Files
P2-split3_prob_Qwen3-8B-Bas…/train_results.json
ModelHub XC 64f5a2c2ca 初始化项目,由ModelHub XC社区提供模型
Model: Hyeongwon/P2-split3_prob_Qwen3-8B-Base_0325-01
Source: Original Platform
2026-06-04 18:21:17 +08:00

9 lines
232 B
JSON

{
"epoch": 3.0,
"total_flos": 607742654087168.0,
"train_loss": 0.4391140220601574,
"train_runtime": 58765.7127,
"train_samples": 24640,
"train_samples_per_second": 1.258,
"train_steps_per_second": 0.013
}