HunyuanCustom

HunyuanCustom is a multi-modal, conditional, and controllable generation model centered on subject consistency, built upon the Hunyuan Video generation framework. It enables the generation of subject-consistent videos conditioned on text, images, audio, and video inputs. Specifically, HunyuanCustom introduces an image-text fusion module based on LLaVA to facilitate interaction between images and text, allowing identity information from images to be effectively integrated into textual descriptions.
HunyuanCustom works by first addressing the image-text conditioned generation task by introducing a text-image fusion module based on LLaVA for enhanced multi-modal understanding, along with an image ID enhancement module that leverages temporal concatenation to reinforce identity features across frames. To enable audio- and video-conditioned generation, HunyuanCustom proposes modality-specific condition injection mechanisms: an AudioNet module that achieves hierarchical alignment via spatial cross-attention, and a video-driven injection module that integrates latent-compressed conditional video through a patchify-based feature-alignment network.
The model ultimately achieves decoupled control over image, audio, and video conditions, demonstrating great potential in subject-centric multi-modal video generation. HunyuanCustom is particularly useful for applications requiring customized video generation with subject consistency, such as video production, advertising, and social media content creation. Its ability to support various input modalities and generate high-quality videos makes it a valuable tool for professionals and creators looking to produce engaging and personalized content.