This repository contains my project notebooks for the Deep Learning course.
- Implement a Two Layers Fully Connected Neural Network in Numpy by hand writing the Feedforward and Backpropagation functions.
- Train and evaluate models on MNIST data set, Comparing the influences of:
- Stochastic / Batch / Mini batch gradient descent
- different hidden layer size
- different learning rate
- sigmoid / softmax output
- Using Tensorboard to visualize the graph and training progress.
- Re-build the FCNN model using Tensorflow.
- Implement a LeNet-5 CNN model in Tensorflow
- Perform some optimizations to get more than 99% of accuracy on MNIST.
- Different Optimizer :Gradient Descent vs AdamOptimizer
- Dropout
- Build a Vanilla-RNN and a GRU by hand writing the formulas in Tensorflow.
- Classify user comments from IMDb, Amazon, and Yelp to two classification(Negative / Positive).