Sharing the learning along the way we been gathering to enable Azure OpenAI at enterprise scale in a secure manner. GPT-RAG core is a Retrieval-Augmented Generation pattern running in Azure, using Azure Cognitive Search for retrieval and Azure OpenAI large language models to power ChatGPT-style and Q&A experiences.
update to the latest version https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/whats-new?view=doc-intel-4.0.0&tabs=csharp#february-2024
changes:
gpt-rag https://github.com/Azure/GPT-RAG/blob/69d3e4b2645d349a651724d4c2a58a188547463a/infra/main.bicep#L104
gpt-rag-ingestion https://github.com/Azure/gpt-rag-ingestion/blob/2a63b1bc747fd207a7c93c7d62d48da70f784a45/README.md?plain=1#L28 https://github.com/Azure/gpt-rag-ingestion/blob/2a63b1bc747fd207a7c93c7d62d48da70f784a45/chunker/chunk_documents_docint.py#L22