Qwen2.5-1.5B-Open-R1-GRPO / eval_results.json
Penghe's picture
End of training
5106575 verified
{
"eval_loss": 0.006964259315282106,
"eval_runtime": 44.1943,
"eval_samples": 99,
"eval_samples_per_second": 2.24,
"eval_steps_per_second": 0.091
}