tensorflow_training.py

import tensorflow as tf
from tensorflow.keras.datasets import mnist
import numpy as np

# Load MNIST dataset
print("Loading MNIST dataset...")
(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Normalize and reshape data
print("Preparing data...")
x_train = x_train.reshape(-1, 784) / 255.0
x_test = x_test.reshape(-1, 784) / 255.0

print(f"Training data shape: {x_train.shape}")
print(f"Test data shape: {x_test.shape}")

# Create model
print("\nBuilding model...")
model = tf.keras.Sequential([
    tf.keras.layers.Dense(128, activation='relu', input_shape=(784,)),
    tf.keras.layers.Dropout(0.2),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dropout(0.2),
    tf.keras.layers.Dense(10, activation='softmax')
])

# Compile model
print("Compiling model...")
model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

# Display model architecture
print("\nModel Architecture:")
model.summary()

# Train model
print("\nTraining model...")
history = model.fit(
    x_train, y_train,
    epochs=10,
    batch_size=32,
    validation_split=0.2,
    verbose=1
)

# Evaluate on test data
print("\nEvaluating on test data...")
test_loss, test_accuracy = model.evaluate(x_test, y_test, verbose=0)
print(f"Test Loss: {test_loss:.4f}")
print(f"Test Accuracy: {test_accuracy:.4f}")

# Make predictions on sample images
print("\nMaking predictions on sample images...")
sample_predictions = model.predict(x_test[:5])
for i, pred in enumerate(sample_predictions):
    predicted_class = np.argmax(pred)
    actual_class = y_test[i]
    confidence = pred[predicted_class]
    print(f"Image {i}: Predicted={predicted_class}, Actual={actual_class}, Confidence={confidence:.4f}")

# Save the model
print("\nSaving model...")
model.save('mnist_model.h5')
print("Model saved as 'mnist_model.h5'")