ML-and-DL Projects

Code for personal projects pertaining to machine and deep learning that I've completed while learning the required libraries.

Files:

1. MNIST using Tensorflow with custom pipeline using openCV2 to classify new images.
2. Titanic (from kaggle) using XGBoost.
3. Sentiment Analysis of tweets using Scikit Learn. Data from competition on Kaggle.
4. MNIST using a CNN in keras, tried to improve performance over a regular FFNN by adjusting the model complexity.
5. Text-generation using RNNs : Used the notebook provided at https://colab.research.google.com/github/tensorflow/docs/blob/master/site/en/tutorials/text/text_generation.ipynb; fit model to generate dialogues/screenplay for a Rick and Morty episode, based on screenplays from Rick and Morty S4.
6. DCGAN model to generate images of Pokemon from noise. Refernces and data sources in notebook.
7. AC-WGAN model to generate microstrucutres from gaussian noise. References in notebook. (Not clean code)
8. TechSoc: Deepfake detector, using a pretrained VGG16 as base. Placed 3nd on Leaderboard. (https://www.kaggle.com/c/dl-hack-track-1-cv)
9. TechSoc: Open ended conditional text generation of abstracts of research papers from Arxiv.org given the titles, using a pretrained GPT-2 as base. Placed 2nd on leaderboard. (https://www.kaggle.com/c/dl-hack-track-2-nlp)