Skip to main content
Version: v1.3.11

FAQs

Enterprise h2oGPTe is an AI-powered search assistant that helps you find answers to questions about your documents, websites, and workplace content - powered by H2O LLM.

  • With Enterprise h2oGPTe, finding what you need has never been easier
  • Ask a question, and Enterprise h2oGPTe will provide relevant results from across provided materials
  • Whether you're looking for a specific document or trying to understand a complex concept, Enterprise h2oGPTe can help you save time and effort

The following sections provide answers to frequently asked questions. If you have additional questions, please send them to cloud-feedback@h2o.ai.

General​

How does Enterprise h2oGPTe handle embedded images in PDFs?​

Enterprise h2oGPTe utilizes optical character recognition (OCR) to extract the text from the images.

Does Enterprise h2oGPTe have an API?​

Yes. To learn more, see Create an API key.

What is the Embedding Model for Enterprise h2oGPTe (1) and can I customize it (2)?​

  1. Enterprise h2oGPTe uses instructor-large as the embedding model.

  2. No. You cannot customize the default embedding model, but AI admins of the Enterprise h2oGPTe environment can select from the following alternative embedding models when setting up the Enterprise h2oGPTe environment: instructor-large , instructor-xl, bge-base-en-v1.5, or bge-large-en-v1.5. H2O.ai recommends using instructor-large as the default embedding model. End users cannot change the embedding model in the UI or Python API.

What is the Vector Database for Enterprise h2oGPTe (1) and can I customize it (2)?​

  1. The Vector Database for Enterprise h2oGPTe is called "Vex" and it is custom built by H2O.ai. Vex is a standalone vector "database" specifically designed for running similarity searches for Enterprise h2oGPTe. It serves as a storage and retrieval system for vector data.

    The current algorithm used in Vex is nmslib/hnswlib. Indexes are stored in hnswlib's native format and are created or loaded by the server on demand. This allows for efficient and fast similarity searches.

    Vex is architecturally similar to Chroma, which also utilizes hnswlib for thread-safety during index-writes. However, Vex includes several performance improvements compared to Chroma:

    1. It does not include unnecessary bindings to Clickhouse and DuckDB.
    2. It uses barebones Uvicorn (the core of FastAPI) instead of FastAPI, eliminating unnecessary overhead.
    3. It uses MsgPack instead of JSON for communication, reducing the bandwidth required for each request. This is particularly beneficial for compressing numpy arrays and UUIDs.
    4. It minimizes allocations in the query fast-path, avoiding unnecessary steps like Pydantic, and directly calling the library for binary deserialization.

  2. Customizability of Vex is limited, as it primarily focuses on delivering efficient similarity searches using its existing features. However, Enterprise h2oGPTe provides additional connectivity options to other vector databases, allowing for more customization possibilities.

    Starting from Enterprise h2oGPTe 1.3.8, users can connect Enterprise h2oGPTe to Elastic Search, Milvus, and Qdrant for storing vector data. Specifically, Enterprise h2oGPTe supports knn (k-nearest neighbors) search for Milvus and Qdrant, while it also supports full-text search for Elastic Search. Although these connectivities are covered by automated tests, they have not been thoroughly tested on real documents.

note

To learn how to customize the Vector Database, contact your admin.

Collections​

Can Enterprise h2oGPTe refer to several Collections to answer a query?​

Enterprise h2oGPTe cannot refer to several Collections when answering a query.

Documents​

What types of files can I use with Enterprise h2oGPTe beyond Documents?​

For a complete list of supported file types, see Supported File Types.

Chats​

Can I chat with a Collection in any language?​

Yes. Enterprise h2oGPTe uses the Multilingual-e5-large model for language translation, therefore any language supported by this model is also supported by Enterprise h2oGPTe. Enterprise h2oGPTe is set to use English by default.

To start chatting in any other language, edit the Chat settings Personality (System Prompt) to include the language you wish to use.

Personality (System Prompt): "I am Enterprise h2oGPTe, an intelligent retrieval-augmented GenAI system developed by H2O.ai. I speak Chinese."

Feedback