def _check_dimensions(self, partition):
provider = cifar10.Provider(seed=4)
input_fn = provider.get_input_fn(partition,
tf.contrib.learn.ModeKeys.TRAIN,
batch_size=3)
data, labels = input_fn()
self.assertIn(cifar10.FEATURES, data)
features = data[cifar10.FEATURES]
init = tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer())
with self.test_session() as sess:
sess.run(init)
self.assertEqual((3, 32, 32, 3), sess.run(features).shape)
self.assertEqual((3, 1), sess.run(labels).shape)
def main(argv):
del argv
run_config = make_run_config()
estimator_builder = adanet_improve_nas.Builder()
hparams = estimator_builder.hparams(FLAGS.batch_size, FLAGS.hparams)
tf.logging.info("Running Experiment with HParams: %s", hparams)
if FLAGS.dataset == "cifar10":
data_provider = cifar10.Provider()
elif FLAGS.dataset == "cifar100":
data_provider = cifar100.Provider()
elif FLAGS.dataset == "fake":
data_provider = fake_data.FakeImageProvider(
num_examples=10,
num_classes=10,
image_dim=32,
channels=3,
seed=42)
else:
raise ValueError("Invalid dataset")
estimator = estimator_builder.estimator(
data_provider=data_provider,
run_config=run_config,
hparams=hparams,
train_steps=FLAGS.train_steps)
train_spec = tf.estimator.TrainSpec(
input_fn=data_provider.get_input_fn(
partition="train",
mode=tf.estimator.ModeKeys.TRAIN,
batch_size=FLAGS.batch_size),
max_steps=FLAGS.train_steps
)
eval_spec = tf.estimator.EvalSpec(
input_fn=data_provider.get_input_fn(
partition="test",
mode=tf.estimator.ModeKeys.EVAL,
batch_size=FLAGS.batch_size),
steps=FLAGS.eval_steps,
start_delay_secs=10,
throttle_secs=1800
)
tf.logging.info("Training!")
tf.estimator.train_and_evaluate(estimator, train_spec, eval_spec)
tf.logging.info("Done training!")
def test_no_preprocess(self):
provider = cifar10.Provider(seed=4)
input_fn = provider.get_input_fn("train",
tf.contrib.learn.ModeKeys.TRAIN,
batch_size=3,
preprocess=False)
data, label = input_fn()
init = tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer())
with self.test_session() as sess:
sess.run(init)
data_result = sess.run(data["x"])
self.assertEqual((3, 32, 32, 3), data_result.shape)
self.assertAllEqual([148, 141, 174], data_result[0][0][0])
self.assertAllEqual([[5], [9], [3]], sess.run(label))
def test_basic_preprocess(self):
provider = cifar10.Provider(params_string="augmentation=basic", seed=4)
input_fn = provider.get_input_fn("train",
tf.contrib.learn.ModeKeys.TRAIN,
batch_size=3,
preprocess=True)
data, label = input_fn()
init = tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer())
with self.test_session() as sess:
sess.run(init)
data_result = sess.run(data["x"])
self.assertEqual((3, 32, 32, 3), data_result.shape)
self.assertAllEqual([0, 0, 0], data_result[0, 0, 0])
self.assertAlmostEqual(0.0, data_result[0, -1, 0, 0], places=3)
self.assertAllEqual([[5], [9], [3]], sess.run(label))