MMR-DR_GRPO-OpenS1 / train_results.json
kangdawei's picture
Model save
e3c6947 verified
{
"total_flos": 0.0,
"train_loss": 0.030813425727123103,
"train_runtime": 83450.6868,
"train_samples": 18615,
"train_samples_per_second": 0.288,
"train_steps_per_second": 0.006
}