|
--- |
|
tags: |
|
- text-to-image |
|
- stable-diffusion |
|
license: apache-2.0 |
|
language: |
|
- en |
|
library_name: diffusers |
|
--- |
|
|
|
# StyleShot Model Card |
|
|
|
|
|
<div align="center"> |
|
|
|
[**Project Page**](https://styleshot.github.io) **|** [**Paper (ArXiv)**](https://arxiv.org/abs/) **|** [**Code**](https://github.com/tencent-ailab/IP-Adapter) |
|
</div> |
|
|
|
--- |
|
|
|
|
|
## Introduction |
|
|
|
we present StyleShot, a generalized plug-to-play style transfer method, capable of generating the high-quality stylized images that match the desired style from any reference image without test-time style-tuning. To the best of our knowledge, To the best of our knowledge, StyleShot is the first work to designate a style-aware encoder based on Stable Diffusion and a content-fusion encoder for better style and content integration, achieving the state-of-the-art text and image-driven style transfer performance compared to existing methods. |
|
|
|
![arch](./framework.png) |
|
|
|
## Models |
|
|
|
### StyleShot for SD 1.5 |
|
- [ip-adapter_sd15.bin](https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter_sd15.bin): use global image embedding from OpenCLIP-ViT-H-14 as condition |
|
- [ip-adapter_sd15_light.bin](https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter_sd15_light.bin): same as ip-adapter_sd15, but more compatible with text prompt |