An autonomous car model using a regression CNN to complete laps in a virtual simulator. Conducted data collection, augmentation, and pre-processing of frames and steering angle to maintain the car's position on the road, including sharp turns.

Code

The project is an interactive web tool designed to help users witness and understand the working of various search algorithms. It allows users to set a starting point, a goal, and obstacles along the path. By doing so, they can watch the algorithms navigate their way to the goal and compare their performance firsthand. The tool provides a hands-on learning experience that enables users to gain a deeper understanding of how search algorithms function in real-world scenarios.

Code

A GAN model on AWS cloud infrastructure to generate anime faces using a discriminator and generative model with 128-bit latent input. Used a Kaggle anime face dataset for the project.

Code

The project serves as an example of how Natural Language Processing models can be used for generating new names or words that adhere to a specific naming convention. It involves using a Seq2Seq model architecture for Natural Language Processing (NLP) based on LSTM (RNN). The model is trained on a dataset of 2000 names using Python and Torch. A sampler was designed to generate 20 names, starting with any alphabet, using the trained model.

Code

This project involved building a multi-class classification model using Convolutional Neural Network (CNN) written in Torch and Python. The model was trained on a dataset of 90,000 images, each with dimensions of 200x200 pixels, and containing 9 classes of different shapes. The model achieved an impressive accuracy rate of over 98%

Code

This project aimed to build a Machine Learning (ML) model that could recognize handwritten digits from 0-9 using Convolutional Neural Networks (CNN) in Keras and Python. The model achieved an impressive accuracy rate of over 99.9% on the well-known MNIST digit dataset. To make the model more accessible, a user-friendly web application was developed using Flask container.

Code