recognizer_simple_nn.py

import numpy as np
import pandas as pd
import tensorflow as tf
from tensorflow.keras import layers, models

(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()

train = pd.read_csv('train.csv')
test = pd.read_csv('test.csv')

y_df = train[['label']]
x_df = train.drop(['label'], axis=1)

y = y_df.to_numpy().flatten()
X = x_df.to_numpy()
X = X.reshape(X.shape[0], 28, 28)

model = models.Sequential([
    layers.Flatten(input_shape=(28, 28)),
    layers.Dense(128, activation='relu'),
    layers.Dense(10, activation='softmax')
])

model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])

model.fit(X, y, epochs=10)

test_loss, test_acc = model.evaluate(test_images, test_labels)
print(f'Test accuracy: {test_acc}')

model.save('digit_recognition_model.keras')

test = test.to_numpy()
test = test.reshape(test.shape[0], 28, 28)

predictions = model.predict(test)
predicted_labels = np.argmax(predictions, axis=1)


data = {
    'ImageId': range(1, len(predicted_labels) + 1),
    'Label': predicted_labels
}

result = pd.DataFrame(data)
result.to_csv('submission.csv', index=False)