python-bytes-distilgpt2/train_results.json

{
    "epoch": 3.0,
    "train_loss": 3.0642820732502996,
    "train_runtime": 1709.0811,
    "train_samples": 10112,
    "train_samples_per_second": 17.75,
    "train_steps_per_second": 4.437
}