MITDeepLearning/introtodeeplearning
Introtodeeplearning
This repository contains the lab materials and Jupyter notebooks for MIT's introductory deep learning course, using TensorFlow and Keras for hands-on exercises. The courseware is delivered as pre-configured notebooks that run on Google Colaboratory's cloud infrastructure, eliminating the need for local software installation.
Learners can toggle the Colab runtime to a GPU-backed hardware accelerator for faster neural network training during lab exercises. A shared Python package provides helper functions that standardize common operations across all notebooks. The course guides students through a defined sequence of steps to format and submit completed lab work for course-hosted deep learning competitions.
Features
- Deep Learning Courses - Provides hands-on deep learning exercises in a structured course setting with pre-configured notebooks and automated grading support.
- Deep Learning Labs - Opens pre-configured Jupyter notebooks in Google Colaboratory so learners can complete exercises without local setup.
- Managed Cloud Notebooks - Opens and runs Jupyter notebooks in Google Colaboratory without local setup, enabling remote deep learning practice.
- Cloud Execution Environments - Provides pre-configured Jupyter notebooks that run entirely on Google Colaboratory's cloud infrastructure.
- Jupyter Notebook Curricula - Delivers a structured deep learning curriculum as pre-configured Jupyter notebooks with embedded exercises.
- GPU Acceleration - Lets learners toggle the Colab runtime to a GPU hardware accelerator for faster neural network training.
- GPU-Accelerated Training - Switches Colab runtimes to GPU hardware for faster neural network training during lab exercises.
- Course Competition Submissions - Guides learners through a defined sequence of steps to format and submit completed lab work for course competitions.
