Diplomado en Inteligencia Artificial y Aprendizaje Profundo ¶
Introducción a la API Sequential de Keras ¶
Profesores¶
- Alvaro Mauricio Montenegro Díaz, ammontenegrod@unal.edu.co
- Daniel Mauricio Montenegro Reyes, dextronomo@gmail.com
- Campo Elías Pardo Turriago, cepardot@unal.edu.co
Asesora Medios y Marketing digital¶
- Maria del Pilar Montenegro, pmontenegro88@gmail.com
Asistentes¶
- Oleg Jarma, ojarmam@unal.edu.co
- Laura Lizarazo, ljlizarazore@unal.edu.co
Contenido¶
Introducción¶
Este es un notebook de Google Colaboratory. Los programas de Python se executan directamente en tu navegador, una gran manera de aprender y utilizar TensorFlow. Para poder seguir este tutorial, ejecuta este notebook en Google Colab presionando el boton en la parte superior de esta pagina.
In [1]:
from __future__ import absolute_import, division, print_function
import tensorflow as tf
print('Version de Tensorflow = ', tf.__version__)
Version de Tensorflow = 2.4.1
Prepara datos de MNIST¶
In [2]:
mnist = tf.keras.datasets.mnist
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train/255.0, x_test/255.0
Modelo Sequential de Keras¶
In [3]:
import tensorflow.keras.layers as layer
model = tf.keras.models.Sequential([
layer.Flatten(input_shape=(28,28)), #784
layer.Dense(128, activation='relu'),
layer.Dropout(0.2),
layer.Dense(10, activation='softmax')
])
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
Summary y Plot Model¶
In [4]:
model.summary()
Model: "sequential" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= flatten (Flatten) (None, 784) 0 _________________________________________________________________ dense (Dense) (None, 128) 100480 _________________________________________________________________ dropout (Dropout) (None, 128) 0 _________________________________________________________________ dense_1 (Dense) (None, 10) 1290 ================================================================= Total params: 101,770 Trainable params: 101,770 Non-trainable params: 0 _________________________________________________________________
In [5]:
from tensorflow.keras.utils import plot_model
plot_model(model, to_file='../Imagenes/mnist_dense.png',
show_shapes=True, show_dtype=False)
Out[5]:
Puede necesitar instalar graphviz.
In [ ]:
#from plot_model import plot_model
#plot_model(model)
Entrenamiento y evaluación del modelo¶
In [6]:
history = model.fit(x_train, y_train, epochs=5)
Epoch 1/5 1875/1875 [==============================] - 20s 10ms/step - loss: 0.4859 - accuracy: 0.8617 Epoch 2/5 1875/1875 [==============================] - 21s 11ms/step - loss: 0.1453 - accuracy: 0.9564 Epoch 3/5 1875/1875 [==============================] - 18s 10ms/step - loss: 0.1066 - accuracy: 0.9675 Epoch 4/5 1875/1875 [==============================] - 19s 10ms/step - loss: 0.0904 - accuracy: 0.9714 Epoch 5/5 1875/1875 [==============================] - 18s 10ms/step - loss: 0.0750 - accuracy: 0.9771
In [7]:
model.evaluate(x_test, y_test, verbose=2)
313/313 - 2s - loss: 0.0738 - accuracy: 0.9786
Out[7]:
[0.07382926344871521, 0.978600025177002]
Primeros gráficos¶
Extrae datos¶
In [20]:
import matplotlib.pyplot as plt
import pandas as pd
hist = pd.DataFrame(history.history)
hist['epoch'] = history.epoch
hist
Out[20]:
| loss | accuracy | epoch | |
|---|---|---|---|
| 0 | 0.293235 | 0.916267 | 0 |
| 1 | 0.140098 | 0.958250 | 1 |
| 2 | 0.105634 | 0.968300 | 2 |
| 3 | 0.088549 | 0.972383 | 3 |
| 4 | 0.076502 | 0.976267 | 4 |
Pérdida¶
In [27]:
plt.figure()
plt.xlabel('Epoca')
plt.ylabel('Loss: Entropía cruzada')
plt.plot(hist['epoch'], hist['loss'],
label='Error en entrenamiento')
plt.ylim([0,0.5])
plt.legend()
plt.show()
Exactitud¶
In [29]:
plt.figure()
plt.xlabel('Epoca')
plt.ylabel('Exactitud')
plt.plot(hist['epoch'], hist['accuracy'],
label='Exactitud en entrenamiento')
plt.ylim([0.9,1])
plt.legend()
plt.show()
In [ ]: