Bypassing of local minima crashes if best model iteration is the base iteration

bdzyubak / tensorflow-sandbox

A repository for studying applications of Deep Learning across fields, and demonstrating samples of my code and project managment

0 stars 0 forks source link

Bypassing of local minima crashes if best model iteration is the base iteration #20

Closed bdzyubak closed 2 years ago

bdzyubak commented 2 years ago

The way it was originally coded, if the best model obtained by spiking the gradient and retraining matches the original, it will try to copy over itself and fail. Do a quick patch to check match with self and take no action.

bdzyubak commented 2 years ago

Patched in fbbc3a21aa381d49701240b7ec7000acb168f307.