qwen3-8b-aaai27-flagship-inpo-avg-s43/train_results.json

{
    "epoch": 0.8599068434252956,
    "total_flos": 0.0,
    "train_loss": 4429.094605034722,
    "train_runtime": 7055.131,
    "train_samples": 16746,
    "train_samples_per_second": 2.041,
    "train_steps_per_second": 0.128
}