tf.data.Dataset: raise TypeError('{!r} is not a callable object'.format(obj)) -> caused-CodePudding

Home > OS > tf.data.Dataset: raise TypeError('{!r} is not a callable object'.format(obj)) -> caused

tf.data.Dataset: raise TypeError('{!r} is not a callable object'.format(obj)) -> caused

Time：06-20

I think that the error is originated by the wrong use of the .batch method (I don't know why it doesn't make the image shape = (128, 32, 32, 3) ) or by the wrong definition of the estimator.

In the original code (taken from this repository

CodePudding user response：

The solution to the wrong shape of the dataset (being (?, 32, 32, 3) instead of (128, 32, 32, 3) ) is to add drop_remainder=True to the args of the batch() method.

This seems to solve the problem.

Page link：https//www.codepudding.com/os/447759.html

Prev:Entity Framework multi level inheritance using fluent

Next:Calculate gradients of variables used in assignment of other variable using tf.GradientTape

Tags：

python

tensorflow

dataset

Links：
CodePudding