llama-factory/saves/Qwen2-7B-Instruct/eval_results.json

{
    "epoch": 5.9786476868327405,
    "eval_loss": 1.6233642101287842,
    "eval_runtime": 2.6127,
    "eval_samples_per_second": 17.607,
    "eval_steps_per_second": 17.607
}