spencerbertsch1/applied-deep-learning-engg192

Independent study focusing in various deep learning topics - Dartmouth College, Winter 2019

Deep-Learning

Independent study focusing in various deep learning topics - Dartmouth College, Winter 2019

Dependencies

The projects in this repository use a variety of deep learning labraries and python packages. In order to run most of the code, you will need to have both TensorFlow and PyTorch installed on your machine. Other important depedencies include Pandas, Numpy, and Matplotlib.

For MacOS

Simply open a command window and run the following commands to install the necessary libraries.

1. Run pip install --upgrade tensorflow to make sure your version of TensorFlow is up to date
2. Run pip install tflearn
3. If you're not using a virtual environment, Run: sudo pip install keras
   If you're using a virtual environment, Run: pip install keras
4. Run pip3 install torch torchvision
5. Run pip3 install Pandas
6. Run pip3 install numpy
7. Run pip3 matplotlib

Projects and Assignments

• Fully Connected Networks

  • NBA Winner Prediction - PyTorch
  • NBA Winner Prediction - T.F. Keras
  • Regularization - Mushroom Toxicity
  • XOR Example

• Optimization and Regularization

  • Mushroom Toxicity Example
  • Breast Histology Classification

• Convolutional Networks

  • MNIST Classification with 2D ConvNet
  • Breast Histology Classification with 3D ConvNet

• Recurrent Networks

  • Sine Wave Prediction
  • Naive Stock Market Data Prediction
  • Slide Deck - Vanishing and Exploding Gradients

• Autoencoders

  • Simple MNIST Autoendocer

• Final Project - Optimizing Seq2seq Translator

  • Optimization notebook
  • Final slide deck
  • Final Paper

• Just for fun!

  • TensorFlow.js Model Approximation