models/validator-cond-1B-orpo-iter1-collab-paper/train_results.json

{
    "epoch": 5.807339449541285,
    "total_flos": 0.0,
    "train_loss": 0.14153494492173194,
    "train_runtime": 748.5587,
    "train_samples": 545,
    "train_samples_per_second": 4.275,
    "train_steps_per_second": 0.534
}