Implementing a model that uses an LSTM to generate music using keras with tensorflow backend.At each step of sampling, this takes as input the activation and cell state from the previous state of the LSTM. and forward propagate by one step. Predicts output using softmax

Trigger Word Detection

Train a trigger word detection model and make predictions by first listening to the data, converting to spectrogram then fitting the model to output the time when the trigger word has been spoken.

Autonomous Driving using YOLO algorithm

Implementing the YOLO algorithm and non max supression technique to find out the objects seen by the dashboard of the vehicle using Convultional Neural Network.

Art generation with Neural Style Transfer

Using the convultional neural net to transfer style of a picture from a source picture to a destination picture.

Solutions to the Problemset to the CS229

Instructed by Andrew Ng, this contained my python solutions to the Machine learning class taught at Stanford.

Other Projects

Ranging from basic neural network from scratch to playing around with Tensorflow, scikit learn to projects such as:
Emojifier- converts your sentence by adding emojies using sentiment analysis, Facial Recognition, Machine Translation and many more. You can find in my attached git repo.