NanoSet is an experimental dataset where the main goal is to create a usable chatbot through less training data.
NanoSet is divded into 3 major sections, containg 36 entries divided into 6 sub-topics. The structure creates 108 total lines of training data, which may be subject to change in the future. The following is a visual on the structure:
108 entries total
3 Sections, each with 36 entries:
Chat Basics (Greetings, Jokes, etc.)… See the full description on the dataset page:
https://huggingface.co/datasets/srcworks-software/nanoset.
