About the `init_saver` of base_model?
imhuay opened this issue · 1 comments
imhuay commented
Why don't direct use the implement in the annotation?
def init_saver(self):
# just copy the following line in your child class
# self.saver = tf.train.Saver(max_to_keep=self.config.max_to_keep)
raise NotImplementedError
MrGemy95 commented
as "tf.train.Saver" takes the current tensorflow variables and save them, so calling this before building the model will save nothing and will lead to a serious bug, so i preferred to force this function to be implemented in the child class.