The left side is the input image, and the right side is the output image.

In my opinion, skipping the refiner model preserves more image details, while using the refiner makes the image smoother but loses fine details. Therefore, if you plan to use HunyuanImage 2.1, I recommend not using the refiner model — it runs slower and results in noticeable detail loss.

https://huggingface.co/QuantStack/HunyuanImage-2.1-Refiner-GGUF