ronpo-qwen3-8b-fair-ronpo-full-expect-s42/train_results.json

{
    "epoch": 0.2860548271752086,
    "total_flos": 0.0,
    "train_loss": 0.430257171690464,
    "train_runtime": 7318.8237,
    "train_samples": 50340,
    "train_samples_per_second": 1.968,
    "train_steps_per_second": 0.123
}