llama-3-8b-base-epsilon-dpo-ultrafeedback-8xh200/train_results.json

{
    "epoch": 0.9989528795811519,
    "total_flos": 0.0,
    "train_loss": 2.463846208664356,
    "train_runtime": 4358.2481,
    "train_samples": 61135,
    "train_samples_per_second": 14.027,
    "train_steps_per_second": 0.109
}