Advanced Functions and Decorators

Duration: 8 min

Master advanced function techniques that are essential for building scalable AI applications.

Function Decorators

Decorators allow you to modify or enhance functions without changing their source code. They're widely used in AI frameworks like TensorFlow and PyTorch.

def timing_decorator(func):
    import time
    def wrapper(*args, **kwargs):
        start = time.time()
        result = func(*args, **kwargs)
        end = time.time()
        print(f"{func.__name__} took {end - start:.4f} seconds")
        return result
    return wrapper

@timing_decorator
def train_model(data):
    # Simulate training
    import time
    time.sleep(2)
    return "Model trained"

train_model([1, 2, 3])

Try it in Google Colab:

train_model took 2.0023 seconds

Higher-Order Functions

Functions that take other functions as arguments or return functions are powerful for creating flexible AI pipelines.

def apply_operation(func, x, y):
    return func(x, y)

def add(a, b):
    return a + b

def multiply(a, b):
    return a * b

result1 = apply_operation(add, 5, 3)
result2 = apply_operation(multiply, 5, 3)
print(f"Add: {result1}, Multiply: {result2}")

Add: 8, Multiply: 15

Lambda Functions

Lambda functions are anonymous functions useful for short operations in data processing pipelines.

numbers = [1, 2, 3, 4, 5]
squared = list(map(lambda x: x**2, numbers))
print(f"Squared: {squared}")

# Filter even numbers
evens = list(filter(lambda x: x % 2 == 0, numbers))
print(f"Evens: {evens}")

Squared: [1, 4, 9, 16, 25]
Evens: [2, 4]

💡 Tip: Use decorators to add logging, caching, or validation to your AI model functions without modifying the core logic.