zephyr-7b-dpo-full / eval_results.json
sfulay's picture
End of training
17ce58f verified
raw
history blame contribute delete
509 Bytes
{
"epoch": 1.0,
"eval_logits/chosen": -2.6333940029144287,
"eval_logits/rejected": -2.5763978958129883,
"eval_logps/chosen": -280.7647705078125,
"eval_logps/rejected": -159.44223022460938,
"eval_loss": 0.6931471228599548,
"eval_rewards/accuracies": 0.0,
"eval_rewards/chosen": 0.0,
"eval_rewards/margins": 0.0,
"eval_rewards/rejected": 0.0,
"eval_runtime": 5.9655,
"eval_samples": 100,
"eval_samples_per_second": 16.763,
"eval_steps_per_second": 0.335
}