Qwen3-1.7B-Base-DPO-Tulu3-decontaminated-masked/train_results.json

{
    "epoch": 1.0,
    "total_flos": 0.0,
    "train_loss": 3.058436307772784,
    "train_runtime": 6320.6844,
    "train_samples": 272585,
    "train_samples_per_second": 43.126,
    "train_steps_per_second": 0.337
}