qwen2.5-0.5b-expo-DPO-ES-10 / all_results.json
hZzy's picture
End of training
ab5fbfb verified
{
"epoch": 2.2673594709494567,
"eval_dpo_loss": 20.566144943237305,
"eval_logits": -0.6148493885993958,
"eval_logps": -80.20514678955078,
"eval_loss": 21.350601196289062,
"eval_objective": 20.566144943237305,
"eval_ranking_idealized": 0.5212215185165405,
"eval_ranking_idealized_expo": 0.5212215185165405,
"eval_ranking_simple": 0.5383023023605347,
"eval_regularize": 20.566144943237305,
"eval_runtime": 314.2807,
"eval_samples": 5790,
"eval_samples_per_second": 18.423,
"eval_steps_per_second": 1.537,
"eval_wo_beta": 6.651296138763428,
"total_flos": 0.0,
"train_loss": 7.659533626437187,
"train_runtime": 21973.9049,
"train_samples": 50802,
"train_samples_per_second": 11.56,
"train_steps_per_second": 0.08
}