I've set myself a little side project to train a customized GPT-3 model, and did all the pythonic cleaning-up of the data on my windows desktop.

But now that it's time to train the model, the OpenAI API is configured so that it can work with Windows, but only with great difficulty (mostly around managing my local environment variables and how well they play, or don't play, with Python on Windows) ... so I'm gonna do the training on my 2011 Linux laptop, where this just works