def testTrainSaveLoad(self):
"""Tests that insures you can save and reload a trained model."""
def _input_fn(num_epochs=None):
features = {
'age':
input_lib.limit_epochs(constant_op.constant([[.8], [.2],
[.1]]),
num_epochs=num_epochs),
'language':
sparse_tensor.SparseTensor(values=input_lib.limit_epochs(
['en', 'fr', 'zh'], num_epochs=num_epochs),
indices=[[0, 0], [0, 1], [2, 0]],
dense_shape=[3, 2])
}
return features, constant_op.constant([[1], [0], [0]],
dtype=dtypes.int32)
model_dir = tempfile.mkdtemp()
classifier = debug.DebugClassifier(
model_dir=model_dir,
n_classes=3,
config=run_config.RunConfig(tf_random_seed=1))
classifier.fit(input_fn=_input_fn, steps=5)
predict_input_fn = functools.partial(_input_fn, num_epochs=1)
predictions1 = classifier.predict_classes(input_fn=predict_input_fn)
del classifier
classifier2 = debug.DebugClassifier(
model_dir=model_dir,
n_classes=3,
config=run_config.RunConfig(tf_random_seed=1))
predictions2 = classifier2.predict_classes(input_fn=predict_input_fn)
self.assertEqual(list(predictions1), list(predictions2))
def test_uid_for_default_whitelist(self):
config = run_config_lib.RunConfig(
tf_random_seed=11,
save_summary_steps=12,
save_checkpoints_steps=13,
save_checkpoints_secs=14,
session_config=config_pb2.ConfigProto(allow_soft_placement=True),
keep_checkpoint_max=16,
keep_checkpoint_every_n_hours=17)
self.assertEqual(11, config.tf_random_seed)
self.assertEqual(12, config.save_summary_steps)
self.assertEqual(13, config.save_checkpoints_steps)
self.assertEqual(14, config.save_checkpoints_secs)
self.assertEqual(config_pb2.ConfigProto(allow_soft_placement=True),
config.session_config)
self.assertEqual(16, config.keep_checkpoint_max)
self.assertEqual(17, config.keep_checkpoint_every_n_hours)
new_config = run_config_lib.RunConfig(
tf_random_seed=21,
save_summary_steps=22,
save_checkpoints_steps=23,
save_checkpoints_secs=24,
session_config=config_pb2.ConfigProto(allow_soft_placement=False),
keep_checkpoint_max=26,
keep_checkpoint_every_n_hours=27)
self.assertEqual(config.uid(), new_config.uid())
# model_dir is not on the default whitelist.
self.assertNotEqual(config.uid(whitelist=[]),
new_config.uid(whitelist=[]))
new_config = new_config.replace(model_dir=ANOTHER_TEST_DIR)
self.assertNotEqual(config.uid(), new_config.uid())
def testTrainSaveLoad(self):
"""Tests that insures you can save and reload a trained model."""
def _input_fn(num_epochs=None):
features = {
'age':
input_lib.limit_epochs(constant_op.constant([[0.8], [0.15],
[0.]]),
num_epochs=num_epochs),
'language':
sparse_tensor.SparseTensor(values=input_lib.limit_epochs(
['en', 'fr', 'zh'], num_epochs=num_epochs),
indices=[[0, 0], [0, 1], [2, 0]],
dense_shape=[3, 2])
}
return features, constant_op.constant([1., 0., 0.2],
dtype=dtypes.float32)
model_dir = tempfile.mkdtemp()
regressor = debug.DebugRegressor(
model_dir=model_dir, config=run_config.RunConfig(tf_random_seed=1))
regressor.fit(input_fn=_input_fn, steps=5)
predict_input_fn = functools.partial(_input_fn, num_epochs=1)
predictions = list(regressor.predict_scores(input_fn=predict_input_fn))
del regressor
regressor2 = debug.DebugRegressor(
model_dir=model_dir, config=run_config.RunConfig(tf_random_seed=1))
predictions2 = list(
regressor2.predict_scores(input_fn=predict_input_fn))
self.assertAllClose(predictions, predictions2)
def test_checkpoint_and_export(self):
model_dir = tempfile.mkdtemp()
config = run_config_lib.RunConfig(save_checkpoints_steps=3)
est = dnn.DNNClassifier(
n_classes=3,
feature_columns=[
feature_column.real_valued_column('feature', dimension=4)
],
hidden_units=[3, 3],
model_dir=model_dir,
config=config)
exp_strategy = saved_model_export_utils.make_export_strategy(
est, 'export_input', exports_to_keep=None)
ex = experiment.Experiment(
est,
train_input_fn=test_data.iris_input_multiclass_fn,
eval_input_fn=test_data.iris_input_multiclass_fn,
export_strategies=(exp_strategy,),
train_steps=8,
checkpoint_and_export=True,
eval_delay_secs=0)
with test.mock.patch.object(ex, '_maybe_export'):
with test.mock.patch.object(ex, '_call_evaluate'):
ex.train_and_evaluate()
# Eval and export are called after steps 1, 4, 7, and 8 (after training
# is completed).
self.assertEqual(ex._maybe_export.call_count, 4)
self.assertEqual(ex._call_evaluate.call_count, 4)
def test_explicitly_specified_values(self):
cluster_spec = {
run_config_lib.TaskType.PS: ["localhost:9990"],
"my_job_name": ["localhost:9991", "localhost:9992", "localhost:0"]
}
tf_config = {
"cluster": cluster_spec,
"task": {
"type": run_config_lib.TaskType.WORKER,
"index": 2
}
}
with patch.dict("os.environ", {"TF_CONFIG": json.dumps(tf_config)}):
config = run_config_lib.RunConfig(
master="localhost:0", evaluation_master="localhost:9991")
self.assertEqual(config.master, "localhost:0")
self.assertEqual(config.task_id, 2)
self.assertEqual(config.num_ps_replicas, 1)
self.assertEqual(config.num_worker_replicas, 0)
self.assertEqual(config.cluster_spec,
server_lib.ClusterSpec(cluster_spec))
self.assertEqual(config.task_type, run_config_lib.TaskType.WORKER)
self.assertFalse(config.is_chief)
self.assertEqual(config.evaluation_master, "localhost:9991")
def testBostonDNN(self):
boston = base.load_boston()
feature_columns = [feature_column.real_valued_column("", dimension=13)]
regressor = dnn.DNNRegressor(
feature_columns=feature_columns,
hidden_units=[10, 20, 10],
config=run_config.RunConfig(tf_random_seed=1))
regressor.fit(boston.data,
boston.target,
steps=300,
batch_size=boston.data.shape[0])
weights = ([regressor.get_variable_value("dnn/hiddenlayer_0/weights")] +
[regressor.get_variable_value("dnn/hiddenlayer_1/weights")] +
[regressor.get_variable_value("dnn/hiddenlayer_2/weights")] +
[regressor.get_variable_value("dnn/logits/weights")])
self.assertEqual(weights[0].shape, (13, 10))
self.assertEqual(weights[1].shape, (10, 20))
self.assertEqual(weights[2].shape, (20, 10))
self.assertEqual(weights[3].shape, (10, 1))
biases = ([regressor.get_variable_value("dnn/hiddenlayer_0/biases")] +
[regressor.get_variable_value("dnn/hiddenlayer_1/biases")] +
[regressor.get_variable_value("dnn/hiddenlayer_2/biases")] +
[regressor.get_variable_value("dnn/logits/biases")])
self.assertEqual(biases[0].shape, (10,))
self.assertEqual(biases[1].shape, (20,))
self.assertEqual(biases[2].shape, (10,))
self.assertEqual(biases[3].shape, (1,))
def _compare_with_sklearn(self, cov_type):
# sklearn version.
iterations = 40
np.random.seed(5)
sklearn_assignments = np.asarray([0, 0, 1, 0, 0, 0, 1, 0, 0, 1])
sklearn_means = np.asarray([[144.83417719, 254.20130341],
[274.38754816, 353.16074346]])
sklearn_covs = np.asarray([[[395.0081194, -4.50389512],
[-4.50389512, 408.27543989]],
[[385.17484203, -31.27834935],
[-31.27834935, 391.74249925]]])
# skflow version.
gmm = gmm_lib.GMM(self.num_centers,
initial_clusters=self.initial_means,
covariance_type=cov_type,
config=run_config.RunConfig(tf_random_seed=2))
gmm.fit(input_fn=self.input_fn(), steps=iterations)
points = self.points[:10, :]
skflow_assignments = []
for item in gmm.predict_assignments(
input_fn=self.input_fn(points=points, batch_size=10)):
skflow_assignments.append(item)
self.assertAllClose(sklearn_assignments,
np.ravel(skflow_assignments).astype(int))
self.assertAllClose(sklearn_means, gmm.clusters())
if cov_type == 'full':
self.assertAllClose(sklearn_covs, gmm.covariances(), rtol=0.01)
else:
for d in [0, 1]:
self.assertAllClose(np.diag(sklearn_covs[d]),
gmm.covariances()[d, :],
rtol=0.01)
def testExport(self):
"""Tests export model for servo."""
def input_fn():
return {
'age':
constant_op.constant([1]),
'language':
sparse_tensor.SparseTensor(values=['english'],
indices=[[0, 0]],
dense_shape=[1, 1])
}, constant_op.constant([[1]])
language = feature_column.sparse_column_with_hash_bucket(
'language', 100)
feature_columns = [
feature_column.real_valued_column('age'),
feature_column.embedding_column(language, dimension=1)
]
classifier = debug.DebugClassifier(config=run_config.RunConfig(
tf_random_seed=1))
classifier.fit(input_fn=input_fn, steps=5)
def default_input_fn(unused_estimator, examples):
return feature_column_ops.parse_feature_columns_from_examples(
examples, feature_columns)
export_dir = tempfile.mkdtemp()
classifier.export(export_dir, input_fn=default_input_fn)
def testIrisDNN(self):
iris = base.load_iris()
feature_columns = [feature_column.real_valued_column("", dimension=4)]
classifier = dnn.DNNClassifier(
feature_columns=feature_columns,
hidden_units=[10, 20, 10],
n_classes=3,
config=run_config.RunConfig(tf_random_seed=1))
classifier.fit(iris.data, iris.target, max_steps=200)
variable_names = classifier.get_variable_names()
self.assertEqual(
classifier.get_variable_value("dnn/hiddenlayer_0/weights").shape,
(4, 10))
self.assertEqual(
classifier.get_variable_value("dnn/hiddenlayer_1/weights").shape,
(10, 20))
self.assertEqual(
classifier.get_variable_value("dnn/hiddenlayer_2/weights").shape,
(20, 10))
self.assertEqual(
classifier.get_variable_value("dnn/logits/weights").shape, (10, 3))
self.assertIn("dnn/hiddenlayer_0/biases", variable_names)
self.assertIn("dnn/hiddenlayer_1/biases", variable_names)
self.assertIn("dnn/hiddenlayer_2/biases", variable_names)
self.assertIn("dnn/logits/biases", variable_names)
def testTrainEvaluateInferDoesNotThrowErrorWithDnnInput(self):
head_fn = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
learner_config = learner_pb2.LearnerConfig()
learner_config.num_classes = 2
learner_config.constraints.max_tree_depth = 3
model_dir = tempfile.mkdtemp()
config = run_config.RunConfig()
est = estimator.CoreDNNBoostedTreeCombinedEstimator(
head=head_fn,
dnn_hidden_units=[1],
dnn_feature_columns=[core_feature_column.numeric_column("x")],
tree_learner_config=learner_config,
num_trees=1,
tree_examples_per_layer=3,
model_dir=model_dir,
config=config,
dnn_steps_to_train=10,
dnn_input_layer_to_tree=True,
tree_feature_columns=[])
# Train for a few steps.
est.train(input_fn=_train_input_fn, steps=1000)
res = est.evaluate(input_fn=_eval_input_fn, steps=1)
self.assertLess(0.5, res["auc"])
est.predict(input_fn=_eval_input_fn)
def testRankingDontThrowExceptionForForEstimator(self):
learner_config = learner_pb2.LearnerConfig()
learner_config.num_classes = 2
learner_config.constraints.max_tree_depth = 1
model_dir = tempfile.mkdtemp()
config = run_config.RunConfig()
head_fn = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
est = estimator.CoreGradientBoostedDecisionTreeRanker(
head=head_fn,
learner_config=learner_config,
num_trees=1,
examples_per_layer=3,
model_dir=model_dir,
config=config,
feature_columns=[
core_feature_column.numeric_column("f1"),
core_feature_column.numeric_column("f2")
],
ranking_model_pair_keys=("a", "b"))
# Train for a few steps.
est.train(input_fn=_ranking_train_input_fn, steps=1000)
est.evaluate(input_fn=_ranking_train_input_fn, steps=1)
est.predict(input_fn=_infer_ranking_train_input_fn)
def testFitAndEvaluateMultiClassFullDontThrowException(self):
learner_config = learner_pb2.LearnerConfig()
learner_config.num_classes = 3
learner_config.constraints.max_tree_depth = 1
learner_config.multi_class_strategy = (
learner_pb2.LearnerConfig.FULL_HESSIAN)
model_dir = tempfile.mkdtemp()
config = run_config.RunConfig()
classifier = estimator.GradientBoostedDecisionTreeClassifier(
learner_config=learner_config,
n_classes=learner_config.num_classes,
num_trees=1,
examples_per_layer=7,
model_dir=model_dir,
config=config,
center_bias=False,
feature_columns=[contrib_feature_column.real_valued_column("x")])
classifier.fit(input_fn=_multiclass_train_input_fn, steps=100)
classifier.evaluate(input_fn=_eval_input_fn, steps=1)
classifier.export(self._export_dir_base)
result_iter = classifier.predict(input_fn=_eval_input_fn)
for prediction_dict in result_iter:
self.assertTrue("classes" in prediction_dict)
def test_fail_output_dir_and_run_config_are_both_set(self):
with self.assertRaisesRegexp(
ValueError, _CANNOT_SET_BOTH_OUTPUT_DIR_AND_CONFIG_MSG):
learn_runner.run(build_experiment,
output_dir=_MODIR_DIR,
schedule="simple_task",
run_config=run_config_lib.RunConfig())
def _experiment_fn(run_config, hparams):
del run_config, hparams # unused.
# Explicitly use a new run_config.
new_config = run_config_lib.RunConfig(model_dir=_MODIR_DIR +
"/123")
return TestExperiment(config=new_config)
def test_run_with_explicit_local_run(self):
run_config = run_config_lib.RunConfig(model_dir=_MODIR_DIR)
self.assertEqual(
"local_run-" + _MODIR_DIR,
learn_runner.run(build_experiment_for_run_config,
run_config=run_config,
schedule="local_run"))
def test_fail_invalid_hparams_type(self):
run_config = run_config_lib.RunConfig(model_dir=_MODIR_DIR)
with self.assertRaisesRegexp(ValueError, _INVALID_HPARAMS_ERR_MSG):
learn_runner.run(build_experiment_for_run_config,
run_config=run_config,
schedule="local_run",
hparams=["hparams"])
def testCustomMetrics(self):
"""Tests custom evaluation metrics."""
def _input_fn(num_epochs=None):
# Create 4 rows, one of them (y = x), three of them (y=Not(x))
labels = constant_op.constant([[1], [0], [0], [0]])
features = {
'x':
input_lib.limit_epochs(array_ops.ones(shape=[4, 1],
dtype=dtypes.float32),
num_epochs=num_epochs),
}
return features, labels
def _my_metric_op(predictions, labels):
# For the case of binary classification, the 2nd column of "predictions"
# denotes the model predictions.
labels = math_ops.cast(labels, dtypes.float32)
predictions = array_ops.strided_slice(predictions, [0, 1], [-1, 2],
end_mask=1)
labels = math_ops.cast(labels, predictions.dtype)
return math_ops.reduce_sum(math_ops.multiply(predictions, labels))
classifier = debug.DebugClassifier(config=run_config.RunConfig(
tf_random_seed=1))
classifier.fit(input_fn=_input_fn, steps=5)
scores = classifier.evaluate(
input_fn=_input_fn,
steps=5,
metrics={
'my_accuracy':
MetricSpec(metric_fn=metric_ops.streaming_accuracy,
prediction_key='classes'),
'my_precision':
MetricSpec(metric_fn=metric_ops.streaming_precision,
prediction_key='classes'),
'my_metric':
MetricSpec(metric_fn=_my_metric_op,
prediction_key='probabilities')
})
self.assertTrue(
set(['loss', 'my_accuracy', 'my_precision',
'my_metric']).issubset(set(scores.keys())))
predict_input_fn = functools.partial(_input_fn, num_epochs=1)
predictions = np.array(
list(classifier.predict_classes(input_fn=predict_input_fn)))
self.assertEqual(_sklearn.accuracy_score([1, 0, 0, 0], predictions),
scores['my_accuracy'])
# Test the case where the 2nd element of the key is neither "classes" nor
# "probabilities".
with self.assertRaisesRegexp(KeyError, 'bad_type'):
classifier.evaluate(input_fn=_input_fn,
steps=5,
metrics={
'bad_name':
MetricSpec(
metric_fn=metric_ops.streaming_auc,
prediction_key='bad_type')
})
def test_run_with_custom_schedule(self):
run_config = run_config_lib.RunConfig(model_dir=_MODIR_DIR)
self.assertEqual(
"simple_task, default=None.",
learn_runner.run(build_experiment_for_run_config,
run_config=run_config,
schedule="simple_task"))
def testFitAndEvaluateMultiClassFullDontThrowException(self):
n_classes = 3
learner_config = learner_pb2.LearnerConfig()
learner_config.num_classes = n_classes
learner_config.constraints.max_tree_depth = 1
learner_config.multi_class_strategy = (
learner_pb2.LearnerConfig.FULL_HESSIAN)
head_fn = estimator.core_multiclass_head(n_classes=n_classes)
model_dir = tempfile.mkdtemp()
config = run_config.RunConfig()
classifier = estimator.CoreGradientBoostedDecisionTreeEstimator(
learner_config=learner_config,
head=head_fn,
num_trees=1,
center_bias=False,
examples_per_layer=7,
model_dir=model_dir,
config=config,
feature_columns=[core_feature_column.numeric_column("x")])
classifier.train(input_fn=_multiclass_train_input_fn, steps=100)
classifier.evaluate(input_fn=_multiclass_train_input_fn, steps=1)
classifier.predict(input_fn=_eval_input_fn)
def test_run_std_server(self, mock_server):
# Arrange.
tf_config = {
'cluster': self._cluster_spec(),
'task': {
'type': run_config_lib.TaskType.PS,
'index': 1
}
}
with test.mock.patch.dict('os.environ',
{'TF_CONFIG': json.dumps(tf_config)}):
config = run_config_lib.RunConfig(
master='host2:2222',
num_cores=15,
gpu_memory_fraction=0.314,)
for est in self._estimators_for_tests(config):
ex = experiment.Experiment(
est, train_input_fn='train_input', eval_input_fn='eval_input')
# Act.
ex.run_std_server()
# Assert.
mock_server.assert_has_calls(
[test.mock.call().start(), test.mock.call().join()])
def testTrainWithWeights(self):
"""Tests training with given weight column."""
def _input_fn_train():
# Create 4 rows, one of them (y = x), three of them (y=Not(x))
# First row has more weight than others. Model should fit (y=x) better
# than (y=Not(x)) due to the relative higher weight of the first row.
labels = constant_op.constant([[1.], [0.], [0.], [0.]])
features = {
'x': array_ops.ones(shape=[4, 1], dtype=dtypes.float32),
'w': constant_op.constant([[100.], [3.], [2.], [2.]])
}
return features, labels
def _input_fn_eval():
# Create 4 rows (y = x)
labels = constant_op.constant([[1.], [1.], [1.], [1.]])
features = {
'x': array_ops.ones(shape=[4, 1], dtype=dtypes.float32),
'w': constant_op.constant([[1.], [1.], [1.], [1.]])
}
return features, labels
regressor = debug.DebugRegressor(
weight_column_name='w',
config=run_config.RunConfig(tf_random_seed=1))
regressor.fit(input_fn=_input_fn_train, steps=5)
scores = regressor.evaluate(input_fn=_input_fn_eval, steps=1)
self.assertIn('loss', scores)
def testLossWithWeights(self):
"""Tests loss calculation with weights."""
def _input_fn_train():
# 4 rows with equal weight, one of them (y = x), three of them (y=Not(x))
# The algorithm should learn (y = 0.25).
labels = constant_op.constant([[1.], [0.], [0.], [0.]])
features = {
'x': array_ops.ones(shape=[4, 1], dtype=dtypes.float32),
'w': constant_op.constant([[1.], [1.], [1.], [1.]])
}
return features, labels
def _input_fn_eval():
# 4 rows, with different weights.
labels = constant_op.constant([[1.], [0.], [0.], [0.]])
features = {
'x': array_ops.ones(shape=[4, 1], dtype=dtypes.float32),
'w': constant_op.constant([[7.], [1.], [1.], [1.]])
}
return features, labels
regressor = debug.DebugRegressor(
weight_column_name='w',
config=run_config.RunConfig(tf_random_seed=1))
regressor.fit(input_fn=_input_fn_train, steps=5)
scores = regressor.evaluate(input_fn=_input_fn_eval, steps=1)
self.assertIn('loss', scores)
def testLearnSineFunction(self):
"""Tests learning a sine function."""
batch_size = 8
num_unroll = 5
sequence_length = 64
train_steps = 250
eval_steps = 20
num_rnn_layers = 1
num_units = [4] * num_rnn_layers
learning_rate = 0.3
loss_threshold = 0.035
def get_sin_input_fn(sequence_length, increment, seed=None):
def input_fn():
start = random_ops.random_uniform((),
minval=0,
maxval=(np.pi * 2.0),
dtype=dtypes.float32,
seed=seed)
sin_curves = math_ops.sin(
math_ops.linspace(start, (sequence_length - 1) * increment,
sequence_length + 1))
inputs = array_ops.slice(sin_curves, [0], [sequence_length])
labels = array_ops.slice(sin_curves, [1], [sequence_length])
return {'inputs': inputs}, labels
return input_fn
seq_columns = [
feature_column.real_valued_column('inputs', dimension=1)
]
config = run_config.RunConfig(tf_random_seed=1234)
dropout_keep_probabilities = [0.9] * (num_rnn_layers + 1)
sequence_estimator = ssre.StateSavingRnnEstimator(
constants.ProblemType.LINEAR_REGRESSION,
num_units=num_units,
cell_type='lstm',
num_unroll=num_unroll,
batch_size=batch_size,
sequence_feature_columns=seq_columns,
learning_rate=learning_rate,
dropout_keep_probabilities=dropout_keep_probabilities,
config=config,
queue_capacity=2 * batch_size,
seed=1234)
train_input_fn = get_sin_input_fn(sequence_length,
np.pi / 32,
seed=1234)
eval_input_fn = get_sin_input_fn(sequence_length,
np.pi / 32,
seed=4321)
sequence_estimator.fit(input_fn=train_input_fn, steps=train_steps)
loss = sequence_estimator.evaluate(input_fn=eval_input_fn,
steps=eval_steps)['loss']
self.assertLess(
loss, loss_threshold,
'Loss should be less than {}; got {}'.format(loss_threshold, loss))
def test_no_schedule_and_non_distributed_runs_train_and_evaluate(self):
tf_config = {"cluster": build_non_distributed_cluster_spec()}
with patch.dict("os.environ", {"TF_CONFIG": json.dumps(tf_config)}):
config = run_config_lib.RunConfig()
self.assertEqual(
"train_and_evaluate-" + _MODIR_DIR,
learn_runner.run(build_experiment_fn_for_output_dir(config),
output_dir=_MODIR_DIR))
def test_model_dir_fail_if_constructor_value_mismatch_tf_config(self):
tf_config = {"model_dir": TEST_DIR}
with patch.dict("os.environ", {"TF_CONFIG": json.dumps(tf_config)}):
with self.assertRaisesRegexp(
ValueError,
"`model_dir` provided in RunConfig .* must have "
"the same value .* in TF_CONFIG"):
run_config_lib.RunConfig(model_dir=TEST_DIR + "/sub_dir")
def testRegression_MatrixData(self):
"""Tests regression using matrix data as input."""
regressor = debug.DebugRegressor(config=run_config.RunConfig(
tf_random_seed=1))
input_fn = test_data.iris_input_logistic_fn
regressor.fit(input_fn=input_fn, steps=200)
scores = regressor.evaluate(input_fn=input_fn, steps=1)
self.assertIn('loss', scores)
def testMultiRNNState(self):
"""Test that state flattening/reconstruction works for `MultiRNNCell`."""
batch_size = 11
sequence_length = 16
train_steps = 5
cell_sizes = [4, 8, 7]
learning_rate = 0.1
def get_shift_input_fn(batch_size, sequence_length, seed=None):
def input_fn():
random_sequence = random_ops.random_uniform(
[batch_size, sequence_length + 1],
0,
2,
dtype=dtypes.int32,
seed=seed)
labels = array_ops.slice(random_sequence, [0, 0],
[batch_size, sequence_length])
inputs = array_ops.expand_dims(
math_ops.cast(
array_ops.slice(random_sequence, [0, 1],
[batch_size, sequence_length]),
dtypes.float32), 2)
input_dict = {
dynamic_rnn_estimator._get_state_name(i): random_ops.random_uniform(
[batch_size, cell_size], seed=((i + 1) * seed))
for i, cell_size in enumerate([4, 4, 8, 8, 7, 7])
}
input_dict['inputs'] = inputs
return input_dict, labels
return input_fn
seq_columns = [feature_column.real_valued_column('inputs', dimension=1)]
config = run_config.RunConfig(tf_random_seed=21212)
cell_type = 'lstm'
sequence_estimator = dynamic_rnn_estimator.DynamicRnnEstimator(
problem_type=constants.ProblemType.CLASSIFICATION,
prediction_type=rnn_common.PredictionType.MULTIPLE_VALUE,
num_classes=2,
num_units=cell_sizes,
sequence_feature_columns=seq_columns,
cell_type=cell_type,
learning_rate=learning_rate,
config=config,
predict_probabilities=True)
train_input_fn = get_shift_input_fn(batch_size, sequence_length, seed=12321)
eval_input_fn = get_shift_input_fn(batch_size, sequence_length, seed=32123)
sequence_estimator.fit(input_fn=train_input_fn, steps=train_steps)
prediction_dict = sequence_estimator.predict(
input_fn=eval_input_fn, as_iterable=False)
for i, state_size in enumerate([4, 4, 8, 8, 7, 7]):
state_piece = prediction_dict[dynamic_rnn_estimator._get_state_name(i)]
self.assertListEqual(list(state_piece.shape), [batch_size, state_size])
def test_defaults_with_no_tf_config(self):
config = run_config_lib.RunConfig()
self.assertEqual(config.master, "")
self.assertEqual(config.task_id, 0)
self.assertEqual(config.num_ps_replicas, 0)
self.assertEqual(config.cluster_spec, {})
self.assertIsNone(config.task_type)
self.assertTrue(config.is_chief)
self.assertEqual(config.evaluation_master, "")
def testLogisticRegression_MatrixData(self):
"""Tests binary classification using matrix data as input."""
classifier = debug.DebugClassifier(config=run_config.RunConfig(
tf_random_seed=1))
input_fn = test_data.iris_input_logistic_fn
classifier.fit(input_fn=input_fn, steps=5)
scores = classifier.evaluate(input_fn=input_fn, steps=1)
self._assertInRange(0.0, 1.0, scores['accuracy'])
self.assertIn('loss', scores)
def testLearnSineFunction(self):
"""Tests learning a sine function."""
batch_size = 8
sequence_length = 64
train_steps = 200
eval_steps = 20
cell_size = [4]
learning_rate = 0.1
loss_threshold = 0.02
def get_sin_input_fn(batch_size, sequence_length, increment, seed=None):
def _sin_fn(x):
ranger = math_ops.linspace(
array_ops.reshape(x[0], []), (sequence_length - 1) * increment,
sequence_length + 1)
return math_ops.sin(ranger)
def input_fn():
starts = random_ops.random_uniform(
[batch_size], maxval=(2 * np.pi), seed=seed)
sin_curves = map_fn.map_fn(
_sin_fn, (starts,), dtype=dtypes.float32)
inputs = array_ops.expand_dims(
array_ops.slice(sin_curves, [0, 0], [batch_size, sequence_length]),
2)
labels = array_ops.slice(sin_curves, [0, 1],
[batch_size, sequence_length])
return {'inputs': inputs}, labels
return input_fn
seq_columns = [
feature_column.real_valued_column(
'inputs', dimension=cell_size[0])
]
config = run_config.RunConfig(tf_random_seed=1234)
sequence_estimator = dynamic_rnn_estimator.DynamicRnnEstimator(
problem_type=constants.ProblemType.LINEAR_REGRESSION,
prediction_type=rnn_common.PredictionType.MULTIPLE_VALUE,
num_units=cell_size,
sequence_feature_columns=seq_columns,
learning_rate=learning_rate,
dropout_keep_probabilities=[0.9, 0.9],
config=config)
train_input_fn = get_sin_input_fn(
batch_size, sequence_length, np.pi / 32, seed=1234)
eval_input_fn = get_sin_input_fn(
batch_size, sequence_length, np.pi / 32, seed=4321)
sequence_estimator.fit(input_fn=train_input_fn, steps=train_steps)
loss = sequence_estimator.evaluate(
input_fn=eval_input_fn, steps=eval_steps)['loss']
self.assertLess(loss, loss_threshold,
'Loss should be less than {}; got {}'.format(loss_threshold,
loss))
