Introduction to SimpleRAG
SimpleRAG is based on WPF and Semantic Kernel implementation of a simple RAG application , can be used to learn and understand how to use Semantic Kernel to build RAG applications .
GitHub Address:/Ming-jiayou/SimpleRAG
Key Features
AI chat
Supports all large language models compatible with the OpenAI format:
Text embedding
Supports all embedded models compatible with the OpenAI format:
Simple RAG Answers
Simple RAG answer effect:
Contrast this with responses that do not use a RAG:
Build from source
git clone to local and open the file:
As shown below:
ChatAI is used to configure the dialog model, Embedding is used to configure the embedding model, and TextChunker is used to configure the document slice size.
Still using SiliconCloud as an example, just fill in your api key and change the filename to, or create a new one.
The configuration is completed as shown below:
IDE:VS2022
NET version: .NET 8
Open the solution and the project structure is shown below:
Run the program:
Test AI Chat:
Test Embedding:
The vectors are saved using Sqlite, which can be found in the Debug folder:
Open this database as shown below:
Test RAG Answer:
Other configurations
You are also free to make other configurations, such as using dialog models in Ollama with embedded models for local offline scenarios, configuring other online dialog models, using embedded models in local Ollama, and so on.
ultimate
If it helps you, a star✨ is the biggest support 😊.
If you've read this guide and are still experiencing problems, feel free to contact me through the public: