10.5K ·GitHub
What is this?
This repository collects tutorials, assignments, and competitions from MIT's Deep Learning related courses, covering semantic segmentation, generative adversarial networks, and reinforcement learning for self-driving cars.
This repository collects tutorials, assignments, and competitions from MIT's Deep Learning related courses, covering semantic segmentation, generative adversarial networks, and reinforcement learning for self-driving cars.Is this one unified course, or a set of separate tutorials and competitions?It bundles three independent pieces: a semantic segmentation tutorial using DeepLab, a GAN tutorial based on BigGAN, and the DeepTraffic deep reinforcement learning competition.
It bundles three independent pieces: a semantic segmentation tutorial using DeepLab, a GAN tutorial based on BigGAN, and the DeepTraffic deep reinforcement learning competition.How do I run the tutorials?Each tutorial links to a ready-made Jupyter Notebook as well as a Google Colab version, so you can run them directly in the browser without any local setup.
Each tutorial links to a ready-made Jupyter Notebook as well as a Google Colab version, so you can run them directly in the browser without any local setup.Do I need TensorFlow installed locally to try the tutorials?Not necessarily: the notebooks use TensorFlow-based models like DeepLab and BigGAN, but the linked Google Colab versions let you run them in the cloud without installing anything.
Not necessarily: the notebooks use TensorFlow-based models like DeepLab and BigGAN, but the linked Google Colab versions let you run them in the cloud without installing anything.What is DeepTraffic?DeepTraffic is a deep reinforcement learning competition where you design a neural network to drive one or more vehicles as fast as possible through dense highway traffic.
DeepTraffic is a deep reinforcement learning competition where you design a neural network to drive one or more vehicles as fast as possible through dense highway traffic.What does the segmentation tutorial show?It demonstrates semantic segmentation using the state-of-the-art DeepLab model on a sample video from the MIT Driving Scene Segmentation Dataset.
It demonstrates semantic segmentation using the state-of-the-art DeepLab model on a sample video from the MIT Driving Scene Segmentation Dataset.What does the GAN tutorial cover?It explores generative adversarial networks starting with BigGAN, a state-of-the-art conditional GAN architecture.
It explores generative adversarial networks starting with BigGAN, a state-of-the-art conditional GAN architecture.Is there research backing the DeepTraffic competition?Yes, DeepTraffic links to its own GitHub repo, website, and a published paper.
Yes, DeepTraffic links to its own GitHub repo, website, and a published paper.What are similar projects?DeepLearning, AirSim, d2l-en, ailearning, reinforcement-learning, TensorFlow-Examples, awesome-tensorflow, AI-For-Beginners
DeepLearning, AirSim, d2l-en, ailearning, reinforcement-learning, TensorFlow-Examples, awesome-tensorflow, AI-For-Beginners