def testLearnMean(self):
"""Test learning to calculate a mean."""
batch_size = 16
sequence_length = 3
train_steps = 200
eval_steps = 20
cell_type = 'basic_rnn'
cell_size = 8
optimizer_type = 'Momentum'
learning_rate = 0.1
momentum = 0.9
loss_threshold = 0.1
def get_mean_input_fn(batch_size, sequence_length, seed=None):
def input_fn():
# Create examples by choosing 'centers' and adding uniform noise.
centers = math_ops.matmul(
random_ops.random_uniform([batch_size, 1],
-0.75,
0.75,
dtype=dtypes.float32,
seed=seed),
array_ops.ones([1, sequence_length]))
noise = random_ops.random_uniform(
[batch_size, sequence_length],
-0.25,
0.25,
dtype=dtypes.float32,
seed=seed)
sequences = centers + noise
inputs = array_ops.expand_dims(sequences, 2)
labels = math_ops.reduce_mean(sequences, reduction_indices=[1])
return {'inputs': inputs}, labels
return input_fn
seq_columns = [
feature_column.real_valued_column('inputs', dimension=cell_size)
]
config = run_config.RunConfig(tf_random_seed=6)
sequence_regressor = dynamic_rnn_estimator.single_value_rnn_regressor(
num_units=cell_size,
sequence_feature_columns=seq_columns,
cell_type=cell_type,
optimizer_type=optimizer_type,
learning_rate=learning_rate,
momentum=momentum,
config=config)
train_input_fn = get_mean_input_fn(batch_size, sequence_length, 121)
eval_input_fn = get_mean_input_fn(batch_size, sequence_length, 212)
sequence_regressor.fit(input_fn=train_input_fn, steps=train_steps)
evaluation = sequence_regressor.evaluate(input_fn=eval_input_fn,
steps=eval_steps)
loss = evaluation['loss']
self.assertLess(
loss, loss_threshold,
'Loss should be less than {}; got {}'.format(loss_threshold, loss))
def testLearnMean(self):
"""Test learning to calculate a mean."""
batch_size = 16
sequence_length = 3
train_steps = 200
eval_steps = 20
cell_type = 'basic_rnn'
cell_size = 8
optimizer_type = 'Momentum'
learning_rate = 0.1
momentum = 0.9
loss_threshold = 0.1
def get_mean_input_fn(batch_size, sequence_length, seed=None):
def input_fn():
# Create examples by choosing 'centers' and adding uniform noise.
centers = math_ops.matmul(
random_ops.random_uniform(
[batch_size, 1], -0.75, 0.75, dtype=dtypes.float32, seed=seed),
array_ops.ones([1, sequence_length]))
noise = random_ops.random_uniform(
[batch_size, sequence_length],
-0.25,
0.25,
dtype=dtypes.float32,
seed=seed)
sequences = centers + noise
inputs = array_ops.expand_dims(sequences, 2)
labels = math_ops.reduce_mean(sequences, reduction_indices=[1])
return {'inputs': inputs}, labels
return input_fn
seq_columns = [
feature_column.real_valued_column(
'inputs', dimension=cell_size)
]
config = run_config.RunConfig(tf_random_seed=6)
sequence_regressor = dynamic_rnn_estimator.single_value_rnn_regressor(
num_units=cell_size,
sequence_feature_columns=seq_columns,
cell_type=cell_type,
optimizer_type=optimizer_type,
learning_rate=learning_rate,
momentum=momentum,
config=config)
train_input_fn = get_mean_input_fn(batch_size, sequence_length, 121)
eval_input_fn = get_mean_input_fn(batch_size, sequence_length, 212)
sequence_regressor.fit(input_fn=train_input_fn, steps=train_steps)
evaluation = sequence_regressor.evaluate(
input_fn=eval_input_fn, steps=eval_steps)
loss = evaluation['loss']
self.assertLess(loss, loss_threshold,
'Loss should be less than {}; got {}'.format(loss_threshold,
loss))
def testLearnMean(self):
"""Test that `_SequenceRegressor` can learn to calculate a mean."""
batch_size = 16
sequence_length = 3
train_steps = 200
eval_steps = 20
cell_type = "basic_rnn"
cell_size = 8
optimizer_type = "Momentum"
learning_rate = 0.5
momentum = 0.9
loss_threshold = 0.1
def get_mean_input_fn(batch_size, sequence_length, seed=None):
def input_fn():
# Create examples by choosing 'centers' and adding uniform noise.
centers = tf.matmul(
tf.random_uniform([batch_size, 1], -0.75, 0.75, dtype=tf.float32, seed=seed),
tf.ones([1, sequence_length]),
)
noise = tf.random_uniform([batch_size, sequence_length], -0.25, 0.25, dtype=tf.float32, seed=seed)
sequences = centers + noise
inputs = tf.expand_dims(sequences, 2)
labels = tf.reduce_mean(sequences, reduction_indices=[1])
return {"inputs": inputs}, labels
return input_fn
config = tf.contrib.learn.RunConfig(tf_random_seed=6)
sequence_regressor = dynamic_rnn_estimator.single_value_rnn_regressor(
num_units=cell_size,
cell_type=cell_type,
optimizer_type=optimizer_type,
learning_rate=learning_rate,
momentum=momentum,
config=config,
)
train_input_fn = get_mean_input_fn(batch_size, sequence_length, 121)
eval_input_fn = get_mean_input_fn(batch_size, sequence_length, 212)
sequence_regressor.fit(input_fn=train_input_fn, steps=train_steps)
evaluation = sequence_regressor.evaluate(input_fn=eval_input_fn, steps=eval_steps)
loss = evaluation["loss"]
self.assertLess(loss, loss_threshold, "Loss should be less than {}; got {}".format(loss_threshold, loss))
