hackduke1_DPO/train_results.json

{
    "epoch": 1.0,
    "train_loss": 0.6225535166428385,
    "train_runtime": 3531.6202,
    "train_samples_per_second": 0.983,
    "train_steps_per_second": 0.123
}