Custom loss functions in TensorFlow

Introduction

Sometimes, the built-in loss functions do not fit your problem well. Custom loss functions let you tell the model exactly how to measure errors in a way that matches your goal.

You want the model to focus more on certain types of mistakes.

Your problem needs a special way to compare predictions and true values.

You want to combine multiple loss calculations into one.

You are working on a new task that standard losses don't cover.

You want to add penalties or rewards based on extra conditions.

Syntax

TensorFlow

def custom_loss(y_true, y_pred):
    loss = some_operations_on(y_true, y_pred)
    return loss

The function takes two inputs: y_true (real values) and y_pred (model predictions).

It returns a single number representing the error for each example or batch.

Examples

This custom loss calculates the average absolute difference between true and predicted values.

TensorFlow

def mean_absolute_error(y_true, y_pred):
    return tf.reduce_mean(tf.abs(y_true - y_pred))

This loss adds a penalty term to the usual squared error to discourage large predictions.

TensorFlow

def squared_error_with_penalty(y_true, y_pred):
    base_loss = tf.reduce_mean(tf.square(y_true - y_pred))
    penalty = 0.1 * tf.reduce_mean(tf.square(y_pred))
    return base_loss + penalty

Sample Model

This program defines a custom mean squared error loss, trains a simple model on linear data, and shows the final loss and predictions.

TensorFlow

import tensorflow as tf

# Define a simple custom loss function
@tf.function
def custom_mse(y_true, y_pred):
    return tf.reduce_mean(tf.square(y_true - y_pred))

# Create some sample data
x = tf.constant([[1.0], [2.0], [3.0], [4.0]])
y = tf.constant([[2.0], [4.0], [6.0], [8.0]])

# Build a simple linear model
model = tf.keras.Sequential([
    tf.keras.layers.Dense(1, input_shape=(1,))
])

# Compile the model with the custom loss
model.compile(optimizer='sgd', loss=custom_mse, metrics=['mse'])

# Train the model
history = model.fit(x, y, epochs=5, verbose=0)

# Make predictions
predictions = model.predict(x)

# Print training loss and predictions
print(f"Final training loss: {history.history['loss'][-1]:.4f}")
print("Predictions:")
print(predictions)

Important Notes

Custom loss functions must return a tensor with the loss value.

Use TensorFlow operations inside the loss to keep it compatible with training.

Decorate the loss function with @tf.function for better performance if needed.

Summary

Custom loss functions let you control how the model learns by defining your own error measure.

They take true and predicted values as input and return a number showing how bad the prediction is.

You can use them to solve special problems or improve model behavior.

Practice

1. What is the main purpose of a custom loss function in TensorFlow?

easy

A. To change the model's architecture layers

B. To generate new training data from existing data

C. To speed up the training process automatically

D. To define a special way to measure how wrong the model's predictions are

Solution

1. Step 1: Understand the role of loss functions

   Loss functions measure how far the model's predictions are from the true values.

2. Step 2: Identify what custom loss functions do

   Custom loss functions let you define your own way to measure this error, which can help in special cases.

3. Final Answer:

   To define a special way to measure how wrong the model's predictions are -> Option D

4. Quick Check:

   Custom loss = custom error measure [OK]

Hint: Loss functions measure error; custom ones define your own error [OK]

Common Mistakes:

• Confusing loss function with model architecture
• Thinking loss speeds up training automatically
• Mixing loss function with data generation

2. Which of the following is the correct way to define a simple custom loss function in TensorFlow using Python?

easy

A. def custom_loss(y_true, y_pred): return tf.add(y_true, y_pred)

B. def custom_loss(y_true, y_pred): return y_true + y_pred

C. def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.square(y_true - y_pred))

D. def custom_loss(y_true, y_pred): return tf.matmul(y_true, y_pred)

Solution

1. Step 1: Check the function signature

   The function must take true and predicted values as inputs.

2. Step 2: Verify the loss calculation

   Loss should measure error, e.g., mean squared error: mean of squared differences.

3. Final Answer:

   def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.square(y_true - y_pred)) -> Option C

4. Quick Check:

   Loss = mean squared error formula [OK]

Hint: Loss = mean of squared differences for regression [OK]

Common Mistakes:

• Returning sum instead of mean
• Adding true and predicted values instead of difference
• Using matrix multiplication instead of error calculation

3. Given the following custom loss function, what will be the output if y_true = [2.0, 3.0] and y_pred = [1.0, 5.0]?

def custom_loss(y_true, y_pred):
    return tf.reduce_sum(tf.abs(y_true - y_pred))

medium

A. 3.0

B. 4.0

C. 5.0

D. 6.0

Solution

1. Step 1: Calculate absolute differences

   abs(2.0 - 1.0) = 1.0 and abs(3.0 - 5.0) = 2.0

2. Step 2: Sum the absolute differences

   1.0 + 2.0 = 3.0

3. Final Answer:

   3.0 -> Option A

4. Quick Check:

   Sum of absolute errors = 3.0 [OK]

Hint: Sum absolute differences for this loss [OK]

Common Mistakes:

• Confusing sum with mean
• Using squared differences instead of absolute
• Mixing up true and predicted values

4. Identify the error in this custom loss function and select the correct fix:

def custom_loss(y_true, y_pred):
    return tf.reduce_mean(y_true - y_pred)

medium

A. Change function to return y_true + y_pred

B. Use absolute difference or squared difference instead of direct subtraction

C. Remove tf.reduce_mean and return raw difference

D. Use tf.matmul instead of subtraction

Solution

1. Step 1: Understand the problem with direct subtraction

   Direct subtraction can give positive and negative values that cancel out, not measuring error properly.

2. Step 2: Correct the loss calculation

   Use absolute value or square the difference to measure error magnitude correctly.

3. Final Answer:

   Use absolute difference or squared difference instead of direct subtraction -> Option B

4. Quick Check:

   Loss must measure error magnitude, not signed difference [OK]

Hint: Loss must be positive; use abs() or square() [OK]

Common Mistakes:

• Returning signed difference causing cancellation
• Removing mean reduces loss meaning
• Using matrix multiplication wrongly

5. You want to create a custom loss function that penalizes under-predictions more than over-predictions. Which of the following custom loss functions achieves this behavior correctly?

hard

A. def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.where(y_pred < y_true, 2 * (y_true - y_pred), y_pred - y_true))

B. def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.abs(y_true - y_pred))

C. def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.square(y_true - y_pred))

D. def custom_loss(y_true, y_pred): return tf.reduce_sum(y_pred - y_true)

Solution

1. Step 1: Understand the requirement

   We want to penalize under-predictions (y_pred < y_true) more than over-predictions.

2. Step 2: Analyze each option

   def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.where(y_pred < y_true, 2 * (y_true - y_pred), y_pred - y_true)) uses tf.where to multiply the error by 2 when y_pred < y_true, increasing penalty for under-predictions. Options B and C treat errors symmetrically. def custom_loss(y_true, y_pred): return tf.reduce_sum(y_pred - y_true) sums signed differences, which is incorrect.

3. Final Answer:

   def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.where(y_pred < y_true, 2 * (y_true - y_pred), y_pred - y_true)) -> Option A

4. Quick Check:

   Conditional penalty for under-prediction = def custom_loss(y_true, y_pred): return tf.reduce_mean(tf.where(y_pred < y_true, 2 * (y_true - y_pred), y_pred - y_true)) [OK]

Hint: Use tf.where to weight errors differently [OK]

Common Mistakes:

• Using symmetric loss ignoring under/over difference
• Summing signed differences causing wrong loss
• Not using conditional logic for penalty