Multi-input and multi-output models in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to define two input layers for a multi-input model.

TensorFlow

input1 = tf.keras.Input(shape=[1])
input2 = tf.keras.Input(shape=(10,))

Drag options to blanks, or click blank then click option'

ANone

B32

C[32]

D(32,)

Attempts:

3 left

💡 Hint

Common Mistakes

Using an integer instead of a tuple for input shape.

Using a list instead of a tuple for input shape.

The input shape must be a tuple representing the shape of the input data excluding the batch size. Here, (32,) means the input has 32 features.

2fill in blank

medium

Complete the code to concatenate two input tensors in the model.

TensorFlow

concat = tf.keras.layers.Concatenate()([input1, [1]])

Drag options to blanks, or click blank then click option'

Ainput2

Binput3

Cinputs

Dinput_2

Attempts:

3 left

💡 Hint

Common Mistakes

Using an undefined variable name.

Using a string instead of a tensor variable.

To concatenate two inputs, you must pass both input tensors. Here, input2 is the second input tensor to concatenate with input1.

3fill in blank

hard

Fix the error in the model output definition by selecting the correct output layer.

TensorFlow

output1 = tf.keras.layers.Dense(1, activation='sigmoid')([1])
output2 = tf.keras.layers.Dense(10, activation='softmax')(dense2)

Drag options to blanks, or click blank then click option'

Ainput2

Binput1

Cconcat

Ddense1

Attempts:

3 left

💡 Hint

Common Mistakes

Passing input layers directly to output layers.

Passing unrelated layer outputs.

The output layer should receive the concatenated tensor 'concat' as input, not the raw inputs or unrelated layers.

4fill in blank

hard

Fill both blanks to compile the multi-output model with appropriate loss functions and metrics.

TensorFlow

model.compile(optimizer='adam', loss=[1], metrics=[2])

Drag options to blanks, or click blank then click option'

A{'output1': 'binary_crossentropy', 'output2': 'categorical_crossentropy'}

B['mse', 'mae']

C{'output1': ['accuracy'], 'output2': ['accuracy']}

D['accuracy', 'accuracy']

Attempts:

3 left

💡 Hint

Common Mistakes

Using lists instead of dictionaries for losses and metrics.

Not matching output names in loss and metrics.

For multi-output models, loss must be a dictionary mapping output names to loss functions. Metrics can also be a dictionary mapping output names to metric lists.

5fill in blank

hard

Fill all three blanks to create a multi-input, multi-output model with correct inputs, outputs, and model definition.

TensorFlow

input_a = tf.keras.Input(shape=[1])
input_b = tf.keras.Input(shape=[2])
model = tf.keras.Model(inputs=[input_a, input_b], outputs=[3])

Drag options to blanks, or click blank then click option'

A(64,)

B(10,)

C[output1, output2]

D(32,)

Attempts:

3 left

💡 Hint

Common Mistakes

Using integers instead of tuples for input shapes.

Passing outputs as a tuple instead of a list.

The inputs must have correct shapes as tuples. The outputs must be a list of output layers for multi-output models.