,
      "completions/min_terminated_length": 32.6,
      "epoch": 1.5,
      "frac_reward_zero_std": 0.2,
      "grad_norm": 0.27529817819595337,
      "kl": 0.01560373855754733,
      "learning_rate": 9.989068136093873e-05,
      "loss": 0.010054035484790802,
      "num_tokens": 69198.0,
      "reward": 1.1558928728103637,
      "reward_std": 0.8465379416942597,
      "rewards/reward_fn/mean": 1.1558928728103637,
      "rewards/reward_fn/std": 1.1404805421829223,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 75.85,
      "completions/clipped_ratio": 0.05,
      "completions/max_length": 220.6,
      "completions/max_terminated_length": 151.0,
      "completions/mean_length": 75.85,
      "completions/mean_terminated_length": 66.14857330322266,
      "completions/min_length": 32.8,
      "completions/min_terminated_length": 32.8,
      "epoch": 2.0,
      "frac_reward_zero_std": 0.35,
      "grad_norm": 0.353487491607666,
      "kl": 0.027986684162169696,
      "learning_rate": 9.944739353007344e-05,
      "loss": -0.030885905027389526,
      "num_tokens": 87334.0,
      "reward": 0.7932142853736878,
      "reward_std": 0.4118983089923859,
      "rewards/reward_fn/mean": 0.7932142853736878,
      "rewards/reward_fn/std": 0.7167340397834778,
      "step": 20
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 90.85,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 248.0,
      "completions/max_terminated_length": 216.4,
      "completions/mean_length": 90.85,
      "completions/mean_terminated_length": 79.819287109375,
      "completions/min_length": 28.0,
      "completions/min_terminated_length": 28.0,
      "epoch": 2.5,
      "frac_reward_zero_std": 0.15,
      "grad_norm": 0.3318491578102112,
      "kl": 0.04692814685404301,
      "learning_rate": 9.86663298624003e-05,
      "loss": 0.05783940553665161,
      "num_tokens": 108330.0,
      "reward": 1.1987500190734863,
      "reward_std": 0.5907092690467834,
      "rewards/reward_fn/mean": 1.1987500190734863,
      "rewards/reward_fn/std": 0.9482298731803894,
      "step": 25
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 91.775,
      "completions/clipped_ratio": 0.125,
      "completions/max_length": 248.2,
      "completions/max_terminated_length": 212.6,
      "completions/mean_length": 91.775,
      "completions/mean_terminated_length": 68.68722763061524,
      "completions/min_length": 25.2,
      "completions/min_terminated_length": 25.2,
      "epoch": 3.0,
      "frac_reward_zero_std": 0.4,
      "grad_norm": 0.24605490267276764,
      "kl": 0.07254143096506596,
      "learning_rate": 9.755282581475769e-05,
      "loss": 0.03934891819953919,
      "num_tokens": 129012.0,
      "reward": 1.0242857217788697,
      "reward_std": 0.5802445828914642,
      "rewards/reward_fn/mean": 1.0242857217788697,
      "rewards/reward_fn/std": 0.8706011414527893,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 59.7875,
      "completions/clipped_ratio": 0.0375,
      "completions/max_length": 175.6,
      "completions/max_terminated_length": 118.4,
      "completions/mean_length": 59.7875,
      "completions/mean_terminated_length": 52.19559555053711,
      "completions/min_length": 25.4,
      "completions/min_terminated_length": 25.4,
      "epoch": 3.5,
      "frac_reward_zero_std": 0.6,
      "grad_norm": 0.2970605790615082,
      "kl": 0.053099883161485194,
      "learning_rate": 9.611448774886924e-05,
      "loss": 0.06977529525756836,
      "num_tokens": 147103.0,
      "reward": 0.9473214387893677,
      "reward_std": 0.3923970699310303,
      "rewards/reward_fn/mean": 0.9473214268684387,
      "rewards/reward_fn/std": 0.649278998374939,
      "step": 35
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 47.85,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 94.8,
      "completions/max_terminated_length": 94.8,
      "completions/mean_length": 47.85,
      "completions/mean_terminated_length": 47.85,
      "completions/min_length": 31.2,
      "completions/min_terminated_length": 31.2,
      "epoch": 4.0,
      "frac_reward_zero_std": 0.85,
      "grad_norm": 0.16646049916744232,
      "kl": 0.056573960930109024,
      "learning_rate": 9.43611409721806e-05,
      "loss": -0.0031644195318222047,
      "num_tokens": 164643.0,
      "reward": 0.7899999976158142,
      "reward_std": 0.12681145668029786,
      "rewards/reward_fn/mean": 0.7899999976158142,
      "rewards/reward_fn/std": 0.5133758962154389,
      "step": 40
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 49.2625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 93.6,
      "completions/max_terminated_length": 93.6,
      "completions/mean_length": 49.2625,
      "completions/mean_terminated_length": 49.2625,
      "completions/min_length": 33.8,
      "completions/min_terminated_length": 33.8,
      "epoch": 4.5,
      "frac_reward_zero_std": 0.6,
      "grad_norm": 0.3159082531929016,
      "kl": 0.0353464649990201,
      "learning_rate": 9.230476262104677e-05,
      "loss": -0.03417131304740906,
      "num_tokens": 182568.0,
      "reward": 0.7924999833106995,
      "reward_std": 0.3821385264396667,
      "rewards/reward_fn/mean": 0.7924999833106995,
      "rewards/reward_fn/std": 0.62625293135643,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completion_length": 58.25,
      "completions/clipped_ratio": 0.0125,
      "completions/max_length": 180.6,
      "completions/max_terminated_length": 153.2,
      "completions/mean_length": 58.25,
      "completions/mean_terminated_length": 55.67250061035156,
      "completions/min_length": 33.2,
      "completions/min_terminated_length": 33.2,
      "epoch": 5.0,
      "frac_reward_zero_std": 0.55,
      "grad_norm": 0.27360936999320984,
      "kl": 0.04483541352674365,
      "learning_rate": 8.995939984474624e-05,
      "loss": 0.026852786540985107,
      "num_tokens": 200660.0,
      "reward": 0.774821424484253,
      "reward_std": 0.32772802114486693,
      "rewards/reward_fn/mean": 0.774821424484253,
      "rewards/reward_fn/std": 0.5432307064533234,
      "step": 50
    }
  ],
  "logging_steps": 5,
  "max_steps": 200,
  "num_input_tokens_seen": 200660,
  "num_train_epochs": 20,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}