MMR-DR_GRPO-lambda-0.5 / train_results.json
kangdawei's picture
Model save
b61d851 verified
{
"total_flos": 0.0,
"train_loss": 0.0024257735135033726,
"train_runtime": 9466.1521,
"train_samples": 7000,
"train_samples_per_second": 2.535,
"train_steps_per_second": 0.053
}