AttributeError: module 'keras._tf_keras.keras.layers' has no attribute 'experimental'

[Shanyageo]

No description provided.

[kayleedirecto]

Hi there! Did you end up finding a solution to this?

[davemaster]

Hi there! Did you end up finding a solution to this?

You have to downgrade TensorFlow to 2.10 if I remember

[kayleedirecto]

Do you remember what version of Python you were running?

[sabbir2609]

3.10.6 or 3.9.6 should work

[davemaster]

Do you remember what version of Python you were running?

Python 3.9.13

and here my packets installed
myrequirements.txt

[davemaster]

SINCERELY, I'm questioning this python way to do Machine Learning, EVERY TIME, those packets needed to do so, every time, there are a lot of problems of compatibility, a MESS DISASTER what really, REALLY WASTE MORE TIME than that used to train a model... MUST BE a better way to sync these hundred of packets, or at least for 6 months, NO CHANGES at all.

[kayleedirecto]

I know, it's really frustrating :/ Were you able to get all the packages installed with that version combination?

[sabbir2609]

SINCERELY, I'm questioning this python way to do Machine Learning, EVERY TIME, those packets needed to do so, every time, there are a lot of problems of compatibility, a MESS DISASTER what really, REALLY WASTE MORE TIME than that used to train a model... MUST BE a better way to sync these hundred of packets, or at least for 6 months, NO CHANGES at all.

Use Anaconda

[davemaster]

Anaconda, another stuff; then what, tiger, lion, rhino, spider.....

better do this using c++ (rules, forever) or c#, and no more wasting time with volatile packages ...

Good luck, AND THANKS anyway...