Files
P2-split4_prob_Qwen3-8B-Bas…/train_results.json
ModelHub XC 7fac2f163b 初始化项目,由ModelHub XC社区提供模型
Model: Hyeongwon/P2-split4_prob_Qwen3-8B-Base_0325-01
Source: Original Platform
2026-06-04 17:42:46 +08:00

9 lines
232 B
JSON

{
"epoch": 3.0,
"total_flos": 605566838833152.0,
"train_loss": 0.4404701322238902,
"train_runtime": 58825.4442,
"train_samples": 24616,
"train_samples_per_second": 1.255,
"train_steps_per_second": 0.013
}