Based on relevant company documents provided we can use LLMs to generate answers for questions in RFI documents.

make clean
make setup
make run

export H2OGPTE_API_TOKEN="" 

export H2OGPTE_URL="https://**.h2ogpte.h2o.ai"

h2o secret create h2o-rfi-qa -l h2ogpte-key="" -l h2ogpte-url="https://***.h2ogpte.h2o.ai"

• Default collection name is "H2O_DEMO_RFI" to be used in H2O Enterprise GPT
• Default doc URLs ingested as the part of above collection are:
  • https://h2o.ai/blog/2022/the-bond-market-ai-how-marketaxess-brings-it-all-together/
  • https://h2o.ai/platform/ai-cloud/managed/
  • https://docs.h2o.ai/haic-documentation/overview/what-is-h2o-ai-cloud/
  • https://docs.h2o.ai/h2o-llmstudio/

This feature enables users to view the current list of documents in a collection that LLMs use to answer RFI queries. Disabled for demo!

This feature automates the RFI answering process. Users can upload an Excel file with a mandatory "Question" field. The LLM models use this to provide answers, which appear in an editable table for user review. Users can modify the answers in the table before downloading them.

• Sample URL: https://raw.githubusercontent.com/narasimhard/demo-data/main/RFIQueriesLLM.xlsx

This feature allows users to interact directly with LLM models.