MMR-DR_GRPO-lambda-0.9 / train_results.json
kangdawei's picture
Model save
cb73f08 verified
{
"total_flos": 0.0,
"train_loss": 0.023616178158205003,
"train_runtime": 58431.9199,
"train_samples": 7000,
"train_samples_per_second": 0.411,
"train_steps_per_second": 0.009
}