5

!pip install tensorflow

import tensorflow as tf




#1. Load and preprocess the CIFAR-10 dataset:

from tensorflow.keras.datasets import cifar10

# Load dataset

(X_train, _), (X_test, _) = cifar10.load_data()

# Normalize the data

X_train = X_train.astype('float32') / 255.0

X_test = X_test.astype('float32') / 255.0

#2. Add noise to the CIFAR-10 images:

import numpy as np

noise_factor = 0.5

X_train_noisy = X_train + noise_factor * np.random.normal(loc=0.0,

scale=1.0, size=X_train.shape)

X_test_noisy = X_test + noise_factor * np.random.normal(loc=0.0,

scale=1.0, size=X_test.shape)

# Clip to keep pixel values between 0 and 1

X_train_noisy = np.clip(X_train_noisy, 0., 1.)

X_test_noisy = np.clip(X_test_noisy, 0., 1.)

#3. Define the convolutional autoencoder (CAE):

model = tf.keras.Sequential([

 # Encoder

 tf.keras.layers.Conv2D(32, (3, 3), activation='relu', padding='same',

input_shape=(32, 32, 3)),

 tf.keras.layers.MaxPooling2D((2, 2), padding='same'),

 tf.keras.layers.Conv2D(64, (3, 3), activation='relu', padding='same'),

 tf.keras.layers.MaxPooling2D((2, 2), padding='same'),

 # Decoder

 tf.keras.layers.Conv2DTranspose(64, (3, 3), activation='relu',

padding='same'),

 tf.keras.layers.UpSampling2D((2, 2)),

tf.keras.layers.Conv2DTranspose(32, (3, 3), activation='relu',

padding='same'),

 tf.keras.layers.UpSampling2D((2, 2)),

 tf.keras.layers.Conv2D(3, (3, 3), activation='sigmoid',

padding='same')

])

model.compile(optimizer='adam', loss='mse')

#4. Train the autoencoder on noisy CIFAR-10 images:

history = model.fit(X_train_noisy, X_train, epochs=50, batch_size=128,

validation_data=(X_test_noisy, X_test))

#5. Visualize original, noisy, and reconstructed images:

import matplotlib.pyplot as plt

# Reconstruct images

decoded_imgs = model.predict(X_test_noisy)

# Display original, noisy, and reconstructed images

n = 10

plt.figure(figsize=(20, 6))

for i in range(n):

 # Display original

 ax = plt.subplot(3, n, i + 1)

 plt.imshow(X_test[i])

 plt.title("Original")

 plt.axis('off')

 # Display noisy

 ax = plt.subplot(3, n, i + 1 + n)

 plt.imshow(X_test_noisy[i])

 plt.title("Noisy")

 plt.axis('off')

 # Display reconstruction

 ax = plt.subplot(3, n, i + 1 + 2*n)

 plt.imshow(decoded_imgs[i])

 plt.title("Reconstructed")

 plt.axis('off')

plt.show()

#6. Evaluate the performance using Mean Squared Error (MSE):

mse = np.mean(np.square(X_test - decoded_imgs))

print(f"Reconstruction MSE: {mse}") 

Comments

Post a Comment

Popular posts from this blog

2

#1 from tensorflow.keras.datasets import fashion_mnist (x_train, y_train), (x_test, y_test) = fashion_mnist.load_data() x_train,x_test=x_train/255.0,x_test/255.0 #2 import tensorflow as tf model = tf.keras.Sequential([   tf.keras.layers.Flatten(input_shape=(28, 28)),   tf.keras.layers.Dense(128, activation='relu'),   tf.keras.layers.BatchNormalization(),   tf.keras.layers.Dropout(0.5),   tf.keras.layers.Dense(64, activation='relu'),   tf.keras.layers.Dense(10, activation='softmax') ]) model.compile(optimizer='adam',               loss='sparse_categorical_crossentropy',               metrics=['accuracy']) #3 lr_schedule = tf.keras.callbacks.LearningRateScheduler(     lambda epoch: 1e-3 * 10**(epoch / 20)) history = model.fit(x_train, y_train, epochs=10,                     validation_data=(x_test, y_test),       ...
Read more

1

#Steps: #1. Set up the environment: !pip install tensorflow import numpy as np #2. Define the OR and XOR gate datasets: # Define OR and XOR datasets X_OR = np.array([[0, 0], [0, 1], [1, 0], [1, 1]], dtype=np.float32) y_OR = np.array([[0], [1], [1], [1]], dtype=np.float32) X_XOR = np.array([[0, 0], [0, 1], [1, 0], [1, 1]], dtype=np.float32) y_XOR = np.array([[0], [1], [1], [0]], dtype=np.float32) #3. Create a single-layer perceptron model: import tensorflow as tf # Function to create and train a single-layer perceptron with improved #accuracy def train_perceptron(X, y, epochs=100, learning_rate=0.1):  model = tf.keras.Sequential([  tf.keras.layers.Dense(1, activation='sigmoid', input_shape=(2,))  ])  model.compile(optimizer=tf.keras.optimizers.SGD(learning_rate=  learning_rate),loss='binary_crossentropy',  metrics=['accuracy'])  model.fit(X, y, epochs=epochs, verbose=0)  return model #4. Train the model on OR and XOR datasets: # Train on OR gate wi...
Read more

3

#1.load amd preprocess the CIFAR-10 dataset import numpy as np import tensorflow as tf from tensorflow.keras.datasets import cifar10 from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense from sklearn.metrics import accuracy_score, precision_score, recall_score # Load the CIFAR-10 dataset (x_train, y_train), (x_test, y_test) = cifar10.load_data() # Normalize pixel values to be between 0 and 1 x_train = x_train.astype('float32') / 255.0 x_test = x_test.astype('float32') / 255.0 # convert class vectors to binary class matrices(one-hot encoding) y_train = tf.keras.utils.to_categorical(y_train,num_classes=10) y_test = tf.keras.utils.to_categorical(y_test,num_classes=10) #2.define a deeper cnn model #define the cnn model model = Sequential() model.add(Conv2D(32, (3, 3), activation='relu', padding='same', input_shape=(32, 32, 3))) model.add(Conv2D(32, (3, 3), activation='relu', padding=...
Read more