qwen25_05b_base_full_ft_lunarlander_a4000/train_results.json

{
    "epoch": 1.0,
    "total_flos": 5.72407910760448e+16,
    "train_loss": 0.07194560643361539,
    "train_runtime": 34224.4204,
    "train_samples_per_second": 5.727,
    "train_steps_per_second": 0.179
}