Missing explanations

[Guitalex]

==> hidden_units = [8, 8]

Hello guys,

Is it possible to add explanations on the reasons for the choices for the hidden units values?
it's not a real issue because everything work perfectly but I don't understand why you had choosen these values..

why not these possibilities : [7, 7] [9, 9] [7, 7, 7] [8,8,8] [9, 9, 9] etc...
(theses exemples are totaly random, sorry if they look to be "dumb values". I'm learning alone, without school, in my house with nobody to help me)

THANKS YOU VERY MUCH!!!!!

Alex

[Guitalex]

OK, I asked my issue to ChatGPT-4 and he tell me that these values are most of the time selected after somes simulations.

My personnal answer is (I try...) : Keras tuner was probably used to choose these values and it's impossible to have 3 values (like my preview random exemple) because we configure a dimension of 2D (so only 2 values)

I hope that my answer is right..

THANKS Guys!!!!!

[SuryanarayanaY]

Hi @Guitalex ,

Could you please confirm which example/tutorial you are referring to ? Yeah sometimes some values work best based on lot of trails.Please provide more context to look into the issue.

Thanks!

[github-actions[bot]]

This issue is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you.

[iAlexMG]

hello @SuryanarayanaY !

for this example : https://keras.io/examples/keras_recipes/bayesian_neural_networks/

THX!

Alex

[github-actions[bot]]

This issue is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you.

[github-actions[bot]]

This issue was closed because it has been inactive for 28 days. Please reopen if you'd like to work on this further.

[SuryanarayanaY]

Hi @Guitalex ,

Thanks for pointing out to exact tutorial. I have gone through the tutorial and the hidden_units are nothing but hidden dense layers with each unit in hidden_units representing no of neurons. Yeah you can alter these values as per your requirement.You can take of size=3 of you want 3 hidden dense layers etc.

To demonstrate the same I tried with hidden_units = [10, 8, 4] and attached gist for results.

[SuryanarayanaY]

@Guitalex ,

I am not reopening the issue again as it is not an issue and my explanation above is enough to address the issue.

If you still have any query regarding this please let me know.

Thanks!