Qwen3-8B-grpo-medmcqa-medi70 / train_results.json
mlxha's picture
Model save
2c4170f verified
{
"total_flos": 0.0,
"train_loss": 22251.08962759373,
"train_runtime": 39419.1987,
"train_samples": 16885,
"train_samples_per_second": 0.428,
"train_steps_per_second": 0.004
}