Qwen-2.5-7B-Simple-RL / all_results.json
Maker-0409's picture
End of training
811e76a verified
{
"eval_loss": 0.03460121154785156,
"eval_runtime": 32312.0729,
"eval_samples": 5000,
"eval_samples_per_second": 0.155,
"eval_steps_per_second": 0.011,
"total_flos": 0.0,
"train_loss": 4841.422249500714,
"train_runtime": 180396.3107,
"train_samples": 7500,
"train_samples_per_second": 0.042,
"train_steps_per_second": 0.003
}