Current GitHub adoption and maintenance signals.
This project implements a GPT model from scratch using PyTorch, covering tokenization, attention mechanisms, transformer blocks, and text generation. It demonstrates pretraining, finetuning for classification, and instruction finetuning, including data handling and evaluation techniques.
The complete text of the Harry Potter saga (Books 1-7) in one file. A dataset for NLP, text analysis, and data exploration. Note: This material is copyrighted and is provided for educational and non-commercial use.
Created_Neural_Networks_From_Scratch demonstrates building neural networks step by step without using high-level libraries. It covers core concepts like forward propagation, backpropagation, and training, providing a foundational understanding of deep learning mechanics.
🦀 Low-level 3D Computer Vision library in Rust
GSoC 2026: Real-time GPU-accelerated Bird's-Eye View (BEV) camera pipeline built on Bubbaloop and kornia-rs
High-performance, zero-copy Rust inference backend for Vision-Language Models (VLMs). Designed for edge deployment utilizing ONNX Runtime and hardware-accelerated execution providers.
A comprehensive Rust implementation of Low-Level Design (LLD), System Architecture, and Design Patterns. mastering object-oriented principles and scalable system design using idiomatic Rust.
A high-performance real-time chat application using a hybrid architecture. Features a Django Rest Framework backend for Auth/Admin, a Go (Golang) server for high-concurrency WebSockets, and a Next.js (TypeScript) frontend. Data is synced via a shared PostgreSQL database.
An AI-powered interview coach built with Streamlit and the Groq API for blazing-fast responses. Practice for behavioral interviews by customizing your role, experience level, and question difficulty. Get instant, detailed feedback on your performance to sharpen your skills and land your dream job.
The AI-Powered Recommendation Engine is a backend project that demonstrates how to build a movie recommendation system using Django, Django REST Framework, and machine learning. The system leverages collaborative filtering techniques to analyze user ratings and movie data, providing personalized recommendations for users.
No repository description provided.
No repository description provided.
A lightweight, command-line string search utility built with Rust. This project implements core grep functionality, featuring efficient file reading, case-sensitive/insensitive search modes, and robust error handling.
This repository contains my implementation and experiments with multi-head attention mechanisms. Here, I have explored how multi-head attention works, tested different configurations, and provided code samples, explanations, and results. The project includes detailed documentation and example use cases.
This notebook covers text tokenization for LLMs, exploring simple tokenizers, handling unknown words, and BPE. It demonstrates creating input-target pairs and implementing a PyTorch data loader for training, including token and positional embeddings.
This code merges token embeddings from the Byte Pair Encoding (BPE) tokenizer with positional embeddings. This allows the model to understand both word meaning (including subwords via BPE) and word order, forming an embedding matrix of shape [batch_size, max_length, output_dim].
No repository description provided.
A rule-based sentiment analysis project implemented in Jupyter Notebook. This repository demonstrates how to analyze text sentiment using predefined rules and heuristics, making it easy to understand and customize for different sentiment analysis tasks.
Explore various chatbot and LLM applications powered by the high-speed Groq API. This repo features Python examples for text generation, streaming responses, keyword extraction, and a complete Retrieval-Augmented Generation (RAG) pipeline built with LangChain, ChromaDB, and Llama3.
📰 Fake News Classification using NLP This project focuses on identifying and classifying fake news articles using Natural Language Processing (NLP) techniques. The goal is to build a reliable model that can differentiate between real and fake news based on textual data such as article titles or content.
NLP_Fundamental_&_Text_Preprocessing: This repository covers the foundational concepts of Natural Language Processing (NLP), including key techniques for text preprocessing such as tokenization, stopword removal, stemming, lemmatization, and vectorization, essential for preparing raw text data for machine learning models.
No repository description provided.
🚀 Intro to Computer Vision with TensorFlow This repository contains introductory examples and notebooks demonstrating basic computer vision concepts using TensorFlow and OpenCV. It covers reading, processing, and visualizing images — laying the foundation for building deep learning models for computer vision tasks.
No repository description provided.
No repository description provided.
No repository description provided.
E-Commerce Backend is a robust solution for developing the backend infrastructure of an e-commerce application. It supports features such as user authentication, product catalog management, order processing, and payment gateway integration, catering to the needs of scalable and efficient online shopping platforms.
No repository description provided.
A Python implementation of the Caesar Cipher to encrypt and decrypt messages by shifting letters in the alphabet.
"Django-based movie recommender web app using machine learning for intelligent suggestions, with both API and static frontend rendering."
🔐 Password Generator using Python A simple yet powerful Password Generator built using Python. This script allows users to generate secure, random passwords with a mix of letters, numbers, and symbols. It supports custom password lengths and ensures each password is strong and unpredictable. ✅ Features: User-defined password length
No repository description provided.
No repository description provided.
No repository description provided.