Creating Document Collections in h2oGPTe for RAG

This lesson walks through creating a document collection in Enterprise h2oGPTe to ground answers in your own files instead of relying on the model's general knowledge. It covers naming a new collection, configuring settings like the embedding model and chunk size, and uploading a PDF so it can be chunked, embedded, and indexed in a vector database. The lesson also covers sharing a collection privately, publicly, or with specific users, and explains that this setup enables retrieval-augmented generation so chats can draw on the uploaded source.