Files
qwen3-8b-base-new-dpo-hh-ha…/all_results.json
ModelHub XC 9268b0b929 初始化项目,由ModelHub XC社区提供模型
Model: jackf857/qwen3-8b-base-new-dpo-hh-harmless-4xh200-batch-64-q_t-0.45-s_star-0.6
Source: Original Platform
2026-05-13 01:12:57 +08:00

23 lines
842 B
JSON

{
"epoch": 0.999244142101285,
"eval_fcm_dpo/beta": 0.047326020896434784,
"eval_logits/chosen": 1.3826223611831665,
"eval_logits/rejected": 1.2558915615081787,
"eval_logps/chosen": -92.67096710205078,
"eval_logps/ref_chosen": -86.90177917480469,
"eval_logps/ref_rejected": -96.69639587402344,
"eval_logps/rejected": -114.26859283447266,
"eval_loss": 0.5416048169136047,
"eval_margin_dpo/margin_mean": 11.803030967712402,
"eval_margin_dpo/margin_std": 18.962507247924805,
"eval_runtime": 42.3076,
"eval_samples": 2303,
"eval_samples_per_second": 54.435,
"eval_steps_per_second": 1.702,
"total_flos": 0.0,
"train_loss": 1.1374638212250148,
"train_runtime": 2122.2138,
"train_samples": 42336,
"train_samples_per_second": 19.949,
"train_steps_per_second": 0.311
}