In [1]:
Copied!
import tensorflow as tf
print("TensorFlow version:", tf.__version__)
import tensorflow as tf
print("TensorFlow version:", tf.__version__)
TensorFlow version: 2.3.0
Loading test database (MNIST)
In [2]:
Copied!
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
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
Building a simple model
In [3]:
Copied!
model = tf.keras.models.Sequential([
tf.keras.layers.Flatten(input_shape=(28, 28)),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dropout(0.2),
tf.keras.layers.Dense(10)
])
model = tf.keras.models.Sequential([
tf.keras.layers.Flatten(input_shape=(28, 28)),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dropout(0.2),
tf.keras.layers.Dense(10)
])
In [4]:
Copied!
model.summary()
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]:
Copied!
predictions = model(x_train[:1]).numpy()
predictions
predictions = model(x_train[:1]).numpy()
predictions
Out[5]:
array([[-1.8647462e-01, 5.3670764e-01, -5.1415348e-01, -1.3579503e-01,
-6.0872316e-01, -5.1377362e-01, -3.4768134e-05, 1.7156357e-01,
-3.9986107e-01, 5.4845825e-02]], dtype=float32)
In [6]:
Copied!
tf.nn.softmax(predictions).numpy()
tf.nn.softmax(predictions).numpy()
Out[6]:
array([[0.09152274, 0.18862666, 0.06595077, 0.09628063, 0.05999966,
0.06597582, 0.11028053, 0.13092513, 0.07393608, 0.11650193]],
dtype=float32)
In [7]:
Copied!
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
In [8]:
Copied!
model.compile(optimizer='adam',
loss=loss_fn,
metrics=['accuracy'])
model.compile(optimizer='adam',
loss=loss_fn,
metrics=['accuracy'])
In [9]:
Copied!
model.fit(x_train, y_train, epochs=5)
model.fit(x_train, y_train, epochs=5)
Epoch 1/5 1875/1875 [==============================] - 2s 879us/step - loss: 0.2947 - accuracy: 0.9137 Epoch 2/5 1875/1875 [==============================] - 2s 925us/step - loss: 0.1437 - accuracy: 0.9580 Epoch 3/5 1875/1875 [==============================] - 2s 954us/step - loss: 0.1102 - accuracy: 0.9661 Epoch 4/5 1875/1875 [==============================] - 2s 900us/step - loss: 0.0900 - accuracy: 0.9730 Epoch 5/5 1875/1875 [==============================] - 2s 860us/step - loss: 0.0766 - accuracy: 0.9758
Out[9]:
<tensorflow.python.keras.callbacks.History at 0x7f1099f839d0>
In [10]:
Copied!
model.evaluate(x_test, y_test, verbose=2)
model.evaluate(x_test, y_test, verbose=2)
313/313 - 0s - loss: 0.0796 - accuracy: 0.9754
Out[10]:
[0.07960231602191925, 0.9753999710083008]
In [11]:
Copied!
probability_model = tf.keras.Sequential([
model,
tf.keras.layers.Softmax()
])
probability_model(x_test[:5])
probability_model = tf.keras.Sequential([
model,
tf.keras.layers.Softmax()
])
probability_model(x_test[:5])
Out[11]:
<tf.Tensor: shape=(5, 10), dtype=float32, numpy=
array([[2.16737632e-07, 1.00639301e-08, 2.08058896e-06, 3.69298097e-04,
8.35847364e-12, 1.86631244e-08, 9.44293539e-14, 9.99598086e-01,
1.12612099e-07, 3.02279786e-05],
[3.15482773e-09, 3.69438885e-06, 9.99993563e-01, 1.00565217e-06,
3.63748135e-15, 1.18580135e-06, 6.52262599e-07, 3.42703665e-12,
2.77176042e-08, 4.78829631e-15],
[7.48431859e-08, 9.98841465e-01, 3.04258778e-04, 2.81406392e-05,
5.11805338e-05, 7.44846739e-06, 2.18839123e-05, 6.27289061e-04,
1.17799733e-04, 3.43982606e-07],
[9.99946833e-01, 1.73406516e-08, 3.81245036e-06, 3.70854252e-07,
2.12491678e-08, 2.05819515e-06, 2.97118731e-06, 1.24853177e-05,
6.72489975e-08, 3.14617173e-05],
[2.61222317e-06, 1.26015087e-09, 1.62396134e-06, 1.11417165e-07,
9.99176919e-01, 1.55488397e-06, 1.43458340e-06, 1.57523991e-05,
2.67988139e-06, 7.97215558e-04]], dtype=float32)>
In [ ]:
Copied!