How to Run Jupyter Notebooks in VS Code for the AI For Beginners Curriculum

You can run the Microsoft AI For Beginners curriculum notebooks in VS Code by cloning the repository, creating the Conda environment from environment.yml, and selecting the ai4beg interpreter to execute cells directly in the editor.

The Microsoft AI For Beginners curriculum is structured as a collection of Jupyter notebooks designed to teach core AI concepts through interactive code examples. VS Code serves as the recommended integrated development environment because it natively supports Python notebooks, automatically detects Conda environments defined in the repository, and allows you to edit and execute code without leaving the editor. This guide walks through the exact workflow documented in lessons/0-course-setup/how-to-run.md and AGENTS.md to configure your local environment.

Prerequisites

Before running notebooks, ensure you have VS Code and Conda (Anaconda or Miniconda) installed on your system. The curriculum dependencies are pinned in environment.yml at the repository root, specifying exact versions of TensorFlow, PyTorch, Keras, OpenCV, and other machine learning libraries required for reproducible execution.

Step-by-Step Setup Process

Clone the Repository and Create the Conda Environment

Start by cloning the repository and creating the isolated environment named ai4beg. According to the setup instructions in lessons/0-course-setup/how-to-run.md (lines 11-16), run the following commands:

git clone https://github.com/microsoft/ai-for-beginners
cd ai-for-beginners
conda env create --name ai4beg --file environment.yml
conda activate ai4beg

This creates a dedicated environment containing all necessary packages, preventing conflicts with your system Python installation.

Open the Repository in VS Code

Launch VS Code from the repository root to ensure the workspace context is properly established:

code .

Upon opening the folder, VS Code will detect the Python project structure. As noted in lessons/0-course-setup/how-to-run.md (lines 20-22), the editor will prompt you to install the Python extension if it is not already present. Install this extension to enable notebook support and interpreter management.

Select the Conda Interpreter

The Python extension automatically discovers the ai4beg environment created in the previous step. You must explicitly select this interpreter to ensure notebook kernels run with the correct packages:

1. Open the Command Palette (⇧ ⌘ P on macOS or Ctrl+Shift+P on Windows/Linux)
2. Type "Python: Select Interpreter"
3. Choose the ai4beg environment from the dropdown list

This step binds VS Code to the Conda environment, ensuring that imports like import tensorflow or import torch resolve to the versions specified in environment.yml.

Open and Execute Notebooks

Navigate to any lesson directory under lessons/ and open a .ipynb file. For example, open lessons/3-NeuralNetworks/03-Perceptron/Perceptron.ipynb to view the notebook UI within VS Code.

To execute code:

• Click on a code cell
• Press Shift + Enter to run the cell in the selected kernel
• View outputs (including plots and DataFrames) directly beneath the cell

You can step through the curriculum at your own pace, modifying parameters and observing results in real time.

VS Code Jupyter Features

Beyond basic cell execution, VS Code provides enhanced notebook functionality through the Jupyter extension. As implemented in the repository architecture described in AGENTS.md (lines 68-73), these features include:

• Variable Explorer: Inspect Python variables and their current values during execution
• Interactive Window: Convert code cells to an interactive Python session for experimentation
• Jupyter Lab Interface: Access the full Jupyter Lab UI within VS Code via the Jupyter extension, providing a file explorer and advanced markdown rendering
• Integrated Debugging: Set breakpoints in notebook cells and step through code execution

Summary

• The AI For Beginners curriculum uses Jupyter notebooks stored in the lessons/ directory, such as lessons/3-NeuralNetworks/03-Perceptron/Perceptron.ipynb.
• Create the ai4beg Conda environment using environment.yml to install pinned dependency versions of TensorFlow, PyTorch, and other libraries.
• VS Code automatically detects the environment when you select the interpreter via Python: Select Interpreter.
• Run cells using Shift + Enter to execute code directly in the editor with outputs rendered below each cell.
• The workflow is documented in lessons/0-course-setup/how-to-run.md and supported by the architecture outlined in AGENTS.md.

Frequently Asked Questions

Do I need to install Jupyter separately for VS Code?

No. When you install the Python extension in VS Code, it automatically installs the Jupyter extension as a dependency. The notebook interface renders natively within VS Code without requiring a separate Jupyter installation or browser window, though the underlying Jupyter protocol still operates through the selected Conda kernel.

Can I use a virtual environment instead of Conda?

While the curriculum officially supports Conda as documented in environment.yml, you can technically use venv or virtualenv if you manually install the packages listed in that file. However, using the Conda environment named ai4beg ensures exact version compatibility with the curriculum's tested configuration and is the recommended approach according to the setup guide.

How do I resolve "Kernel not found" errors in VS Code?

If VS Code cannot locate the ai4beg kernel, first ensure the environment is activated in your terminal (conda activate ai4beg). Then restart VS Code and reselect the interpreter via ⇧ ⌘ P → "Python: Select Interpreter" → ai4beg. If the issue persists, run python -m ipykernel install --user --name=ai4beg within the activated environment to explicitly register the kernel with Jupyter.

Is it possible to run these notebooks in JupyterLab instead of VS Code?

Yes. The curriculum notebooks are standard .ipynb files compatible with any Jupyter client. You can launch JupyterLab from the activated ai4beg environment by running jupyter lab in the terminal. However, the curriculum authors recommend VS Code for its integrated editing and debugging features, as highlighted in the repository's agent documentation.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}