Expert-level academic advice for working with Mnist datasets and Convolutional Neural Networks
import matplotlib.pyplot as plt
from tensorflow.keras.datasets import mnist
from tensorflow import keras
from tensorflow.keras.layers import Dense, Flatten, Dropout, Conv2D, MaxPooling2D
(x_train, y_train), (x_test, y_test) = mnist.load_data() # Load Mnist datasets from the tensorflow.keras.datasets library
# Standardize the input data
x_train = x_train / 255
x_test = x_test / 255
y_train_cat = keras.utils.to_categorical(y_train, 10) # Convert labels to categorical using keras utils
y_test_cat = keras.utils.to_categorical(y_test, 10)
x_train = np.expand_dims(x_train, axis=3) # Add a dimension to the input data to work with Conv2D layers
x_test = np.expand_dims(x_test, axis=3)
print(x_train.shape)
model = keras.Sequential([
Conv2D(32, (3,3), padding='same', activation='relu', input_shape=(28,28,1)),
MaxPooling2D((2,2), strides=2),
Conv2D(64, (3,3), padding='same', activation='relu'),
MaxPooling2D((2,2), strides=2),
Flatten(),
Dense(128, activation='relu'),
Dense(10, activation='softmax')
]) # Create a sequential model with Conv2D, MaxPooling2D, and Dense layers
print(model.summary()) # Print model summary for reference
model.compile(optimizer='adam',
loss='categorical_crossentropy',
metrics=['accuracy']) # Compile the model using Adam optimizer, categorical crossentropy loss, and accuracy metrics
history = model.fit(x_train, y_train_cat, batch_size=32, epochs=5, validation_split=0.2) # Train the model on the training data with batch size of 32, 5 epochs, and using 20% of the data as validation set
model.evaluate(x_test, y_test_cat) # Evaluate the model performance on the test set
