MKJia/MGVQ
[Arxiv'25] MGVQ: Could VQ-VAE Beat VAE? A Generalizable Tokenizer with Multi-group Quantization
This project offers a sophisticated image tokenizer that transforms high-resolution images into compact digital tokens, which can then be used to reconstruct images with exceptional quality. It handles various image resolutions, from standard to ultra-high-definition, delivering superior detail and accuracy compared to other methods. This is ideal for professionals in computer vision or digital media who work with advanced image generation and reconstruction tasks.
No commits in the last 6 months.
Use this if you need to efficiently tokenize and reconstruct images, especially at ultra-high resolutions, with leading accuracy for both visual fidelity and semantic content.
Not ideal if you are looking for a basic image compression tool or a solution for general-purpose image editing.
Stars
55
Forks
4
Language
Python
License
—
Category
Last pushed
Sep 16, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/MKJia/MGVQ"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
jxhe/vae-lagging-encoder
PyTorch implementation of "Lagging Inference Networks and Posterior Collapse in Variational...
chaitanya100100/VAE-for-Image-Generation
Implemented Variational Autoencoder generative model in Keras for image generation and its...
taldatech/soft-intro-vae-pytorch
[CVPR 2021 Oral] Official PyTorch implementation of Soft-IntroVAE from the paper "Soft-IntroVAE:...
lavinal712/AutoencoderKL
Train Your VAE: A VAE Training and Finetuning Script for SD/FLUX
Rayhane-mamah/Efficient-VDVAE
Official Pytorch and JAX implementation of "Efficient-VDVAE: Less is more"