DRA-GRPO / train_results.json
LLucass's picture
Model save
b50d192 verified
{
"total_flos": 0.0,
"train_loss": 3.568841239768084e-05,
"train_runtime": 22061.4799,
"train_samples": 7000,
"train_samples_per_second": 0.435,
"train_steps_per_second": 0.005
}