We are building a chatbot framework from scratch, that uses simple mathematical principles to train on small amounts of data and connect these in simple but effective ways, and creatively interact with users. Thinking out of the box, if you will.
We originally intended using C, but we are switchin to rust. However, we acknowledge that numpy in python would probably be the easiest way to go about the task, we prefer compiled to interpreted (just some quirky preferences lol)
bAInellm is still in it's initial research and testing, and it has shown to be greatly fast and effective on very little amounts of data.