by huanngzh
Open source · 16k downloads · 61 likes
The MV-Adapter is an innovative tool designed to transform text-to-image generation models into multi-view generators, enabling the creation of high-fidelity images from multiple angles. It performs exceptionally well with various custom models such as DreamShaper or Animagine, as well as techniques like LCM or ControlNet, offering great flexibility in use. By generating multi-view images from either text or images, it simplifies 3D reconstruction or 3D texture generation with geometric cues. Its key strength lies in its ability to produce arbitrary views, catering to diverse needs in visual creation and 3D modeling.
Project Page | Paper (ArXiv) | Paper (HF) | Code | Gradio demo
Create High-fidelity Multi-view Images with Various Base T2I Models and Various Conditions.
MV-Adapter is a creative productivity tool that seamlessly transfer text-to-image models to multi-view generators.
Highlights:
| Model | Base Model | HF Weights | Demo Link |
|---|---|---|---|
| Text-to-Multiview | SD2.1 | mvadapter_t2mv_sd21.safetensors | |
| Text-to-Multiview | SDXL | mvadapter_t2mv_sdxl.safetensors | General / Anime |
| Image-to-Multiview | SD2.1 | mvadapter_i2mv_sd21.safetensors | |
| Image-to-Multiview | SDXL | mvadapter_i2mv_sdxl.safetensors | Demo |
| Text-Geometry-to-Multiview | SDXL | mvadapter_tg2mv_sdxl.safetensors | |
| Image-Geometry-to-Multiview | SDXL | mvadapter_ig2mv_sdxl.safetensors | |
| Image-to-Arbitrary-Views | SDXL |
Refer to our Github repository.
If you find this work helpful, please consider citing our paper:
@article{huang2024mvadapter,
title={MV-Adapter: Multi-view Consistent Image Generation Made Easy},
author={Huang, Zehuan and Guo, Yuanchen and Wang, Haoran and Yi, Ran and Ma, Lizhuang and Cao, Yan-Pei and Sheng, Lu},
journal={arXiv preprint arXiv:2412.03632},
year={2024}
}