Qwen2.5-7B-Open-R1-GRPO-math-lighteval-1epochstop-withformat/all_results.json

{
    "total_flos": 0.0,
    "train_loss": 0.02949339959234225,
    "train_runtime": 14062.8623,
    "train_samples": 7500,
    "train_samples_per_second": 0.533,
    "train_steps_per_second": 0.005
}