Bayesian-Optimization
Bayesian-Optimization copied to clipboard
initial guess added to DOE
I wrote this in VSCode browser so haven't gotten to test yet.
Basically just allows you to pass an np.ndarray with initial guess values to first DoE creation.
Does not enforce any constraints, does not take away from already specified DoE size (but may do so if dups are found and dropped).
Ok so I did some debugging and got it to work with a list of values instead of numpy array because it seems like that's what the ask and tell use?
Works enough for my use case, but if you'd like to suggest something to add like the (currently commented out):
assert self.initial_guess.shape[0] == self.dim or self.initial_guess.shape[1] == self.dim assert all([isinstance(_, float) for _ in self.initial_guess[:, self.r_index].ravel()]) assert all([isinstance(_, int) for _ in self.initial_guess[:, self.i_index].ravel()]) assert all([isinstance(_, str) for _ in self.initial_guess[:, self.d_index].ravel()])
which i took from the warm data, but matching list
instead