Qwen3-8B-base-Open-R1-GRPO_dapo_acc_16384_nokl/train_results.json

{
    "total_flos": 0.0,
    "train_loss": -0.00041466000025666894,
    "train_runtime": 232889.004,
    "train_samples": 14116,
    "train_samples_per_second": 1.924,
    "train_steps_per_second": 0.004
}