Files
llama-3-8b-base-sft-hh-harm…/all_results.json
ModelHub XC 38db2426ef 初始化项目,由ModelHub XC社区提供模型
Model: W-61/llama-3-8b-base-sft-hh-harmless-4xh200
Source: Original Platform
2026-04-25 14:10:50 +08:00

14 lines
417 B
JSON

{
"epoch": 0.9975786924939467,
"eval_loss": 1.4828004837036133,
"eval_runtime": 4.0547,
"eval_samples": 2303,
"eval_samples_per_second": 183.984,
"eval_steps_per_second": 5.919,
"total_flos": 7.598970652485222e+16,
"train_loss": 1.8085910475369795,
"train_runtime": 916.9479,
"train_samples": 42336,
"train_samples_per_second": 14.402,
"train_steps_per_second": 0.225
}