QLora/all_results.json

{
    "epoch": 1.0,
    "train_loss": 0.24151210376558752,
    "train_runtime": 349438.2347,
    "train_samples_per_second": 1.019,
    "train_steps_per_second": 1.019
}