My solution to DeepTraffic competition organized by MIT.

My deep learning architecture attaining 72.5 mph of max 80 mph using Deep Q-learning, Convnet.js popular convolutional neural network library - in DeepTraffic competition organized by MIT course 6.S094: Deep Learning for Self-Driving Cars

Various architectures attaining different speeds are uploaded to git.