qwen2.5-0.5b-expo-DPO-noES-0.1 / all_results.json
hZzy's picture
End of training
2290771 verified
{
"epoch": 2.992914501653283,
"eval_dpo_loss": 0.8653034567832947,
"eval_logits": -1.8165048360824585,
"eval_logps": -132.85667419433594,
"eval_loss": 0.849328875541687,
"eval_objective": 0.8653034567832947,
"eval_ranking_simple": 0.534679114818573,
"eval_regularize": 0.8653034567832947,
"eval_runtime": 374.0313,
"eval_samples": 5790,
"eval_samples_per_second": 15.48,
"eval_steps_per_second": 1.291,
"eval_wo_beta": 10.941845893859863,
"total_flos": 0.0,
"train_loss": 0.37872021553586377,
"train_runtime": 34657.2213,
"train_samples": 50802,
"train_samples_per_second": 4.398,
"train_steps_per_second": 0.03
}