Introduction to Deep Learning with Tensorflow

Tutorials with Tensorflow implementations. Initially designed for ENSAI SID 2017.

A .ova file (Ubuntu 16.04 - ~ 3.5 Go) is available on a USB key where Python 2.7 and Tensorflow (0.12.0) are already installed. Go to VirtualBox and 'import a virtual environment' and select the .ova file.

Slides available on my website: "Intro to deep learning with Tensorflow"

Feel free to install directly Tensorflow on your laptop TensorFlow Installation Guide

Introduction

• Hello World (code)
• Basic Maths (code)

Understanding Stochastic Gradient Descent

• Linear Regression (exo) - (solution)
• Binary Classification (exo) - (solution)

MNIST

• Softmax (exo) - (solution)
• Mulilayer Perceptron (exo) - (solution)
• One Conv + Max Pool (exo) - (solution)
• Make your life easier => SLIM ;)
• LeNet (exo) - (solution)
• Autoencoder (exo) - (solution)
• Conv-Deconv Autoencoder (exo) - (solution)
• RNN (exo) - (solution)
• GAN (coming...)
• DCGAN (coming...)

The Nature Conservancy Fisheries Monitoring (Kaggle)

• Have a look at the README of this repository first
• Inception v3 Features Extraction (code)
• Classification from DeepFeatures (exo) - (softmax regression)

Frozen Lake

• Q-Learning with neural network (exo) - (solution)

Thanks

Code inspired by other great tutorials from @aymericdamien (SGD and MNIST parts), @awjuliani (RL part), @pkmital (conv-deconv autoencoder).
Here are my favorite deep learning blogs: colah.github.io, karpathy.github.io, wildml.com

Dependencies

tensorflow (>0.12.1)
numpy
matplotlib
gym (Frozen Lake only)

An update one the repo will be done soone (python 3.5 and Tensorflow 1.2)