def test_deprecated_illegal_args(self):
instructions = "This is how you update..."
with self.assertRaisesRegexp(ValueError, "date"):
deprecation.deprecated(None, instructions)
with self.assertRaisesRegexp(ValueError, "date"):
deprecation.deprecated("", instructions)
with self.assertRaisesRegexp(ValueError, "YYYY-MM-DD"):
deprecation.deprecated("07-04-2016", instructions)
date = "2016-07-04"
with self.assertRaisesRegexp(ValueError, "instructions"):
deprecation.deprecated(date, None)
with self.assertRaisesRegexp(ValueError, "instructions"):
deprecation.deprecated(date, "")
def test_deprecated_namedtuple(self, mock_warning):
date = "2016-07-04"
instructions = "This is how you update..."
mytuple = deprecation.deprecated(date, instructions, warn_once=True)(
collections.namedtuple("my_tuple", ["field1", "field2"]))
mytuple(1, 2)
self.assertEqual(1, mock_warning.call_count)
mytuple(3, 4)
self.assertEqual(1, mock_warning.call_count)
self.assertIn("IS DEPRECATED", mytuple.__doc__)
from tensorflow.python.training import monitored_session
from tensorflow.python.training import queue_runner
from tensorflow.python.training import saver as tf_saver
from tensorflow.python.training import session_manager as session_manager_lib
from tensorflow.python.training import summary_io
from tensorflow.python.training import supervisor as tf_supervisor
from tensorflow.python.util.deprecation import deprecated
# Singleton for SummaryWriter per logdir folder.
_SUMMARY_WRITERS = {}
# Lock protecting _SUMMARY_WRITERS
_summary_writer_lock = threading.Lock()
_graph_action_deprecation = deprecated(
'2017-02-15',
'graph_actions.py will be deleted. Use tf.train.* utilities instead. '
'You can use learn/estimators/estimator.py as an example.')
@_graph_action_deprecation
def clear_summary_writers():
"""Clear cached summary writers. Currently only used for unit tests."""
return summary_io.SummaryWriterCache.clear()
def get_summary_writer(logdir):
"""Returns single SummaryWriter per logdir in current run.
Args:
logdir: str, folder to write summaries.
m = np.max(a.shape[-2:].as_list())
else:
m = tf.reduce_max(tf.shape(a)[-2:])
eps = np.finfo(dtype_util.as_numpy_dtype(a.dtype)).eps
tol = (eps * tf.cast(m, a.dtype) *
tf.reduce_max(s, axis=-1, keepdims=True))
return tf.reduce_sum(tf.cast(s > tol, tf.int32), axis=-1)
try:
matrix_rank = tf.linalg.matrix_rank
except AttributeError:
pass
matrix_rank = deprecation.deprecated(
'2019-10-01',
'tfp.math.matrix_rank is deprecated. Use tf.linalg.matrix_rank instead',
warn_once=True)(matrix_rank)
def cholesky_concat(chol, cols, name=None):
"""Concatenates `chol @ chol.T` with additional rows and columns.
This operation is conceptually identical to:
```python
def cholesky_concat_slow(chol, cols): # cols shaped (n + m) x m = z x m
mat = tf.matmul(chol, chol, adjoint_b=True) # batch of n x n
# Concat columns.
mat = tf.concat([mat, cols[..., :tf.shape(mat)[-2], :]], axis=-1) # n x z
# Concat rows.
mat = tf.concat([mat, tf.linalg.matrix_transpose(cols)], axis=-2) # z x z
return tf.linalg.cholesky(mat)
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Functions for computing statistics of samples."""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from tensorflow_probability.python import stats
from tensorflow.python.util import deprecation
__all__ = [
"auto_correlation",
"percentile",
]
auto_correlation_deprecator = deprecation.deprecated(
"2018-10-01",
"auto_correlation is moved to the `stats` namespace. Access it via: "
"`tfp.stats.auto_correlation`.",
warn_once=True)
auto_correlation = auto_correlation_deprecator(stats.auto_correlation)
percentile_deprecator = deprecation.deprecated(
"2018-10-01",
"percentile is moved to the `stats` namespace. Access it via: "
"`tfp.stats.percentile`.",
warn_once=True)
percentile = percentile_deprecator(stats.percentile)
scale = tf.convert_to_tensor(value=scale,
name="softplus_scale",
dtype=dtype)
super(_InverseGammaWithSoftplusConcentrationScale, self).__init__(
concentration=tf.nn.softplus(concentration,
name="softplus_concentration"),
scale=tf.nn.softplus(scale, name="softplus_scale"),
validate_args=validate_args,
allow_nan_stats=allow_nan_stats,
name=name)
self._parameters = parameters
_rate_deprecator = deprecation.deprecated(
"2019-06-05",
"InverseGammaWithSoftplusConcentrationRate is deprecated, use "
"InverseGamma(concentration=tf.nn.softplus(concentration), "
"scale=tf.nn.softplus(scale)) instead.",
warn_once=True)
# pylint: disable=invalid-name
InverseGammaWithSoftplusConcentrationRate = _rate_deprecator(
_InverseGammaWithSoftplusConcentrationScale)
_scale_deprecator = deprecation.deprecated(
"2019-06-05",
"InverseGammaWithSoftplusConcentrationScale is deprecated, use "
"InverseGamma(concentration=tf.nn.softplus(concentration), "
"scale=tf.nn.softplus(scale)) instead.",
warn_once=True)
InverseGammaWithSoftplusConcentrationScale = _scale_deprecator(
_InverseGammaWithSoftplusConcentrationScale)
from tensorflow_probability.python.math.root_search import secant_root
from tensorflow_probability.python.math.scan_associative import scan_associative
from tensorflow_probability.python.math.sparse import dense_to_sparse
from tensorflow_probability.python.math.special import erfcinv
from tensorflow_probability.python.math.special import lambertw
from tensorflow_probability.python.math.special import lambertw_winitzki_approx
from tensorflow_probability.python.math.special import lbeta
from tensorflow_probability.python.math.special import log_gamma_correction
from tensorflow_probability.python.math.special import log_gamma_difference
from tensorflow_probability.python.math.special import round_exponential_bump_function
from tensorflow_probability.python.random import rademacher as random_rademacher
from tensorflow_probability.python.random import rayleigh as random_rayleigh
from tensorflow.python.util import deprecation # pylint: disable=g-direct-tensorflow-import
random_rademacher = deprecation.deprecated(
'2020-09-20', 'Use tfp.random.rademacher')(random_rademacher)
random_rayleigh = deprecation.deprecated(
'2020-09-20', 'Use tfp.random.rayleigh')(random_rayleigh)
_allowed_symbols = [
'round_exponential_bump_function',
'batch_interp_regular_1d_grid',
'batch_interp_regular_nd_grid',
'bessel_iv_ratio',
'bessel_ive',
'bessel_kve',
'cholesky_concat',
'cholesky_update',
'clip_by_value_preserve_gradient',
'custom_gradient',
'dense_to_sparse',
from tensorflow_probability.python.internal.reparameterization import FULLY_REPARAMETERIZED
from tensorflow_probability.python.internal.reparameterization import NOT_REPARAMETERIZED
from tensorflow_probability.python.internal.reparameterization import ReparameterizationType
# Deprecated:
from tensorflow_probability.python.experimental.substrates.numpy.math.generic import reduce_weighted_logsumexp as _reduce_weighted_logsumexp
from tensorflow_probability.python.experimental.substrates.numpy.math.generic import softplus_inverse as _softplus_inverse
from tensorflow_probability.python.experimental.substrates.numpy.math.linalg import fill_triangular as _fill_triangular
from tensorflow_probability.python.experimental.substrates.numpy.math.linalg import fill_triangular_inverse as _fill_triangular_inverse
from tensorflow_probability.python.experimental.substrates.numpy.util.seed_stream import SeedStream as _SeedStream
# Module management:
from tensorflow_probability.python.experimental.substrates.numpy.distributions.kullback_leibler import augment_kl_xent_docs
from tensorflow.python.util import deprecation # pylint: disable=g-direct-tensorflow-import
_deprecated = deprecation.deprecated('2019-10-01',
'This function has moved to `tfp.math`.')
fill_triangular = _deprecated(_fill_triangular)
fill_triangular_inverse = _deprecated(_fill_triangular_inverse)
softplus_inverse = _deprecated(_softplus_inverse)
reduce_weighted_logsumexp = _deprecated(_reduce_weighted_logsumexp)
class SeedStream(_SeedStream):
__init__ = deprecation.deprecated(
'2019-10-01',
'SeedStream has moved to `tfp.util.SeedStream`.')(_SeedStream.__init__)
class SeedStream(_SeedStream):
__init__ = deprecation.deprecated(
'2019-10-01',
'SeedStream has moved to `tfp.util.SeedStream`.')(_SeedStream.__init__)
Deprecated: please see the new location of this module at `tfx.types.artifact`
and `tfx.types.artifact_utils`.
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from typing import Dict, List, Text
from tensorflow.python.util import deprecation # pylint: disable=g-direct-tensorflow-import
from tfx.types import artifact_utils
from tfx.types.artifact import Artifact
TfxType = deprecation.deprecated( # pylint: disable=invalid-name
None,
'tfx.utils.types.TfxType has been renamed to tfx.types.Artifact as of '
'TFX 0.14.0.')(Artifact)
TfxArtifact = deprecation.deprecated( # pylint: disable=invalid-name
None,
'tfx.utils.types.TfxArtifact has been renamed to tfx.types.Artifact as of '
'TFX 0.14.0.')(Artifact)
@deprecation.deprecated(
None, 'tfx.utils.types.parse_tfx_type_dict has been renamed to '
'tfx.types.artifact_utils.parse_artifact_dict as of TFX 0.14.0.')
def parse_tfx_type_dict(json_str: Text) -> Dict[Text, List[Artifact]]:
return artifact_utils.parse_artifact_dict(json_str)
name="InverseGammaWithSoftplusConcentrationScale"):
if rate is not None:
scale = rate
parameters = dict(locals())
with tf.name_scope(name, values=[concentration, scale]) as name:
dtype = dtype_util.common_dtype([concentration, scale])
concentration = tf.convert_to_tensor(value=concentration,
name="softplus_concentration",
dtype=dtype)
scale = tf.convert_to_tensor(value=scale,
name="softplus_scale",
dtype=dtype)
super(InverseGammaWithSoftplusConcentrationScale, self).__init__(
concentration=tf.nn.softplus(concentration,
name="softplus_concentration"),
scale=tf.nn.softplus(scale, name="softplus_scale"),
validate_args=validate_args,
allow_nan_stats=allow_nan_stats,
name=name)
self._parameters = parameters
_rate_deprecator = deprecation.deprecated(
"2019-05-08",
"InverseGammaWithSoftplusConcentrationRate is deprecated, use "
"InverseGammaWithSoftplusConcentrationScale instead.",
warn_once=True)
# pylint: disable=invalid-name
InverseGammaWithSoftplusConcentrationRate = _rate_deprecator(
InverseGammaWithSoftplusConcentrationScale)
from tensorflow.python.training import coordinator
from tensorflow.python.training import queue_runner
from tensorflow.python.training import saver as tf_saver
from tensorflow.python.training import session_manager as session_manager_lib
from tensorflow.python.training import summary_io
from tensorflow.python.training import supervisor as tf_supervisor
from tensorflow.python.util.deprecation import deprecated
# Singleton for SummaryWriter per logdir folder.
_SUMMARY_WRITERS = {}
# Lock protecting _SUMMARY_WRITERS
_summary_writer_lock = threading.Lock()
_graph_action_deprecation = deprecated(
'2017-02-15',
'graph_actions.py will be deleted. Use tf.train.* utilities instead. '
'You can use learn/estimators/estimator.py as an example.')
@_graph_action_deprecation
def clear_summary_writers():
"""Clear cached summary writers. Currently only used for unit tests."""
return summary_io.SummaryWriterCache.clear()
@deprecated(None, 'Use `SummaryWriterCache.get` directly.')
def get_summary_writer(logdir):
"""Returns single SummaryWriter per logdir in current run.
Args:
logdir: str, folder to write summaries.
'pipeline_config_path', '',
'Path to a pipeline_pb2.TrainEvalPipelineConfig config '
'file. If provided, other configs are ignored')
flags.DEFINE_string('eval_config_path', '',
'Path to an eval_pb2.EvalConfig config file.')
flags.DEFINE_string('input_config_path', '',
'Path to an input_reader_pb2.InputReader config file.')
flags.DEFINE_string('model_config_path', '',
'Path to a model_pb2.DetectionModel config file.')
flags.DEFINE_boolean(
'run_once', False, 'Option to only run a single pass of '
'evaluation. Overrides the `max_evals` parameter in the '
'provided config.')
FLAGS = flags.FLAGS
deprecated(None, 'Use object_detection/model_main.py.')
def main(unused_argv):
assert FLAGS.checkpoint_dir, '`checkpoint_dir` is missing.'
assert FLAGS.eval_dir, '`eval_dir` is missing.'
tf.gfile.MakeDirs(FLAGS.eval_dir)
if FLAGS.pipeline_config_path:
configs = config_util.get_configs_from_pipeline_file(
FLAGS.pipeline_config_path)
tf.gfile.Copy(FLAGS.pipeline_config_path,
os.path.join(FLAGS.eval_dir, 'pipeline.config'),
overwrite=True)
else:
configs = config_util.get_configs_from_multiple_files(
model_config_path=FLAGS.model_config_path,
from tensorflow.python.ops import gen_sparse_ops
from tensorflow.python.ops import gen_spectral_ops
from tensorflow.python.platform import tf_logging as logging
# go/tf-wildcard-import
# pylint: disable=wildcard-import
from tensorflow.python.ops.gen_math_ops import *
# pylint: enable=wildcard-import
from tensorflow.python.util import compat
from tensorflow.python.util import deprecation
from tensorflow.python.util import nest
from tensorflow.python.util.tf_export import tf_export
# Aliases for some automatically-generated names.
linspace = gen_math_ops.lin_space
arg_max = deprecation.deprecated(None, "Use `argmax` instead")(arg_max) # pylint: disable=used-before-assignment
arg_min = deprecation.deprecated(None, "Use `argmin` instead")(arg_min) # pylint: disable=used-before-assignment
tf_export("arg_max")(arg_max)
tf_export("arg_min")(arg_min)
# This is set by resource_variable_ops.py. It is included in this way since
# there is a circular dependency between math_ops and resource_variable_ops
_resource_variable_type = None
def _set_doc(doc):
def _decorator(func):
func.__doc__ = doc
return func
return _decorator
