Files
Qwen3-4B_Paper_Impact_datas…/train_results.json
ModelHub XC 82c479f443 初始化项目,由ModelHub XC社区提供模型
Model: FlyPig23/Qwen3-4B_Paper_Impact_dataset_SFT_1ep
Source: Original Platform
2026-04-20 19:34:24 +08:00

8 lines
208 B
JSON

{
"epoch": 1.0,
"total_flos": 2.587386052566057e+17,
"train_loss": 0.13282569709108838,
"train_runtime": 1624.0926,
"train_samples_per_second": 2.629,
"train_steps_per_second": 0.041
}