Files
qwen3-8b-base-beta-dpo-hh-h…/train_results.json

9 lines
232 B
JSON
Raw Permalink Normal View History

{
"epoch": 0.999244142101285,
"total_flos": 0.0,
"train_loss": 1.1663504292943294,
"train_runtime": 3087.1314,
"train_samples": 42336,
"train_samples_per_second": 13.714,
"train_steps_per_second": 0.214
}