GPTYourData 1.5

This release overhauls the web interface giving it a much more professional look.
It also expands the new web interface to easily allow uploading files to the knowledge base without having to run Embed separately.
It now uses GPT3_5_Turbo_16K which has more context and is cheaper than DaVinci 003.

Bottom line,

• still very fast -around 5 seconds vs 2-5 minutes from localGPT or privateGPT,
• far more accurate in its answers than localGPT, privateGPT or GPT4All
• more professional looking
• and now even cheaper to maintain.

GPTYourData 1.1

This release adds a new web interface for querying.
It also improves the embeddings format to create far less files.

If you have embeddings created in version 1.0, please use the UpgradeEmbeddingsToV1.1 application.
Just run it and it will create a new Consolidated folder. Backup your embeddings and move the contents from Consolidated folder to Embeddings folder.

GPTYourData 1.0

Welcome to GPTYourData! This project is a simple but fast and effective C# vector search engine that utilizes OpenAI's GPT models for generating embeddings and answering questions based on local files.