Qwen-2.5-7B-Simple-RL / all_results.json
kkish's picture
End of training
d01f41d verified
{
"eval_loss": 0.00022496927704196423,
"eval_runtime": 4404.9486,
"eval_samples": 5000,
"eval_samples_per_second": 1.135,
"eval_steps_per_second": 0.01,
"total_flos": 0.0,
"train_loss": 0.01571170037346621,
"train_runtime": 7483.8014,
"train_samples": 7500,
"train_samples_per_second": 1.002,
"train_steps_per_second": 0.008
}