This repo is a SD-Webui extension for Composable T2I-Adapter (CoAdapter).
- Install the stable diffusion webui.
- Open "Extensions" tab.
- Open "Install from URL" tab in the tab.
- Enter URL of this repo to "URL for extension's git repository".
- Press "Install" button.
- Reload/Restart Web UI.
Once installed, the UI looks like:
Remember to check ``Enable'' box to activate the function of CoAdapter.
- Using functions in "txt2img" or "img2img" tab.
- Import the conditions you need into the window of Coadapter.
- Choosing a sampler in the "sampling method". Empirically, DDIM is all-right.
- For the output resolution, you can choose "Consistent with the condition map" or "Controlled by resize sliders".
Pre-trained models are downloaded automatically from https://huggingface.co/TencentARC/T2I-Adapter and placed in the "stable-diffusion-webui\extensions\T2I-Adapter\models" folder.
Downloading models from T2I-Adapter model zoom: https://github.com/TencentARC/T2I-Adapter/blob/main/docs/AdapterZoo.md
**Put models in your "stable-diffusion-webui\extensions\T2I-Adapter\models".
| Sketch | Canny | Depth | Color (Spatial) | Style | Results |
|---|---|---|---|---|---|
 |
 |
 |
|||
|
 |
 |
|||
 |
 |
 |
|||
|
 |
|
 |
||
|
 |
|
 |