search

knakamura13 / mlrose-ky

A highly optimized fork of the popular mlrose-hiive package. For Machine Learning, Randomized Optimization and SEarch algorithms.
https://nkapila6.github.io/mlrose-ky/
BSD 3-Clause "New" or "Revised" License
20 stars 8 forks source link

Investigate GA loss #28

Open nkapila6 opened 1 month ago

nkapila6 commented 1 month ago

Issue Summary GA despite having high loss performs well on the validation set which is not expected.

Steps to Reproduce / Describe the Request

Expected Behavior / Desired Outcome With the loss that GA gives should be perform badly on the test set.

Proposed Solution (Optional) I'm assuming GA is reporting wrong loss values but arriving at the same weights.