An LLM from scratch in pure C++/CUDA

An automatic differentiation engine in C++.

This project aims to develop a machine learning-based system for classifying chest X-ray images into two categories: pneumonia and normal. The system will be trained on a dataset of labeled chest X…

A program to control the speed of a 5V DC motor using the AVR ATMega162 IC and 2 input sources, a keypad and an ADC.

Broadcasting 3D torch tensors using CUDA(this functionality is already available in pytorch)

Code for a decoder only multimodal transformer model which uses token fusion to enhance inter-modal token representations.

Just a decoder only architecture with mixture of depths

Fine tuning Llama-3-8b using SFT. Unfortunately, the trainer does not allow 8 bit cpu offloading while training, so my 4gb vram is prolly not enough. Try training it in colab.

Medical chatbot

repository dedicated to data science and ML projects

A repository dedicated to data preparation and eventual bacteria detection model training.

Solutions for leetcode problems.

A dataset for electric components, specially for object detection.

Repository for bots

An entire management system for hotels

repository dedicated to data science and visulation tasks

Quantum computation algorithms

PCA and clustering on the given datasets

A very simple flask app which uses 2 models to predict the named entities within a CV

visualization of geospatial data