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

{
    "epoch": 5.9786476868327405,
    "total_flos": 1.7533662523283866e+17,
    "train_loss": 1.8585307757059732,
    "train_runtime": 2890.6103,
    "train_samples_per_second": 9.303,
    "train_steps_per_second": 0.073
}