Installation¶

In most cases, we recommend using automatic installation via an easy-install script.

Download and execute easy_install.py script:

Note

It will clone this repository into the current folder and perform the installation.

After installation you can always run easy_install from the "scripts" folder.

With wget:

wget -O easy_install.py https://raw.githubusercontent.com/Visionatrix/Visionatrix/main/scripts/easy_install.py && python3 easy_install.py

or with curl:

curl -o easy_install.py https://raw.githubusercontent.com/Visionatrix/Visionatrix/main/scripts/easy_install.py && python3 easy_install.py

Manual Installation¶

For those who want to install everything manually, here you will find step-by-step instructions on what the script does.

First clone the repository with git:

git clone https://github.com/Visionatrix/Visionatrix.git && cd Visionatrix

Setup the virtual environment with python:

python -m venv venv

Activate Virtual Environment(Linux/macOS) with source:

source venv/bin/activate

Activate Virtual Environment(Windows) with powershell:

.\venv\Scripts\Activate.ps1

Note

On macOS currently no action is needed.

For AMD graphic cards on Linux install ROCM version of PyTorch using pip:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.1

For AMD graphics cards on Windows install PyTorch with DirectML support using pip:

pip install torch-directml

Python3.10 is the only currently supported version by torch-directml.

For NVIDIA graphics cards on Linux install PyTorch with next pip command:

pip install torch torchvision torchaudio

For NVIDIA graphics cards on Windows install PyTorch using pip specifying PyTorch and CUDA version:

pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu121

Install Visionatrix from the previously cloned sources using pip:

pip install .

Run Visionatrix initialization command using python:

python -m visionatrix install

Execute from the activated virtual environment run command using python:

python -m visionatrix run --ui

Note

On Windows this method currently does not supported.

With the easy_install.py script:

python3 scripts/easy_install.py

and select option Update (2)

Note

On Windows this method requires installed git and Visual Studio Compilers

Pull last changes from repository with git:
```
git pull
```
Execute update command from activated virtual environment with python:
```
python -m visionatrix update
```

Development versions are updated only to development versions, release versions only to release ones.

Note

If you are not a developer, you are better off using the release version, as they should be more stable.

The update scheme in easy_install.py is quite simple; everything is done with ordinary Git commands.

If the current version is a dev release or the current branch is main then:
1. Check out the main branch.
2. Pull the latest changes from the remote repository.
If the current version is a tagged release version:
1. Determine the latest tag for the current major version, and if a newer version tag is found, check out the latest version tag within the current major version.
2. If no newer version is found within the current major version, check for the next major version.
3. If a newer major version tag is found, prompt the user to update to this newer major version.
After checking out the appropriate version, executes a pip install command to update the Python packages.
Finally, executes the python3 -m visionatrix update command to ensure that any additional necessary updates are applied (ComfyUI, custom nodes, flows).