ToxyBorg/Hugging-Face-Hub-Langchain-Document-Embeddings

Implements a complete RAG (Retrieval-Augmented Generation) pipeline: documents are loaded from multiple formats, chunked with LangChain's CharacterTextSplitter, converted to embeddings via HuggingFace Hub, and indexed in a FAISS vector store for similarity search. The system combines retrieved documents with Google's Flan-UL2 language model to generate contextual answers, with configurable storage for chunks, embeddings, and vector indices across the workflow.

No commits in the last 6 months.