Files
qwen3-8b-base-epsilon-dpo-h…/train_results.json
ModelHub XC 3b8f28089f 初始化项目,由ModelHub XC社区提供模型
Model: jackf857/qwen3-8b-base-epsilon-dpo-hh-helpful-4xh200-batch-64-20260424-040306
Source: Original Platform
2026-05-14 01:00:02 +08:00

9 lines
218 B
JSON

{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 1.0212064078200755,
"train_runtime": 3308.3388,
"train_samples": 43598,
"train_samples_per_second": 13.178,
"train_steps_per_second": 0.206
}