mistral-7b-base-epsilon-dpo-hh-harmless-4xh200-batch-64/train_results.json

{
    "epoch": 0.999244142101285,
    "total_flos": 0.0,
    "train_loss": 1.0336073330418967,
    "train_runtime": 2985.5995,
    "train_samples": 42336,
    "train_samples_per_second": 14.18,
    "train_steps_per_second": 0.221
}