InTowards AIbyTapan Babbar·Aug 4Building an AI Race Engineer (Part 1): From Race Replay to Crew ChiefPhases 0 to 2 of shipping a measurable F1 strategy agent: world model, baseline ML, then structured LLM decision making
InTowards AIbyTapan Babbar·Dec 28, 2024Organise Photo Dumps With AI: Face Recognition & Reverse Image SearchThis project showcases the use of YuNet, VGG-16, and FAISS for efficient photo recognition and reverse image search
InTowards AIbyTapan Babbar·Nov 15, 2024Build the Smallest LLM From Scratch With PyTorch (And Generate Pokémon Names!)This project builds a simple character-level model to generate Pokémon-inspired names by predicting one character at a time.A response icon1A response icon1
InTowards AIbyTapan Babbar·Oct 27, 2024Enhance OCR with Llama 3.2-Vision using OllamaThis project upgrades book cover recognition by using Llama 3.2-Vision to seamlessly extract and complete titles and author names from…A response icon5A response icon5
InTowards AIbyTapan Babbar·Oct 25, 2024Speech-To-SQL Using OpenAI’s Whisper and Ollama (Llama3)This project uses Whisper for speech-to-text and Llama 3 for converting transcriptions into SQL queries, enabling automatic database…A response icon1A response icon1
Tapan Babbar·Oct 22, 2024How to Run Llama 3.2-Vision Locally With Ollama: A Game Changer for Edge AIA quick guide to running llama 3.2-vision locally using Ollama with a hands-on demo.A response icon5A response icon5
Tapan Babbar·Oct 18, 2024Build an AI Image Similarity Search with Transformers — ViT, CLIP, DINO-v2, and BLIP-2This project uses vision models to generate image embeddings and performs similarity searches with FAISS.A response icon8A response icon8
Tapan Babbar·Oct 2, 2024Enhance OCR with Custom Yolov10 & Ollama (Llama 3.1)This project enhances text recognition workflows by integrating a custom YOLOv10 model with EasyOCR and refining the results with an LLMA response icon6A response icon6
InInsiderFinance WirebyTapan Babbar·Sep 24, 2024Real-Time AI Stock Advisor with Ollama (Llama 3) & StreamlitThis project uses LLM-powered insights to fetch stock data every minute, and provide real-time, easy-to-understand explanationsA response icon5A response icon5
Tapan Babbar·Sep 9, 2024Build an AI Bartender with RAG, pgvector, Ollama & StreamlitWith Chroma vector databases and LangChain’s similarity search dominating the RAG space, I wanted to try a simpler approach using SQL…