Confusing example in documentation

[rishijain07]

This example is quite confusing. It is more complex than traditional logistic regression. In a true logistic regression, there would typically be a single output neuron with a sigmoid activation for binary classification, or multiple output neurons without a hidden layer for multi-class problems. It might be better to replace this example with something else, or clarify the documentation by adding "input layer with 32 neurons connected with 16 fully connected neurons (dense layer)."

[heydaari]

Hello there
Keras Input Layers dont have any learnable parameters , so you can not call them "hidden layer" . Input layers are just model gates for receiving data and are helpful for knowing the input shape of whatever you want to feed the model . also they are helpful when you have multiple inputs when the model architecture is functional ( see the documents)

however , in this example , you can simply use keras Sequential API and only use one Dense layer as you said for logistic regression , but you have to specify the input shape with the input_shape parameter in that Dense layer

[rishijain07]

Ohh got it thanks :)