HITsz-TMG/Uni-MoE
A Mixture-of-Experts-based omnimodal large language model that understands and generates across multiple modalities.
Not currently ranked — collecting fresh signals.
star history
Uni-MoE is a family of large multimodal models based on Mixture of Experts architecture. It supports understanding and generation across text, image, audio, and other modalities. The project provides multiple model versions with corresponding training code, pretrained checkpoints, and academic publications released through Hugging Face.
Frequently asked
- What is HITsz-TMG/Uni-MoE?
- A Mixture-of-Experts-based omnimodal large language model that understands and generates across multiple modalities.
- Is Uni-MoE open source?
- Yes — HITsz-TMG/Uni-MoE is an open-source project tracked on heatdrop.
- What language is Uni-MoE written in?
- HITsz-TMG/Uni-MoE is primarily written in Python.
- How popular is Uni-MoE?
- HITsz-TMG/Uni-MoE has 1.1k stars on GitHub.
- Where can I find Uni-MoE?
- HITsz-TMG/Uni-MoE is on GitHub at https://github.com/HITsz-TMG/Uni-MoE.