deep_learning_cookbook icon indicating copy to clipboard operation
deep_learning_cookbook copied to clipboard

Published 20 hours ago • verified publisher icon DOsinga

Update ch03

Open ohtaman opened this issue 5 years ago • 4 comments

  1. Sort import statements
  2. Change keras to tensorflow.keras
    • Currentry, original Keras, not tf.keras, is not suitable for TensorFlow 2.0. So we want to use tf.keras everywhere.
  3. Add gzip module and remove some code to call zcat
    • I prefer standard module to external commands. But this change may not be necessary.
  4. Add note about a error caused by cache.
    • I'm not English native. Please point out if the grammar is wrong

ohtaman avatar May 20 '19 21:05 ohtaman

TensorFlow 2.0 and 1.0 are not compatible, so it may be better to have another branch instead of master. @DOsinga How do you think?

ohtaman avatar May 21 '19 20:05 ohtaman

It would be a good idea to move to 2.0, but yeah, it would need careful testing. It seems you are better at that then I am, so sure, why not. But maybe like you suggest we can put this together in a feature branch that we work on together?

DOsinga avatar May 29 '19 01:05 DOsinga

Any thoughts on how to best progress on this?

DOsinga avatar Jun 18 '19 12:06 DOsinga

Dear @DOsinga, are you still planning to update to TensorFlow 2.x? Or do you know any fork for which it was done yet? There are some readers (of the german translation) who ask for.

MarcusFra avatar Aug 03 '21 04:08 MarcusFra