qwen2.5-0.5b-expo-DPO-noES5-1 / all_results.json
hZzy's picture
End of training
e2cf1b0 verified
raw
history blame contribute delete
606 Bytes
{
"epoch": 1.995276334435522,
"eval_dpo_loss": 1.8786616325378418,
"eval_logits": -0.8508970141410828,
"eval_logps": -81.16841125488281,
"eval_loss": 1.9399291276931763,
"eval_objective": 1.8786616325378418,
"eval_ranking_simple": 0.534679114818573,
"eval_runtime": 616.6991,
"eval_samples": 5790,
"eval_samples_per_second": 9.389,
"eval_steps_per_second": 0.783,
"total_flos": 0.0,
"train_loss": 0.7561936149881645,
"train_runtime": 36938.8609,
"train_samples": 50802,
"train_samples_per_second": 2.751,
"train_steps_per_second": 0.019
}