MMR-GRPO-lambda-0.6 / train_results.json
kangdawei's picture
Model save
d62811f verified
{
"total_flos": 0.0,
"train_loss": 0.0016669491224483837,
"train_runtime": 56169.988,
"train_samples": 7000,
"train_samples_per_second": 0.427,
"train_steps_per_second": 0.009
}