In [1]:
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
In [2]:
# Load MNIST data
mnist = input_data.read_data_sets("MNIST_data/")
In [3]:
train_data = mnist.train.images
train_labels = mnist.train.labels
test_data = mnist.test.images
test_labels = mnist.test.labels
print("train_data", train_data.shape)
print("train_labels", train_labels.shape)
In [4]:
def my_model_fn(features, labels, mode):
"""Model function for our CNN"""
net = tf.reshape(features['x'], [-1, 28, 28, 1])
for _ in range(3):
net = tf.layers.conv2d(
inputs=net,
filters=32,
kernel_size=[3, 3],
padding="same",
activation=tf.nn.relu
)
net = tf.layers.max_pooling2d(
inputs=net,
pool_size=[2, 2],
strides=[2,2]
)
net = tf.layers.flatten(net)
net = tf.layers.dense(inputs=net, units=64)
logits = tf.layers.dense(inputs=net, units=10)
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32), depth=10)
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels,
logits=logits)
if mode == tf.estimator.ModeKeys.TRAIN:
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001)
train_op = optimizer.minimize(loss=loss, global_step=tf.train.get_global_step())
return tf.estimator.EstimatorSpec(mode=mode, loss=loss, train_op=train_op)
eval_metric_ops = {
"accuracy": tf.metrics.accuracy(labels=labels,
predictions=tf.argmax(input=logits, axis=1))
}
return tf.estimator.EstimatorSpec(mode=mode, loss=loss, eval_metric_ops=eval_metric_ops)
In [5]:
mnist_estimator = tf.estimator.Estimator(
model_fn=my_model_fn,
model_dir="E:\\temp\\mnist_estimator"
)
In [6]:
train_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": train_data},
y=train_labels,
batch_size=64,
num_epochs=10,
shuffle=True
)
In [ ]:
mnist_estimator.train(
input_fn=train_input_fn)
In [8]:
test_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": test_data},
y=test_labels,
num_epochs=1,
shuffle=False)
In [ ]:
test_results = mnist_estimator.evaluate(input_fn=test_input_fn)
print(test_results)
In [11]:
train_spec = tf.estimator.TrainSpec(
input_fn=train_input_fn,
max_steps=50000
)
test_spec = tf.estimator.EvalSpec(
input_fn=test_input_fn,
steps=50, throttle_secs=60
)
tf.estimator.train_and_evaluate(mnist_estimator, train_spec, test_spec)
In [ ]: