Learning Timeline
Key Insights
Python Version Tip
Avoid using Python 3.13 for AI projects for now. Many AI libraries like Torch are not yet fully stable on that version. Use 3.11 or 3.12 for the best stability.
Important Step: Environment Path
If the 'conda' command is not recognized, ensure you have manually added the /Scripts folder path to your System Environment Variables.
Pre-update Sequence
Always run 'conda activate Wan2GP' before executing 'pip install'. If you skip this step, libraries will be installed globally and could break your computer's original Python setup.
Step by Step
Steps to Install Git and Clone the Repository
- Visit the official Git website and download the '64-bit Git for Windows Setup' installer.
- Open the downloaded .exe file and click 'Next' on all default settings until the installation is complete.
- Go to your Desktop, right-click and select 'Open in Terminal' or type 'cmd' in the folder address bar.
- Type the command 'git clone https://github.com/GP-Unit/Wan2GP.git' (or the relevant repo link) to download the source code.
- Enter the project folder by typing 'cd Wan2GP' in the Command Prompt.
Setting Up the Virtual Environment with Miniconda
- Download Miniconda for Python 3.11 or 3.12 from repo.anaconda.com (avoid 3.13 due to AI model compatibility issues).
- Run the installer, select 'All Users', and check the 'Clear the package cache' option before clicking finish.
- Locate the 'Scripts' folder in your Miniconda installation directory (e.g., C:\ProgramData\miniconda3\Scripts).
- Copy the folder path using 'Copy as path'.
- Open 'Environment Variables' in Windows, find 'Path' under System Variables, click 'Edit', then 'New', and paste the path.
- Open a new Command Prompt and type 'conda --version' to verify the installation.
- Type 'conda create -n Wan2GP python=3.11' and press 'Y' to build a dedicated environment.
Installing Dependencies and Model Libraries
- Activate the environment by typing 'conda activate Wan2GP'.
- Ensure the environment name '(Wan2GP)' appears at the beginning of your prompt line.
- Install PyTorch and Torchvision using the command 'pip install torch torchvision torchaudio' (file size is ~3GB).
- Install the remaining required libraries by typing 'pip install -r requirements.txt'.
- If you are updating an older version, type 'git pull' before running 'pip install -r requirements.txt'.
