Gemma-RAG

Retrieval Augmented Generation (RAG) using FAISS and Gemma 2b instruction tuned model from HuggingFace for answering your questions regarding movies.

Details

• Dataset: "MongoDB/embedded_movies"
• Embedding: "thenlper/gte-small"
• VectorDB: FAISS
• Generator LLM: "google/gemma-2b-it"

Libraries

• accelerate
• bitsandbytes
• transformers
• sentence-transformers
• faiss-gpu
• datasets
• langchain