Qwen2.5-1.5B-Open-R1-GRPO / eval_results.json
bushou's picture
End of training
ce79677 verified
{
"eval_loss": 0.03138502687215805,
"eval_runtime": 85.4883,
"eval_samples": 99,
"eval_samples_per_second": 1.158,
"eval_steps_per_second": 0.047
}