qwen3-8b-aaai27-flagship-ht-mnpo-conciseness-s42/train_results.json

{
    "epoch": 0.8594449418084154,
    "total_flos": 0.0,
    "train_loss": 0.7468786084651947,
    "train_runtime": 6934.2752,
    "train_samples": 16755,
    "train_samples_per_second": 2.077,
    "train_steps_per_second": 0.13
}