Functional API basics in TensorFlow - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output shape of a Functional API model

Consider the following TensorFlow Functional API model. What is the shape of the output tensor?

TensorFlow

import tensorflow as tf
inputs = tf.keras.Input(shape=(32,))
x = tf.keras.layers.Dense(64, activation='relu')(inputs)
outputs = tf.keras.layers.Dense(10)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
print(model.output_shape)

A (32, 64)

B (32, 10)

C (None, 64)

D (None, 10)

💡 Hint

Remember the first dimension is batch size and is None by default.

The model input shape is (32,), so the output shape is (None, 10) where None is the batch size.

❓ Model Choice

intermediate

Choosing the correct Functional API model for multiple inputs

You want to build a model that takes two inputs: one with shape (10,) and another with shape (5,). Which Functional API model code correctly defines this?

A

input1 = tf.keras.Input(shape=(10,))
input2 = tf.keras.Input(shape=(5,))
concat = tf.keras.layers.Concatenate()([input1, input2])
output = tf.keras.layers.Dense(1)(concat)
model = tf.keras.Model(inputs=[input1, input2], outputs=output)

B

input1 = tf.keras.Input(shape=(10,))
input2 = tf.keras.Input(shape=(5,))
concat = tf.keras.layers.Dense(15)(input1)
output = tf.keras.layers.Dense(1)(concat)
model = tf.keras.Model(inputs=[input1, input2], outputs=output)

C

input1 = tf.keras.Input(shape=(10,))
input2 = tf.keras.Input(shape=(5,))
concat = tf.keras.layers.Multiply()([input1, input2])
output = tf.keras.layers.Dense(1)(concat)
model = tf.keras.Model(inputs=[input1, input2], outputs=output)

D

input1 = tf.keras.Input(shape=(10,))
input2 = tf.keras.Input(shape=(5,))
concat = tf.keras.layers.Add()([input1, input2])
output = tf.keras.layers.Dense(1)(concat)
model = tf.keras.Model(inputs=[input1, input2], outputs=output)

💡 Hint

Concatenate layer joins inputs along last axis; Add and Multiply require same shape inputs.

Option A correctly concatenates two inputs of different shapes. Add and Multiply layers require inputs of the same shape, so A and C are invalid. Option A ignores input2.

❓ Hyperparameter

advanced

Effect of activation function in Functional API model

In a Functional API model, you replace the activation in a Dense layer from 'relu' to 'linear'. What is the most likely effect on the model's training?

A The model will train faster because linear activation is simpler.

B The model may train slower or fail to learn complex patterns due to lack of non-linearity.

C The model will always overfit because linear activation increases complexity.

D The model will have better accuracy because linear activation avoids vanishing gradients.

💡 Hint

Activation functions add non-linearity needed for learning complex data.

Replacing 'relu' with 'linear' removes non-linearity, limiting the model's ability to learn complex patterns, often causing slower or failed training.

🔧 Debug

advanced

Identifying error in Functional API model definition

What error will this Functional API code raise when run?

TensorFlow

import tensorflow as tf
inputs = tf.keras.Input(shape=(20,))
x = tf.keras.layers.Dense(10)(inputs)
outputs = tf.keras.layers.Dense(5)(inputs)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer='adam', loss='mse')

A TypeError: Inputs and outputs must be layers or tensors connected in a graph.

B ValueError: Output tensor must be connected to the input tensor.

C No error; model compiles successfully.

D RuntimeError: Model outputs must come from the last layer applied to inputs.

💡 Hint

Outputs can be any tensor derived from inputs, not necessarily last layer.

The code defines outputs directly from inputs, which is valid. Model compiles without error.

🧠 Conceptual

expert

Understanding model summary in Functional API

Given a Functional API model with two input layers merged by Concatenate and followed by two Dense layers, what does the model.summary() show for the total params if the first Dense layer has 8 units and the second has 4 units? Assume input shapes are (3,) and (2,).

A Total params = (3+2)*8 + 8 + 8*4 + 4 = 84

B Total params = (3+2)*8 + 8 + (3+2)*4 + 4 = 72

C Total params = (3+2)*8 + 8 + 8*4 + 4*2 = 70

D Total params = (3+2)*8 + 8 + 8*4 + 4*1 = 66

💡 Hint

Calculate params as (input_dim * units) + bias per Dense layer, sum all layers.

Concatenate merges inputs to dimension 5. First Dense(8): 5*8 + 8 = 48 params. Second Dense(4): 8*4 + 4 = 36 params. Total: 84. Option A shows the correct calculation.