Qwen-2.5-7B-base-Simple-RL / eval_results.json
kkish's picture
End of training
adfb47d verified
{
"eval_loss": 0.001339427661150694,
"eval_runtime": 3901.9909,
"eval_samples": 5000,
"eval_samples_per_second": 1.281,
"eval_steps_per_second": 0.012
}