This repository provides the official implementation of HVI-CIDNet, a low-light image enhancement method published at CVPR 2025 and NTIRE 2025. The approach introduces a new Hue Value Intensity color space designed to better represent color and brightness relationships in low-light conditions. CIDNet employs a transformer architecture to learn enhancement mappings in this color space, and the project includes training code, pretrained models, and a Gradio demo interface.