Qwen2.5-1.5B-Open-R1-GRPO / eval_results.json
Emo3331's picture
End of training
cd927a6 verified
raw
history blame contribute delete
167 Bytes
{
"eval_loss": 0.27600744366645813,
"eval_runtime": 4.1811,
"eval_samples": 99,
"eval_samples_per_second": 23.678,
"eval_steps_per_second": 0.957
}