IQA-PyTorch is a comprehensive image quality assessment library offering both full-reference metrics like PSNR, SSIM, MS-SSIM, LPIPS, and FID, as well as no-reference metrics including NIQE, BRISQUE, NIMA, DBCNN, MUSIQ, and TOPIQ. The toolbox is built entirely in Python and PyTorch with GPU acceleration, providing implementations calibrated against official MATLAB scripts. It supports both CLI usage and programmatic API for evaluating image quality, aesthetics, and perceptual similarity.