mit-han-lab/efficientvit
Efficient vision foundation models for high-resolution image generation using diffusion architectures and vision transformers.
Not currently ranked — collecting fresh signals.
star history
This repository provides efficient vision foundation models including Deep Compression Autoencoders (DC-AE) for high-resolution diffusion models and vision transformer architectures. It supports tasks such as image generation on ImageNet at 512x512 resolution and segmentation. The project includes implementations of models like SANA (text-to-image) and USiT for state-of-the-art generation quality.
Frequently asked
- What is mit-han-lab/efficientvit?
- Efficient vision foundation models for high-resolution image generation using diffusion architectures and vision transformers.
- Is efficientvit open source?
- Yes — mit-han-lab/efficientvit is open source, released under the Apache-2.0 license.
- What language is efficientvit written in?
- mit-han-lab/efficientvit is primarily written in Python.
- How popular is efficientvit?
- mit-han-lab/efficientvit has 3.3k stars on GitHub.
- Where can I find efficientvit?
- mit-han-lab/efficientvit is on GitHub at https://github.com/mit-han-lab/efficientvit.