# Powering your products with ChatGPT and your own data
The Chatbot Kickstarter is a starter repo to get you used to building basic a basic Chatbot using the ChatGPT API and your own knowledge base. The flow you're taken through was originally presented with [these slides](https://drive.google.com/file/d/1dB-RQhZC_Q1iAsHkNNdkqtxxXqYODFYy/view?usp=share_link), which may come in useful to refer to.
The Chatbot Kickstarter is a starter repo to get you used to building a basic Chatbot using the ChatGPT API and your own knowledge base. The flow you're taken through was originally presented with [these slides](https://drive.google.com/file/d/1dB-RQhZC_Q1iAsHkNNdkqtxxXqYODFYy/view?usp=share_link), which may come in useful to refer to.
This repo contains one notebook and two basic Streamlit apps:
- `powering_your_products_with_chatgpt_and_your_data.ipynb`: A notebook containing a step by step process of tokenising, chunking and embedding your data in a vector database, and building simple Q&A and Chatbot functionality on top.
@ -32,4 +32,4 @@ Once you've run the notebook and tried the two Streamlit apps, you should be in
- This is a simple starting point - if you hit issues deploying your use case you may need to tune (non-exhaustive list):
- The prompt and parameters for the model for it to answer accurately
- Your search to return more relevant results
- Your chunking/embedding approach to store the most relevant content effectively for retrieval
- Your chunking/embedding approach to store the most relevant content effectively for retrieval
Midori
1 year ago
committed by
GitHub