Skip to content

Setup

Here's a guide for installing the project necessities, with separate tutorials for Windows and macOS.

Python Installation

Windows

1. Install Git

Choose one of the following methods:

  • Using winget tool in PowerShell or Command Prompt (requires terminal restart afterwards to take effect): 

    winget install --id Git.Git -e --source winget

  • Or download the standalone installer from https://git-scm.com/download/win

2. Navigate to Project Folder

cd <project folder>

3. Check Installed Python Versions

To see which Python versions are installed on your system, run: 

py -0p

4. Create a Virtual Environment

  • If Python 3.11 is installed along with other versions: 
    py -3.11 -m venv .venv
  • If only Python 3.11 is installed: 
    py -m venv .venv

5. Resolve Execution Policy Error

If you encounter an "about_Execution_Policies" error, run this command and retry step 4.: 

Set-ExecutionPolicy Unrestricted -Scope CurrentUser

6. Activate the Virtual Environment

.venv\Scripts\activate

7. Install Requirements

pip install -r requirements_win.txt

macOS

1. Install Git

Choose one of the following methods:

2. Navigate to Project Folder

cd <project folder>

3. Create a Virtual Environment

  • For Python 3.11 (replace with your desired version if different): 
    python3.11 -m venv .venv
  • If you want to use the default Python 3 version: 
    python3 -m venv .venv

4. Activate the Virtual Environment

source .venv/bin/activate

5. Install Requirements

pip install -r requirements_mac.txt

MATLAB Installation

This is how to set up Python for MATLAB.

1. Download Python 3.11.9

  • Download from the official site (DO NOT download from MS Store):
  • Python 3.11.9 for Windows (amd64)
  • During installation, ensure you check the box for "Add python.exe to PATH".

2. Navigate to the MATLAB Project Folder via Windows Command Prompt

cd <project folder>

3. Check Installed Python Versions

py -0p

4. Create a Virtual Environment

  • Make a virtual environment to store Python modules by typing: 
    py -3.11 -m venv .venv
  • If multiple Python versions are installed, use: 
    py -3.11 -m venv .venv
  • If only Python 3.11 is installed, use: 
    py -m venv .venv

5. Activate the Virtual Environment

  • Activate the virtual environment via: 
    .venv\Scripts\activate

6. Install Required Packages

  • Install the necessary packages with CUDA support into the new virtual environment: 
    pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121  
pip install numpy einops timm torchsummary

7. Set the Python Environment in MATLAB

  • In MATLAB console, specify the Python environment: 
    pyenv('Version', 'project_folder\.venv\Scripts\pythonw.exe')
  • The result should look like this: ```matlab ans =

PythonEnvironment with properties:

      Version: "3.11"
   Executable: "project_folder\.venv\Scripts\pythonw.exe"
      Library: "C:\Users\user\AppData\Local\Programs\Python\Python311\python311.dll"
         Home: "project_folder\.venv\Scripts\.venv"
       Status: NotLoaded
ExecutionMode: InProcess

8. Restart MATLAB (if necessary)

  • check if pyenv is set correctly in MATLAB console 
    pyenv

Troubleshooting

For any of the installation steps above, you may need to uninstall the previous installation. Sometimes PyTorch cannot make use of CUDA due to incompatibility.

First option - Uninstall

pip uninstall torch torchvision 
pip cache purge
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121

Second option - Remove Virtual Environment