Files
qwen3-4b-think-baseline-ful…/train_results.json
ModelHub XC 84ba2b4166 初始化项目,由ModelHub XC社区提供模型
Model: modrill/qwen3-4b-think-baseline-full-sft
Source: Original Platform
2026-07-26 22:34:10 +08:00

8 lines
205 B
JSON

{
"epoch": 2.0,
"total_flos": 1548137571287040.0,
"train_loss": 0.6697529787664481,
"train_runtime": 98109.4565,
"train_samples_per_second": 0.546,
"train_steps_per_second": 0.009
}