🧠 FAISS Flashcards
🔍 What is FAISS?
FAISS (Facebook AI Similarity Search) is a library for efficient similarity search and clustering of dense vectors.
⚡ What is FAISS used for?
FAISS is commonly used in vector search engines, recommendation systems, and LLM-based retrieval pipelines.
🧠 What types of indexes does FAISS support?
FAISS supports flat (brute-force), IVF, HNSW, and PQ indexes for approximate nearest neighbor search.
📚 How do you install FAISS?
You can install FAISS using pip (`pip install faiss-cpu`) or build from source for GPU support.
⚙️ Does FAISS support GPU?
Yes, FAISS has a CUDA version for GPU-accelerated indexing and search, ideal for large datasets.
💡 What’s the advantage of FAISS?
It provides high-speed nearest neighbor search with minimal memory usage, even on millions of vectors.
🔁 Can FAISS be used with LangChain?
Yes. FAISS is one of the supported vector stores in LangChain for Retrieval-Augmented Generation (RAG).
🧪 Does FAISS support clustering?
Yes, FAISS supports k-means clustering, making it suitable for embedding analysis and compression.
🧰 How does FAISS compare to Pinecone?
FAISS is a self-hosted library, while Pinecone is a managed service. FAISS offers more control but requires setup.
🌐 Can FAISS be used at scale?
Yes, with GPU acceleration and advanced indexing strategies, FAISS can handle billions of vectors efficiently.