Introduction

Building your first Neural Network is one of the most exciting steps in your Deep Learning journey. In this tutorial, you'll learn how to create, train, and evaluate a Neural Network using Keras and TensorFlow.

Project Goal

We will create a Neural Network that learns to classify whether a number is even or odd based on input data. This simple example demonstrates the full Deep Learning pipeline.

• Input → Single numeric feature
• Output → 0 (even) or 1 (odd)
• Goal → Train a model to predict correctly

Prerequisites

pip install tensorflow numpy

Neural Network Building Workflow

Step 1 — Import Required Libraries

import numpy as np
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense

Step 2 — Prepare the Dataset

We'll create a small dataset to predict even (0) and odd (1) numbers.

X = np.array([2, 3, 4, 5, 6, 7, 8, 9, 10, 11], dtype=float)
y = np.array([0, 1, 0, 1, 0, 1, 0, 1, 0, 1], dtype=float)
print("Features:", X)
print("Labels:", y)

Step 3 — Build the Neural Network

Now we'll build a simple Neural Network with two layers:

model = Sequential([
    Dense(8, activation="relu", input_shape=(1,)),
    Dense(1, activation="sigmoid")
])

model.summary()

Step 4 — Compile the Model

The compile step prepares the model for training.

model.compile(
    optimizer="adam",
    loss="binary_crossentropy",
    metrics=["accuracy"]
)

Step 5 — Train the Model

model.fit(X, y, epochs=200, verbose=0)
print("Training Complete!")

Step 6 — Evaluate the Model

loss, accuracy = model.evaluate(X, y, verbose=0)
print(f"Accuracy: {accuracy * 100:.2f}%")

Step 7 — Make Predictions

predictions = model.predict(np.array([3, 8, 15]))
print(predictions)

• Values close to 1 → Odd
• Values close to 0 → Even

Full Code — Building Your First Neural Network

import numpy as np
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense

# Step 1: Prepare data
X = np.array([2, 3, 4, 5, 6, 7, 8, 9, 10, 11], dtype=float)
y = np.array([0, 1, 0, 1, 0, 1, 0, 1, 0, 1], dtype=float)

# Step 2: Build model
model = Sequential([
    Dense(8, activation="relu", input_shape=(1,)),
    Dense(1, activation="sigmoid")
])

# Step 3: Compile
model.compile(optimizer="adam", loss="binary_crossentropy", metrics=["accuracy"])

# Step 4: Train
model.fit(X, y, epochs=200, verbose=0)

# Step 5: Evaluate
loss, accuracy = model.evaluate(X, y, verbose=0)
print(f"Accuracy: {accuracy * 100:.2f}%")

# Step 6: Predict
predictions = model.predict(np.array([3, 8, 15]))
print("Predictions:", predictions)

Save and Load Your Model

model.save("first_neural_network.h5")

from tensorflow.keras.models import load_model
loaded_model = load_model("first_neural_network.h5")

Behind the Scenes — What Happens?

Visual Idea

Real-Life Analogy

When to Add More Layers?

• When the problem becomes complex.
• When patterns are non-linear.
• When working with images, audio, or text.
• When more data is available.

Real-World Use Cases of Neural Networks

Advantages of This Approach

• Easy to build with Keras.
• Trains quickly on simple data.
• Helps you understand Deep Learning basics.
• Reusable in larger projects.
• Strong foundation for future models.

Common Mistakes to Avoid

Best Practices

Importance of Building Your First Neural Network

Golden Rule