Mini-ML-Projects

This repository contains a collection of mini machine-learning projects for educational and practice purposes. Each project is designed to demonstrate specific concepts and techniques in machine learning and deep learning. These were created by me during the Summer School on AI by MLR, DTU.

Projects:

1.) Playing With Images

Various image processing, manipulation techniques and applying various types of filters from scratch were demonstrated using Python, OpenCV, pyplot, and PIL.

2.) K-Nearest Neighbors (KNN)

A simple KNN classifier implemented from scratch to classify CIFAR10 images based on their neighbours using PyTorch.

3.) MNIST

Created networks(FNN, CNN, FNN+CNN, RNN, LSTM) to classify handwritten digits from the MNIST dataset using PyTorch

4.) Fake News Detection

A project to classify news articles as real or fake by RNN and LSTM using Natural Language Processing (NLP) techniques using PyTorch

5.) GPT

Implementation of Generative Pre-trained Transformer 2 (GPT) for text generation tasks from scratch implemented by learning from Andrej Karpathy's implementation.

6.) Image Captioning

An image captioning model that generates descriptive captions for Flickr 8k images using a combination of CNN and RNN/LSTM architectures.

7.) Vision Transformer

Implemented a Vision Transformer (ViT) from scratch and trained it on the CIFAR10 dataset. This project demonstrates the power of transformer-based architectures in computer vision tasks, particularly image classification.