README.md

---
license: mit
datasets:
- imagenet-1k
language:
- en
- zh
---




# Model Card for VAR (Visual AutoRegressive) Transformers 🔥

<!-- Provide a quick summary of what the model is/does. -->

[![arXiv](https://img.shields.io/badge/arXiv%20papr-2404.02905-b31b1b.svg)](https://arxiv.org/abs/2404.02905)[![arXiv](https://img.shields.io/badge/demo%20website-VAR%20demo-lightblue)](https://var.vision)


VAR is a new visual generation framework that makes GPT-style models surpass diffusion models **for the first time**🚀, and exhibits clear power-law Scaling Laws📈 like large language models (LLMs).

<p align="center">
<img src="https://cdn-uploads.huggingface.co/production/uploads/60e73ffd06ad9ae5bbcfc52c/FusWBHW8uJgYWO02HFNGz.png" width=93%>
<p>
  
VAR redefines the autoregressive learning on images as coarse-to-fine "next-scale prediction" or "next-resolution prediction", diverging from the standard raster-scan "next-token prediction".

<p align="center">
<img src="https://github.com/FoundationVision/VAR/assets/39692511/3e12655c-37dc-4528-b923-ec6c4cfef178" width=93%>
<p>


This repo is used for hosting VAR's checkpoints.

For more details or tutorials see https://github.com/FoundationVision/VAR.