François Chollet comments

Results 342 comments of


                                            François Chollet

Categorical_crossentropy() performance gain

One drive-by comment. If you care about speed, you should use a tf.function. Something like ``` import tensorflow as tf import numpy as np from tensorflow.keras.activations import softmax from tensorflow...

Adds show_trainable option to plot_model()

Thanks for the PR! Can you provide a few screenshots of what the plots look like?

Applications: Updated f-string methods in all possible scripts

The codebase has been reformatted previously. Thanks!

MAINT: Updated f-string method

The codebase has been reformatted previously. Thanks!

Updated f-string method

The codebase has been reformatted previously. Thanks!

PR for solving issue #16797

Sorry for the delay in reviewing this. > On the other hand, I thought about another option I think simply having the content of the current `def _update_trackables(self):` get called...

Migrating Evaluating and exporting scikit-learn metrics in a Keras callback example to Keras 3

> For other backends, validation_split was a problem. So, I tried with validation_data instead but still got an attribute error. What was the error in either case? Can you share...

Migrating Evaluating and exporting scikit-learn metrics in a Keras callback example to Keras 3

Thanks for the report. It looks like we only support np arrays and tf tensors for validation_split. This can be fixed easily, actually. Let me just do that in Keras.

Migrating Evaluating and exporting scikit-learn metrics in a Keras callback example to Keras 3

Done. You can retry after building Keras from sources.

Migrating Evaluating and exporting scikit-learn metrics in a Keras callback example to Keras 3

Just replace `.numpy()` with `ops.convert_to_numpy` which should work with any tensor.