pip install tensorflow
python
import tensorflow as tf
dataset = ...
def preprocess_fn(data):
...
dataset = dataset.map(preprocess_fn)
train_dataset = dataset.take(train_size)
test_dataset = dataset.skip(train_size)
train_dataset = train_dataset.shuffle(buffer_size).batch(batch_size)
test_dataset = test_dataset.batch(batch_size)
train_iterator = train_dataset.make_initializable_iterator()
test_iterator = test_dataset.make_initializable_iterator()
train_data = train_iterator.get_next()
test_data = test_iterator.get_next()
model_input = ...
...
...
...
...
python
import tensorflow as tf
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
x_train = x_train.astype('float32') / 255.0
x_test = x_test.astype('float32') / 255.0
model = tf.keras.Sequential([...])
model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
model.fit(x_train, y_train, epochs=5, batch_size=64)
python
import tensorflow as tf
model = tf.keras.Sequential([...])
model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
model.fit(train_data, epochs=5, batch_size=64)
loss, accuracy = model.evaluate(test_data)
python
import tensorflow as tf
cluster_spec = tf.train.ClusterSpec({
'worker': ['worker0:port', 'worker1:port', ...],
'ps': ['ps0:port', 'ps1:port', ...]
})
server = tf.train.Server(cluster_spec, job_name='worker', task_index=0)
with tf.device('/job:worker/task:0'):
...
