The Digital Life Project is an innovative endeavor that utilizes cutting-edge technologies in computer vision (CV) to build a digital version of oneself. This project aims to bridge the gap between the physical and digital worlds, offering a unique and personal digital experience.
The documentations are under revision, you can try the demo at Digital-Life-Project
- Face Manipulation
- Face Swap: Advanced computer vision technology enabling users to swap faces in images or videos, creating a more personalized digital avatar.
- Video Retalking: Upcoming feature to manipulate video content so that the digital avatar can mimic user speech, enhancing realism and personalization.
- Audio Manipulation
- Text-to-Speech (TTS): Converts user-input text into natural-sounding speech, enhancing the interactivity of the digital avatar.
- Background Music (BGM) Separation: Sophisticated audio processing technology to separate background music from vocals, allowing for clear and distinct audio outputs.
-
Sound Cloning by Sovits: A future enhancement to clone a user's voice using Sovits technology, providing a more authentic and personalized audio experience for the digital avatar.
-
LLM-based chatting experience: A future enhancement to chat with a user's voice using Sovits technology, providing a more authentic and personalized chatting experience for the digital avatar.
To get started with the Digital Life Project, follow these steps:
- Installation:
-
Basic Installation:
# Clone the repository git clone https://github.com/deerestFarther/Digital-Life-Project.git # Navigate to the project directory cd Digital-Life-Project # Create and activate a new conda environment conda create -n dlp python=3.10 conda activate dlp # Install dependencies python -m pip install --no-cache-dir -r requirements.txt
-
For Nvidia GPU user, you may need to reinstall torch by:
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia --force-reinstall
This command on works on cuda-11.8 and within conda environment. For more installation command, please check out PyTorch.org
-
For M-series MacOS user, you may need to install pytorch-nightly to use torch on M-SoC MacOS by:
conda install pytorch torchvision torchaudio -c pytorch-nightly
- Usage:
- Starting at local port: 50954, you can change it inside "main.py"
python main.py
- On Win10 with one Nvidia RTX 4090, it requires at least 6G VRAM to use the features mentioned above
- Still testing on MacOS and cpu-win system
We welcome contributions to the Digital Life Project!
This project is licensed under the [MIT License] - see the LICENSE file for details.
- Contact information for the project maintainers.
-
Answers to common questions about the project.
-
Triton安装问题:XPixelGroup/DiffBIR#24