QLoRa/all_results.json

{
    "epoch": 1.0,
    "train_loss": 1.0734075052920875,
    "train_runtime": 27866.917,
    "train_samples_per_second": 0.461,
    "train_steps_per_second": 0.461
}