How can I train and use a fine-tuning model with CUDA GPU?

The fine-tuning isn’t happening on your computer, it’s happening on OpenAI servers and just displaying the progress. So no control over hardware/speed of process.