How lexfridman/mit-deep-learning works
This repository is a collection of hands-on MIT deep-learning tutorials rather than an application runtime. Learners open notebooks and follow demonstrations: the basics notebook covers feed-forward and convolutional networks; the driving tutorial applies semantic segmentation to sample video; and the GAN tutorial explores BigGAN. Earlier notebooks cover introductory perceptrons, TensorFlow, MNIST, and traffic-light classification. The README also describes DeepTraffic, but identifies it as a separate competition project. The tree and README support these relationships; notebook internals were not available for inspection.
Main entry points: tutorial_driving_scene_segmentation.ipynb Scene segmentation · 5_tensorflow_traffic_light_classification.ipynb Traffic-light classification · mit_driveseg_sample.mp4 Sample driving video · 5_tensorflow_traffic_light_images Traffic-light images
Components 12
Core tutorials
- Deep learning basics
tutorial_deep_learning_basics/deep_learning_basics.ipynb - Scene segmentation
tutorial_driving_scene_segmentation/tutorial_driving_scene_segmentation.ipynb - Sample driving video
tutorial_driving_scene_segmentation/mit_driveseg_sample.mp4 - GANs and BigGAN
tutorial_gans/tutorial_gans.ipynb
Earlier tutorials
- Python perceptron
tutorials_previous/1_python_perceptron.ipynb - TensorFlow intro
tutorials_previous/2_tensorflow_intro.ipynb - Simple neural network
tutorials_previous/3_tensorflow_simple_neural_network.ipynb - TensorFlow MNIST
tutorials_previous/4_tensorflow_mnist.ipynb - Traffic-light classification
tutorials_previous/5_tensorflow_traffic_light_classification.ipynb - Traffic-light images
tutorials_previous/5_tensorflow_traffic_light_images
Competition
- DeepTraffic competition
Other components
- Learner
Connections 11
- Learner → Deep learning basics: follows
- Learner → Scene segmentation: follows
- Learner → GANs and BigGAN: follows
- Sample driving video → Scene segmentation: input video
- Learner → Python perceptron: follows
- Learner → TensorFlow intro: follows
- Learner → Simple neural network: follows
- Learner → TensorFlow MNIST: follows
- Learner → Traffic-light classification: follows
- Traffic-light images → Traffic-light classification: image data
- Learner → DeepTraffic competition: competes in