LittleLittleCloud
commented
2 months ago Thanks for creating this issue, can you describe the solution you want, a consumption pseduo code would be extremely helpful here.
The primary reason why AutoGen.Net lacks built-in support for RAG is because retrieval and augmentation are usually tightly coupled with application/using scenarios. For starter, retrieval is coupled with data source and augmentation is coupled with data scheme. This coupling creates a challenge in devising a RagAgent capable of satisfying all conditions across diverse applications and scenarios. So please tell us your ideal solution and maybe we can come up with a good solution to mitigate the challenge above
vdamov
ekzhu
MaxAkbar
Is your feature request related to a problem? Please describe.
Currently, AutoGen for .NET enables the development of conversational agents but lacks the ability to incorporate contextually rich, data-driven responses dynamically. This limitation becomes apparent in scenarios where agents need to retrieve and leverage specific information to formulate responses, affecting the quality and applicability of the agent's interactions in real-world applications. Adding RAG support would address these limitations by enhancing the system's ability to access and utilize external data sources effectively during conversations.
Describe the solution you'd like
This feature request proposes the integration of Retriever-Augmented Generation (RAG) capabilities into AutoGen for .NET. The addition of RAG would significantly enhance the ability of .NET-based conversational agents to retrieve and utilize contextually relevant information during interactions, thereby improving response accuracy and relevance. The implementation would expand the utility of AutoGen, making it a more robust tool in scenarios requiring dynamic information retrieval and usage.
Additional context
No response