Files
llama3-8b-full-pretrain-was…/train_results.json
ModelHub XC 40d1f746eb 初始化项目,由ModelHub XC社区提供模型
Model: shuoxing/llama3-8b-full-pretrain-wash-c4-3-9m-bs4
Source: Original Platform
2026-06-12 17:28:17 +08:00

8 lines
201 B
JSON

{
"epoch": 3.0,
"total_flos": 16450670346240.0,
"train_loss": 1.665625898082319,
"train_runtime": 7008.0374,
"train_samples_per_second": 3.532,
"train_steps_per_second": 0.883
}