https://github.com/adammikulis/EdgeRag

EdgeRag is a program that runs large language models and vector databases on your local device
https://github.com/adammikulis/EdgeRag

Last synced: 9 months ago
JSON representation

EdgeRag is a program that runs large language models and vector databases on your local device

• Host: GitHub
• URL: https://github.com/adammikulis/EdgeRag
• Owner: adammikulis
• License: mit
• Created: 2024-02-26T19:41:51.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-03-07T21:48:14.000Z (over 2 years ago)
• Last Synced: 2024-03-08T00:42:04.508Z (over 2 years ago)
• Language: C#
• Homepage:
• Size: 992 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• StarryDivineSky - adammikulis/EdgeRag

README

          
# EdgeRag

EdgeRag is a Retrieval Augmented Generative (RAG) Artificial Intelligence (AI) pipeline that is based on [LLamaSharp](https://github.com/SciSharp/LLamaSharp) and Microsoft's [Kernel Memory](https://github.com/microsoft/kernel-memory). It uses Godot 4.2.1 mono and requires .NET8. You can run many popular models (llama, mistral, mixtral, phi) and add nearly any sort of document to the vector database to query. 

Download a Mistral model from here: https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF

Download a Phi model from here: https://huggingface.co/TheBloke/phi-2-GGUF