Sideline projects I built outside of my day job.
An image-to-image translation model using Conditional GANs that takes a black and white anime sketch and gives you a colored version of it.
Trained on the Anime Sketch-Colorization Pair dataset from Kaggle, which has 14.2k sketch-color pairs. I wrote the full walkthrough as part two of my GANs series.
View project →
A bot for helping reduce toxicity levels on Discord servers to zero.
Built with discord.js and TensorFlow.js for a Microsoft Learn Student Ambassadors session, and deployed on Azure App Services with continuous WebJobs so it stays up.
View project →
A Streamlit web app that hosts a housing price prediction model you can tweak on the web itself. You can change the number of neurons, the learning rate and the epochs.
It's an exploratory tutorial on the California Housing Price dataset, with the model built using TensorFlow in Python.
A deep learning based web platform for cervical cancer detection and patient-doctor interaction.
We achieved top in the class accuracy in classification of cervical cells in Pap smear images using transfer learning. Built with TensorFlow and deployed on GCP.