Manual Installation
In most cases, we recommend using automatic installation via an easy-install
script.
For those who want to install everything manually, here you will find step-by-step instructions on what the script does.
Virtual Environment creation
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
PyTorch installation
Note
On macOS with Apple Silicon 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/rocm6.0
For AMD graphics cards on Windows install PyTorch with DirectML support using pip:
pip install torch-directml
Note
Python3.10 is the only currently supported version by torch-directml.
For NVIDIA graphics cards on both Linux or Windows install PyTorch using pip:
pip install torch torchvision torchaudio
Install Visionatrix
Install Visionatrix from the previously cloned sources using pip:
pip install .
Run Visionatrix initialization command using python:
python -m visionatrix install
Run Visionatrix
Execute from the activated virtual environment run command using python:
python -m visionatrix run --ui
Manual Update
Pull last changes from repository with git:
git pull
Execute update command from activated virtual environment with python:
python -m visionatrix update