Today I created a sample project called SKWinterOlympics.
The sample uses Microsoft Semantic Kernel and it includes a C# console app that lets users ask any question related to the 2022 Winter Olympics to ChatGPT.
💡
Why is this interesting you might ask: it is because ChatGPT does not include any information about 2022 Winter Olympics and such data has to be loaded from local storage before we can ask our questions :)
The sample is a port of one of the Python samples in the openai-cookbook and it does the following:
- It loads a big file (213Mb) with Wikipedia information about the 2022 Winter Olympics (the file is located here if you are interested).
- It indexes the data in a local (in-memory) semantic index (IMemoryStore, in SK's terminology).
- It asks the user to ask a questions about the 2022 Winter Olympics until the application is terminated.
This is how the application behaves:
If you are interested in learning more about this sample and see the source code, please check the SKWinterOlympics repository and maybe add a Star to the repo while you're there :)
