llama-factory/saves/Phi-3.5-mini-instruct/eval_results.json

{
    "epoch": 5.9786476868327405,
    "eval_loss": 1.6644203662872314,
    "eval_runtime": 2.6994,
    "eval_samples_per_second": 17.041,
    "eval_steps_per_second": 17.041
}