AvatarBooth - GitLite

KO | EN

gitlite — search

#react #ai #llm #deep-learning #linux #docker #java #awesome #vue #awesome-list #security #ios #openai #open-source

AvatarBooth

(★ 57)

Official implementation of “AvatarBooth: High-Quality and Customizable 3D Human Avatar Generation”

파일 탐색기

최종 버전 다운로드 (.zip)

- exhibit.jpg
- trump_dances.gif
- - obama.conf
- - transforms_train.json
- - transforms_train.json
- zero_beta_smpl.obj
- - dataset.cpython-38.pyc
  - dataset0.cpython-38.pyc
  - dataset2.cpython-38.pyc
  - embedder.cpython-38.pyc
  - fields.cpython-38.pyc
  - renderer.cpython-38.pyc
  - utils.cpython-38.pyc
- dataset.py
- dataset0.py
- dataset2.py
- embedder.py
- fields.py
- renderer.py
- utils.py
- stand_pose.npy
- zero_beta_stand_pose.pth
- zero_beta_stand_pose_small.pth
- - __init__.cpython-38.pyc
  - base.cpython-38.pyc
  - callbacks.cpython-38.pyc
  - config.cpython-38.pyc
  - misc.cpython-38.pyc
  - ops.cpython-38.pyc
  - processors.cpython-38.pyc
  - rasterize.cpython-38.pyc
  - saving.cpython-38.pyc
  - typing.cpython-38.pyc
- __init__.py
- base.py
- callbacks.py
- config.py
- misc.py
- ops.py
- processors.py
- rasterize.py
- saving.py
- typing.py
config.conf
drive.py
main.py
README.md
requirements.txt
simple_sd.py

# 설치 가이드

1. 코드 내려받기

git clone https://github.com/zeng-yifei/AvatarBooth

깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.

cd AvatarBooth

방금 내려받은 프로젝트 폴더 안으로 이동합니다.

2. Python

쉬움 추천

사전 준비물

Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
Python 3 설치 시 'Add Python to PATH' 옵션을 꼭 체크하세요 (Windows).

pip install -r requirements.txt

requirements.txt 등에 명시된 파이썬 라이브러리를 설치합니다.

✅ 에러 메시지 없이 실행되고 터미널에 안내 문구가 출력되면 정상입니다.

이 레포의 README에 적힌 실제 명령어를 그대로 가져왔습니다.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<h1>AvatarBooth: High-Quality and Customizable 3D Human Avatar Generation</h1>

<div>
    <a href='https://github.com/zeng-yifei?tab=repositories/' target='_blank'>Yifei Zeng</a><sup>1</sup>&emsp;
    <a href='https://github.com/YuanxunLu' target='_blank'>Yuanxun Lu</a><sup>1</sup>&emsp;
    <a href='https://github.com/jixinya' target='_blank'>Xinya Ji</a><sup>1</sup>&emsp;
    <a href='https://yoyo000.github.io/' target='_blank'>Yao Yao</a><sup>1</sup>&emsp;
    <a href='https://zhuhao-nju.github.io/home/' target='_blank'>Hao Zhu</a><sup>1+</sup>&emsp;
    <a href='https://cite.nju.edu.cn/People/Faculty/20190621/i5054.html' target='_blank'>Xun Cao</a><sup>1</sup>
</div>
<div>
    <sup>1</sup>Nanjing University
</div>
<div>
    <sup>+</sup>corresponding author
</div>

<h4 align="center">
  <a href="https://zeng-yifei.github.io/avatarbooth_page/" target='_blank'>[Project Page]</a> •
</h4>

# Install

For package installation, ensure that you have installed pytorch (tested on pytorch 2.1 cuda121 and pytorch 1.13 cuda 117):
```bash
pip install -r requirements.txt
```

For data preparation, register and download SMPL models [here](https://smpl.is.tue.mpg.de/). Put the downloaded models in the folder `smpl_models`. The folder structure should look like

```
./
├── ...
└── smpl_models/
    ├── smpl/
        ├── SMPL_FEMALE.pkl
        ├── SMPL_MALE.pkl
        └── SMPL_NEUTRAL.pkl
```

# Usage

To generate avatars. you can use:

```bash
python main.py --mode train --conf confs/examples/obama.conf
```

To use personalized model like LoRA or DreamBooth model, you can assign the corresponding file path in config file like:
```
general {
    sd_path = ... # assign DreamBooth path for whole body in huggingface format, e.g. stabilityai/stable-diffusion-2-1-base or stablediffusionapi/realistic-vision(recommanded and by default)
    sd_face_path = ... # assign DreamBooth path for face in huggingface format

lora_path = ... # assign lora path with safetensors, e.g. ./pretrained_models/A.safetensors
}
```

To animate the avatar, you can refer to the AvatarCLIP. With the same procedure, you can obtain a animatable fbx after processing the A-pose ply model.

# Teaser
<img src='./assets/exhibit.jpg' height='60%'>

# Animation
<img src='./assets/trump_dances.gif' height='60%'>

## Citation
If you find our work useful for your research, please consider citing the paper:
```
@inproceedings{Zeng2023AvatarBoothHA,
  title={AvatarBooth: High-Quality and Customizable 3D Human Avatar Generation},
  author={Yifei Zeng and Yuanxun Lu and Xinya Ji and Yao Yao and Hao Zhu and Xun Cao},
  year={2023}
}
```

# Acknowledgement
The code is built upon AvatarCLIP and Stable DreamFusion, we express great appreciation to the authors for their great work.