hybrid-knowledge-base-chain

- on AWS

Sketch Flow

Demo

Current Features:

• Amazon Kendra Index as private Knowledge Base

• Private LLM on SageMaker Endpoint with LMI (Large Model Inference) container

• WebSearch as a chain Tool

• 'Hybrid Chain' - Internal Answer (Internal Knowledge) + web search => final answer

Developing:

• Larger model with better understanding of task prompts

• Retrieval similarity filtering

• Description-based Tool retrieving

References:

• https://aws.amazon.com/cn/blogs/machine-learning/quickly-build-high-accuracy-generative-ai-applications-on-enterprise-data-using-amazon-kendra-langchain-and-large-language-models/

• https://github.com/aws-samples/amazon-kendra-langchain-extensionsa

• https://github.com/imClumsyPanda/langchain-ChatGLM

• https://rapidapi.com/microsoft-azure-org-microsoft-cognitive-services/api/bing-web-search1/

Licensing

Inherit LICENSE from amazon-kendra-langchain-extensions