27 lines
1018 B
JSON
27 lines
1018 B
JSON
|
{
|
||
|
|
"epoch": 0.9989528795811519,
|
||
|
|
"eval_KL/chosen_KL_mean": -41.074798583984375,
|
||
|
|
"eval_KL/mean": -67.0528793334961,
|
||
|
|
"eval_KL/rejected_KL_mean": -93.03097534179688,
|
||
|
|
"eval_KL/std": 70.80274963378906,
|
||
|
|
"eval_fcm_dpo/beta": 0.005364115349948406,
|
||
|
|
"eval_logits/chosen": 1.4891531467437744,
|
||
|
|
"eval_logits/rejected": 1.5578936338424683,
|
||
|
|
"eval_logps/chosen": -321.4914855957031,
|
||
|
|
"eval_logps/ref_chosen": -280.41668701171875,
|
||
|
|
"eval_logps/ref_rejected": -264.6455383300781,
|
||
|
|
"eval_logps/rejected": -357.676513671875,
|
||
|
|
"eval_loss": 0.5913359522819519,
|
||
|
|
"eval_margin_dpo/margin_mean": 51.956180572509766,
|
||
|
|
"eval_margin_dpo/margin_std": 81.990966796875,
|
||
|
|
"eval_runtime": 88.3876,
|
||
|
|
"eval_samples": 2000,
|
||
|
|
"eval_samples_per_second": 22.628,
|
||
|
|
"eval_steps_per_second": 2.828,
|
||
|
|
"total_flos": 0.0,
|
||
|
|
"train_loss": 2.690188397901363,
|
||
|
|
"train_runtime": 4372.8359,
|
||
|
|
"train_samples": 61135,
|
||
|
|
"train_samples_per_second": 13.981,
|
||
|
|
"train_steps_per_second": 0.109
|
||
|
|
}
|