MMR-DR_GRPO-lambda-0.8 / train_results.json
kangdawei's picture
Model save
3948a7a verified
{
"total_flos": 0.0,
"train_loss": 0.023119730316335334,
"train_runtime": 57324.27,
"train_samples": 7000,
"train_samples_per_second": 0.419,
"train_steps_per_second": 0.009
}