Weight initialization strategies in TensorFlow - Model Pipeline Trace

Model Pipeline - Weight initialization strategies

This pipeline shows how different weight initialization methods affect the training of a simple neural network. Proper initialization helps the model learn faster and better by starting with good initial weights.

Data Flow - 5 Stages

1Data Input

1000 rows x 20 columns→Load and normalize input features→1000 rows x 20 columns

[0.12, 0.45, 0.33, ..., 0.78]

↓

2Train/Test Split

1000 rows x 20 columns→Split data into training (80%) and testing (20%) sets→Train: 800 rows x 20 columns, Test: 200 rows x 20 columns

Train sample: [0.12, 0.45, ..., 0.78], Test sample: [0.22, 0.55, ..., 0.68]

↓

3Model Initialization

800 rows x 20 columns→Initialize model weights using He initialization→Model weights shape: Layer1 (20, 64), Layer2 (64, 10)

Layer1 weights sample: [0.15, -0.22, 0.05, ...]

↓

4Model Training

800 rows x 20 columns→Train model for 10 epochs→Trained model weights updated

Epoch 1 loss: 0.65, accuracy: 0.70

↓

5Model Evaluation

200 rows x 20 columns→Evaluate model on test data→Test accuracy: 0.82

Predicted labels vs true labels

Training Trace - Epoch by Epoch

Loss
0.7 |****
0.6 |****
0.5 |***
0.4 |**
0.3 |**
0.2 |*
    +----------------
     1 2 3 4 5 6 7 8 9 10 Epochs

Epoch	Loss ↓	Accuracy ↑	Observation
1	0.65	0.70	Model starts learning with moderate loss and accuracy.
2	0.52	0.76	Loss decreases and accuracy improves as weights adjust.
3	0.43	0.81	Training progresses well with steady improvement.
4	0.37	0.84	Model continues to learn effectively.
5	0.33	0.86	Loss decreases further, accuracy rises.
6	0.30	0.87	Training stabilizes with good performance.
7	0.28	0.88	Model converges with small improvements.
8	0.26	0.89	Loss and accuracy plateau near optimal values.
9	0.25	0.90	Final tuning of weights.
10	0.24	0.91	Training completes with high accuracy.

Prediction Trace - 5 Layers

Layer 1: Input Layer

Layer 2: Dense Layer 1 with He initialization

Layer 3: ReLU Activation

Layer 4: Dense Layer 2 with He initialization

Layer 5: Softmax Activation

Model Quiz - 3 Questions

Why is He initialization useful for layers with ReLU activation?

A It initializes weights with very large values to speed up learning.

B It keeps the variance of activations stable to avoid vanishing or exploding gradients.

C It sets all weights to zero for faster training.

D It randomly drops some weights during training.

Key Insight

Good weight initialization like He initialization helps the model start training with well-scaled activations. This prevents problems like vanishing or exploding gradients, allowing the model to learn faster and reach higher accuracy.

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