SmolLM2-1.7B-16k/train_results.json

{
    "epoch": 1.0,
    "total_flos": 123288491655168.0,
    "train_loss": 0.7956359339436115,
    "train_runtime": 17588.8577,
    "train_samples_per_second": 2.952,
    "train_steps_per_second": 0.092
}