Qwen2-0.5B-Reward_debug_mas / all_results.json
Shahradmz's picture
End of training
b614c4a verified
{
"epoch": 1.0,
"eval_accuracy": 1.0,
"eval_loss": 2.145418825233447e-12,
"eval_runtime": 1.0751,
"eval_samples_per_second": 88.365,
"eval_steps_per_second": 11.162
}