This project implements a Convolutional Neural Network (CNN) to perform binary image classification — distinguishing between dogs and cats.

This project implements a CycleGAN to translate natural photos into the artistic style of Claude Monet using the Kaggle GAN Getting Started competition.

Astraltech Bionic Implants is an ambitious RimWorld mod that introduces a new tier of advanced bionics.

This project aims to classify tweets as describing real disasters (1) or not (0) using natural language processing (NLP) techniques.

Deep learning pipeline for detecting metastatic cancer in small histopathology image patches.

This project builds a supervised regression model to predict individual medical insurance charges based on demographic and lifestyle features. Using the Medical Insurance Cost Dataset, we explore d…

A flexible and extensible PyTorch pretraining script built atop the timm library.

Lumenspark2 is the second-generation implementation of Lumenspark, a lightweight transformer for efficient large-scale language modeling.

A custom PyTorch implementation of EfficientNet with Squeeze-and-Excitation and MBConv blocks. Lightweight, modular, and easy to extend for research and experiments.

This project trains a Mask R-CNN with an EfficientViT-FPN backbone to segment brain MRI FLAIR abnormalities from the TCGA-LGG dataset. It includes data augmentation, custom anchors, warmup+exponent…

A diffusion model trained on FairFace with discrete conditioning on age, gender, and race. Generates high-quality face images using a custom UNet and DDPM scheduler.

A minimal PyTorch script for training an unconditional diffusion model on your own images. Includes safe image loading, mixed precision, and periodic sample generation.

A PyTorch image classifier using RegNetY and Albumentations on the Fashion MNIST dataset. Trains with TQDM progress, plots loss curves, and supports clean modular design.

Generate vivid, human-like captions for portrait images using the Qwen2.5-VL-7B model. Outputs dense descriptions covering emotion, posture, clothing, and environment.

Mouse Entropy is a CLI tool for capturing and storing raw mouse movement data on Arch-based Linux systems.

This repository contains a training script for a custom computer vision model using PyTorch Lightning. It use MLflow for robust experiment tracking.

A high-performance Rust utility that converts large image datasets into chunked Apache Arrow files for efficient storage and processing.

Lumenspark is a lightweight Linformer-based Language Model Trained from Scratch

RSNA 2024 Lumbar Spine Degenerative Classification Training Script

A custom implementation of a Convolutional Neural Network (CNN) from scratch in Rust

Pretraining the EfficientViT-B4 model on the ImageNet-1k dataset

Application designed to streamline the process of organizing and labeling images for image classification datasets.

A fast and flexible token bucket rate limiter for FastAPI applications.