Files
OpenRS-GRPO-S/train_results.json

8 lines
200 B
JSON
Raw Permalink Normal View History

{
"total_flos": 0.0,
"train_loss": -0.17849066271579692,
"train_runtime": 2375.2815,
"train_samples": 7000,
"train_samples_per_second": 3.536,
"train_steps_per_second": 0.147
}