Setup

Before you begin, make sure you have WSL, Docker, and NVIDIA drivers installed.

1. Follow the instructions here to install WSL.

2. Follow the instructions here to install Docker Desktop for Windows.

3. Follow the instructions here to install NVIDIA drivers.

Usage

Now we're ready to setup the Jupyter Server in the Docker container using WSL.

1. Start by cloning this repository to your local machine.

   # Clone this repository
   git clone https://github.com/Auxority/tensorflow-jupyter.git

2. Open your favorite terminal and run wsl install to create a new WSL instance. Just follow the prompts and you should be good to go.

   # Create a new WSL instance
   wsl install

   Note: If you run into any issues, you can use wsl --shutdown followed by wsl -l and wsl --unregister <distro> to shutdown, list, and delete WSL instances respectively.

3. Once you are in your WSL instance, navigate to this repository. Make sure to replace the path with the path to your repository. You can use the /mnt/c prefix to access files on your C drive.

   # Navigate to the repository
   cd /mnt/c/Users/<username>/Documents/GitHub/tensorflow-jupyter
   

4. Run make init to initialize your .env file.

   # Initialize your .env file
   make init

5. Open the .env file you just created and configure your token there. You will need this token later.

6. Run make run to start the Docker container. This will take a while the first time you run it, but subsequent runs will be much faster.

   # Start the Docker container
   make run

7. Once the container is running, you can access Jupyter Notebook by navigating to http://localhost:8888?token={YOUR_TOKEN} in your browser. Make sure to replace {YOUR_TOKEN} with the actual token you set in your .env file during step 5.

8. You can now start coding! You can find some example notebooks in the notebooks folder.