mistral-7b-grok/train_results.json

{
    "epoch": 1.0,
    "train_loss": 0.9725383741046311,
    "train_runtime": 5277.8235,
    "train_samples": 211055,
    "train_samples_per_second": 26.46,
    "train_steps_per_second": 0.103
}