When does X go on the outside of a layer versus the inside?

[JoshK68]

Hi,

Can you remind me, when we do put (x) on the outside vs on the inside as in:

x = layers.GlobalAveragePooling2D(name="pooling_layer")(x) ### Use 2D b/c our shapes are 2 Dimensional (height * Width)
x = layers.Dense(len(class_names))(x) ### want one output neuron per class

vs. something with x inside (can't think of an example but like this (x) )

Curious when (x) goes on the outside/inside and why.

Thank you,
Josh

[mrdbourke]

Inside

If it's a model (tf.keras.Model), the x goes on the inside.

For example,

import tensorflow as tf
model = tf.keras.applications.EfficientNetB0(include_top=False) # <- this is a model
model.trainable = False
x = tf.keras.layers.Input(shape=(224, 224, 3), name="input_layer")
x = model(x) # <- inside

Outside

If it's a layer (tf.keras.layers.Layer), the x goes on the outside.

For exmaple,

import tensorflow as tf
model = tf.keras.applications.EfficientNetB0(include_top=False)
model.trainable = False
inputs = tf.keras.layers.Input(shape=(224, 224, 3), name="input_layer")
x = model(inputs)
x = tf.keras.layers.GlobalAveragePooling2D()(x) # <- outside, since it's a layer

[JoshK68]

Finally I understand. Thank you very much! :)