I created this repository to explore various search relevance techniques by developing tutorials 🏗️, creating Hugging Face 🤗 spaces and demos, and compiling useful resources 📚.
Quora Similar Questions Finder
: Hugging Face 🤗 Space. A demonstration of constructing a vector index database with Qdrant using questions from Quora. It then employs semantic search to identify the top K questions most similar to the user's input.Semantic Search on Simple Wikipedia Through Retrieval and Reranking
: Hugging Face 🤗 Space. Based on the user's input question or query, the algorithm searches through ~250k passages to find the top K relevant passages to the query, then reranks the results and outputs the top article with the relevant passage highlighted. Qdrant is used as a vector store, and SentenceTransformers are used for generating embeddings and reranking through cross-encoders. Finally, the demo is powered by Gradio.
Feel free to contribute by sharing your insights, improvements, or any useful resources related to search relevance.
Star or watch this repository for updates on new tutorials and resources. Your support helps this project grow.