train_results.json

{
    "epoch": 4.903225806451613,
    "num_input_tokens_seen": 1207760,
    "total_flos": 5.438488809413018e+16,
    "train_loss": 0.49016160156086125,
    "train_runtime": 2575.226,
    "train_samples_per_second": 9.626,
    "train_steps_per_second": 0.074
}