repository for multiview character generation based on SDv1.5
- torch
- torchvision
- torchaudio
- diffusers
- accelerate
- xformers
- triton
- transformers
- 16GB GPU
git clone https://github.com/ArcherFMY/Multiview-Avatar.git
cd Multiview-Avatar
pip install -r requirements.txt
Download models from Baidu Disk or Google Drive. Unzip models.zip into the root directory of the project.
${ROOT}
|-- data
| |-- test.jpg
|-- models
| |-- base-model
| |-- control-imgs
| |-- controlnet-lineart
| |-- controlnet-pose
| |-- wildcards
|-- mvavatar
|...
Or download the models from hugging face (link1 or link2).
from mvavatar import MVAvatar
model = MVAvatar.from_pretrained('models', device='cuda')
image = model.inference('1 girl')
image.save('data/test.jpg')
see more in demo.py
This code is released under the Apache License 2.0 (refer to the LICENSE file for details).