Lab 1: Introduction to PyTorch

📘 MIT Introduction to Deep Learning - 2025

Welcome to Lab 1 of the MIT Introduction to Deep Learning course. This lab introduces the basics of PyTorch, one of the most popular deep learning frameworks. It includes hands-on experience with tensors, neural networks, and automatic differentiation, setting the foundation for future deep learning labs and projects.

🧠 What You Will Learn

• Fundamentals of PyTorch tensors and operations

• Defining and manipulating tensors of different shapes and ranks

• Creating computation graphs and custom operations

• Building and testing neural networks using:

  • Manual layer implementation
  • nn.Sequential
  • Subclassing nn.Module

• Understanding and using automatic differentiation with autograd

• Implementing gradient descent from scratch

🔧 Technologies Used

• PyTorch
• NumPy
• Matplotlib
• Google Colab
• mitdeeplearning Python package

🚀 Getting Started

Run on Google Colab

You can open and run this notebook in Google Colab using the link below: 👉 Open in Colab

Local Setup (Optional)

If you prefer to run it locally:

1. Clone the repository:

   git clone https://github.com/MITDeepLearning/introtodeeplearning.git
   cd introtodeeplearning/lab1

2. Install dependencies:

   pip install torch mitdeeplearning matplotlib

3. Open the notebook:

   jupyter notebook PT_Part1_Intro.ipynb

📂 File Structure

• PT_Part1_Intro.ipynb: Main notebook for Lab 1 covering PyTorch basics.
• MIT Deep Learning Repo: Contains all labs, documentation, and utilities.

📌 Notes

• Throughout the notebook, you'll encounter TODO blocks where you are expected to fill in code to complete exercises.
• Make sure to read the comments and explanations carefully to understand the concepts being applied.

© License

This lab is part of the MIT Introduction to Deep Learning curriculum. It is licensed under the MIT License. Use and/or modification outside the scope of the course must reference:

© MIT Introduction to Deep Learning  
http://introtodeeplearning.com