Files
llama-3-8b-base-margin-dpo-…/train_results.json
ModelHub XC 15c4f36648 初始化项目,由ModelHub XC社区提供模型
Model: W-61/llama-3-8b-base-margin-dpo-ultrafeedback-8xh200
Source: Original Platform
2026-04-24 11:32:05 +08:00

9 lines
233 B
JSON

{
"epoch": 0.9989528795811519,
"total_flos": 0.0,
"train_loss": 1.1476524381017785,
"train_runtime": 4456.6592,
"train_samples": 61135,
"train_samples_per_second": 13.718,
"train_steps_per_second": 0.107
}