Files
Qwen2.5-1.5B-SFT-Tulu3-deco…/all_results.json

8 lines
216 B
JSON
Raw Permalink Normal View History

{
"total_flos": 2402699861753856.0,
"train_loss": 0.7249908164597705,
"train_runtime": 27625.1008,
"train_samples": 936509,
"train_samples_per_second": 67.801,
"train_steps_per_second": 0.53
}