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

{
    "epoch": 1.0,
    "total_flos": 0.0,
    "train_loss": 1.5426912418834826,
    "train_runtime": 5367.6535,
    "train_samples": 122270,
    "train_samples_per_second": 22.779,
    "train_steps_per_second": 0.178
}