A way to evaluate a gpt model for how well it can improve AI/ML code. It will iterate and make changes, and execute based on observations it has made about previous runs.

It may be necessary to use a unique dataset other than the breast cancer for this problem in order to avoid solving it via training data, but I use it here as an example.

Working on similar looping solutions, follow me on twitter here to stay up on the latest - https://twitter.com/QRDL