tinyllama-1.1b-dpo-pku-saferlhf_2/train_results.json

{
    "epoch": 0.9997600191984641,
    "total_flos": 0.0,
    "train_loss": 0.49830314429746014,
    "train_runtime": 2918.5416,
    "train_samples_per_second": 11.421,
    "train_steps_per_second": 0.714
}