Weight initialization strategies in TensorFlow - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why use He initialization in deep neural networks?

He initialization is often recommended for deep networks with ReLU activations. Why is this the case?

A It randomly sets some weights to zero to create sparsity.

B It sets all weights to zero to simplify training.

C It initializes weights with very large values to speed up convergence.

D It helps keep the variance of activations constant across layers, preventing vanishing or exploding gradients.

💡 Hint

Think about how activation variance affects gradient flow in deep networks.

He initialization scales weights to maintain activation variance through layers, especially with ReLU, which helps avoid vanishing or exploding gradients.

❓ Predict Output

intermediate

Output shape after Xavier initialization in TensorFlow

What is the shape of the weights tensor initialized by the following code?

TensorFlow

import tensorflow as tf
initializer = tf.keras.initializers.GlorotUniform()
weights = initializer(shape=(64, 128))
print(weights.shape)

A (64,)

B (128, 64)

C (64, 128)

D (128,)

💡 Hint

Look at the shape argument passed to the initializer.

The initializer creates a tensor with the exact shape specified, here (64, 128).

❓ Hyperparameter

advanced

Choosing initialization for sigmoid activation

Which weight initialization strategy is best suited for a network using sigmoid activation functions to reduce vanishing gradients?

A He initialization

B Xavier (Glorot) initialization

C Random normal with mean 0 and stddev 1

D All zeros initialization

💡 Hint

Consider the activation function's output range and how initialization affects gradient flow.

Xavier initialization is designed to keep variance stable for activations like sigmoid and tanh, helping reduce vanishing gradients.

❓ Metrics

advanced

Effect of poor weight initialization on training metrics

What is the most likely effect on training loss and accuracy if weights are initialized with very large random values?

A Training loss will be unstable and accuracy will be low due to exploding gradients.

B Training loss will decrease smoothly and accuracy will improve quickly.

C Training loss will be zero from the start and accuracy will be perfect.

D Training loss and accuracy will not change because initialization does not affect training.

💡 Hint

Think about how large initial weights affect gradient calculations.

Large initial weights cause activations and gradients to explode, making training unstable and accuracy poor.

🔧 Debug

expert

Identify the error in this custom weight initializer

What error will this TensorFlow custom initializer code raise?

TensorFlow

import tensorflow as tf
class CustomInit(tf.keras.initializers.Initializer):
    def __call__(self, shape, dtype=None):
        return tf.random.uniform(shape, minval=-1, maxval=1, dtype=dtype)

initializer = CustomInit()
weights = initializer(shape=(32, 32))

A No error, code runs correctly

B AttributeError because __call__ method is missing

C ValueError because shape is not a tuple

D TypeError because dtype is not passed to tf.random.uniform

💡 Hint

tf.random.uniform has a default dtype=tf.float32, so no error occurs.

No error occurs because tf.random.uniform uses dtype=tf.float32 by default when not specified.

Practice

1. Which weight initialization strategy is best suited for layers using ReLU activation in TensorFlow?

easy

A. Zeros initializer

B. GlorotUniform initializer

C. RandomNormal initializer with mean 0

D. HeNormal initializer

Solution

1. Step 1: Understand activation and initialization relation

   ReLU activation benefits from initial weights that maintain variance through layers to avoid dying neurons.

2. Step 2: Match initializer to activation

   HeNormal initializer is designed for ReLU, scaling weights to keep variance stable.

3. Final Answer:

   HeNormal initializer -> Option D

4. Quick Check:

   ReLU uses HeNormal [OK]

Hint: Use HeNormal for ReLU activations to keep variance stable [OK]

Common Mistakes:

• Choosing GlorotUniform for ReLU layers
• Using zeros initializer causing no learning
• RandomNormal without scaling causes slow training

2. Which of the following is the correct syntax to initialize weights with GlorotUniform in TensorFlow?

easy

A. tf.keras.initializers.glorot_uniform()

B. tf.initializers.GlorotUniform()

C. tf.keras.initializers.GlorotUniform()

D. tf.keras.initializers.Glorot_uniform()

Solution

1. Step 1: Recall TensorFlow initializer module

   Initializers are accessed via tf.keras.initializers with exact class names.

2. Step 2: Check correct class name and casing

   The correct class name is GlorotUniform with capital G and U, called as a class with parentheses.

3. Final Answer:

   tf.keras.initializers.GlorotUniform() -> Option C

4. Quick Check:

   Correct class name and module path [OK]

Hint: Use exact class name with correct casing and parentheses [OK]

Common Mistakes:

• Using lowercase function names instead of class
• Missing parentheses after class name
• Using wrong module path tf.initializers

3. What will be the output shape and initializer type of the following TensorFlow layer?

layer = tf.keras.layers.Dense(10, kernel_initializer=tf.keras.initializers.HeNormal())
print(layer.kernel_initializer)

medium

A. GlorotUniform initializer object, output shape (None, 10)

B. HeNormal initializer object, output shape (None, 10)

C. HeNormal initializer object, output shape (10, 10)

D. RandomNormal initializer object, output shape (None, 10)

Solution

1. Step 1: Identify initializer and layer output

   The layer uses HeNormal initializer explicitly for kernel weights and has 10 output units.

2. Step 2: Understand output shape of Dense layer

   Dense layers output shape is (batch_size, units), so shape is (None, 10) before input is given.

3. Final Answer:

   HeNormal initializer object, output shape (None, 10) -> Option B

4. Quick Check:

   HeNormal used and output units = 10 [OK]

Hint: Dense output shape is (None, units); initializer matches given class [OK]

Common Mistakes:

• Confusing output shape with kernel shape
• Assuming default initializer instead of HeNormal
• Mixing initializer types

4. You wrote this code to initialize weights but get an error:

layer = tf.keras.layers.Dense(5, kernel_initializer='he_normal')
model = tf.keras.Sequential([layer])
model.build(input_shape=(None, 3))
print(layer.kernel_initializer())

What is the cause of the error?

medium

A. Calling kernel_initializer as a function without arguments

B. Using string 'he_normal' instead of initializer object

C. Not specifying bias_initializer

D. Missing input_shape in layer constructor

Solution

1. Step 1: Understand kernel_initializer attribute type

   kernel_initializer is an initializer object, not a callable function without arguments.

2. Step 2: Identify error cause

   Calling kernel_initializer() without arguments causes error because it expects shape and dtype parameters.

3. Final Answer:

   Calling kernel_initializer as a function without arguments -> Option A

4. Quick Check:

   Initializer needs shape args, can't call empty () [OK]

Hint: Don't call kernel_initializer() without shape and dtype [OK]

Common Mistakes:

• Confusing string initializer with object
• Ignoring required arguments for initializer call
• Assuming bias_initializer causes this error

5. You want to initialize weights for a model with mixed activations: ReLU in some layers and sigmoid in others. Which strategy best improves training speed and accuracy?

hard

A. Use HeNormal for ReLU layers and GlorotUniform for sigmoid layers

B. Use GlorotUniform for all layers regardless of activation

C. Use RandomNormal with mean 0 and stddev 1 for all layers

D. Use Zeros initializer for all layers to start from zero

Solution

1. Step 1: Match initializers to activation functions

   HeNormal suits ReLU to maintain variance; GlorotUniform suits sigmoid/tanh to avoid saturation.

2. Step 2: Understand impact on training

   Proper matching prevents slow or stuck training and improves accuracy by keeping gradients healthy.

3. Final Answer:

   Use HeNormal for ReLU layers and GlorotUniform for sigmoid layers -> Option A

4. Quick Check:

   Match initializer to activation for best training [OK]

Hint: Match HeNormal to ReLU, GlorotUniform to sigmoid/tanh [OK]

Common Mistakes:

• Using same initializer for all activations
• Using zeros initializer causing no learning
• Using RandomNormal without scaling