Qwen2.5-0.5B-Open-R1-GRPO / all_results.json
coco3143's picture
End of training
c4167a7 verified
{
"eval_loss": 0.006717957090586424,
"eval_runtime": 82.9232,
"eval_samples": 99,
"eval_samples_per_second": 1.194,
"eval_steps_per_second": 0.109,
"total_flos": 0.0,
"train_loss": 0.020143496153197993,
"train_runtime": 113002.1292,
"train_samples": 72441,
"train_samples_per_second": 0.641,
"train_steps_per_second": 0.013
}