Open erp12 opened 4 years ago
@erp12 This indeed seems to be a big limitation of the library/API.
Are you planning to work on this?
The last commit was 4 months ago, it seems that you're not planning to work on this and the other issues any time soon.
This project is not abandoned :)
I dedicate some of my free time to the development of PyshGP (and other projects) and I am currently the only contributor. We all have to choose how to best spend our time, and it isn't every day that this library reaches the top of my priorities.
That said, contributions are always welcome. If you would like to work on a pull request for this, I would be happy to assist you. I can provide more details on what parts of the codebase would need to change, if you are interested.
Evaluators should support custom loss functions to compare the actual and expected outputs of program.
The
.fit()
method of the estimator should support either:.fit(X, y, loss_fn)
whereloss_fn
is a custom loss function..fit(evaluator)
whereevaluator
is an instance of an evaluator to use in place of a dataset.