repo stringlengths 7 55 | path stringlengths 4 223 | func_name stringlengths 1 134 | original_string stringlengths 75 104k | language stringclasses 1
value | code stringlengths 75 104k | code_tokens listlengths 19 28.4k | docstring stringlengths 1 46.9k | docstring_tokens listlengths 1 1.97k | sha stringlengths 40 40 | url stringlengths 87 315 | partition stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
deepmind/sonnet | sonnet/python/modules/batch_norm_v2.py | BatchNormV2._build_statistics | def _build_statistics(self, input_batch, use_batch_stats, stat_dtype):
"""Builds the statistics part of the graph when using moving variance.
Args:
input_batch: Input batch Tensor.
use_batch_stats: Boolean to indicate if batch statistics should be
calculated, otherwise moving averages are r... | python | def _build_statistics(self, input_batch, use_batch_stats, stat_dtype):
"""Builds the statistics part of the graph when using moving variance.
Args:
input_batch: Input batch Tensor.
use_batch_stats: Boolean to indicate if batch statistics should be
calculated, otherwise moving averages are r... | [
"def",
"_build_statistics",
"(",
"self",
",",
"input_batch",
",",
"use_batch_stats",
",",
"stat_dtype",
")",
":",
"# Set up our moving statistics. When connecting in parallel, this is shared.",
"if",
"self",
".",
"MOVING_MEAN",
"not",
"in",
"self",
".",
"_initializers",
"... | Builds the statistics part of the graph when using moving variance.
Args:
input_batch: Input batch Tensor.
use_batch_stats: Boolean to indicate if batch statistics should be
calculated, otherwise moving averages are returned.
stat_dtype: TensorFlow datatype to use for the moving mean and ... | [
"Builds",
"the",
"statistics",
"part",
"of",
"the",
"graph",
"when",
"using",
"moving",
"variance",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/batch_norm_v2.py#L217-L285 | train |
deepmind/sonnet | sonnet/python/modules/batch_norm_v2.py | BatchNormV2._build_update_ops | def _build_update_ops(self, mean, variance, is_training):
"""Builds the moving average update ops when using moving variance.
Args:
mean: The mean value to update with.
variance: The variance value to update with.
is_training: Boolean Tensor to indicate if we're currently in
training ... | python | def _build_update_ops(self, mean, variance, is_training):
"""Builds the moving average update ops when using moving variance.
Args:
mean: The mean value to update with.
variance: The variance value to update with.
is_training: Boolean Tensor to indicate if we're currently in
training ... | [
"def",
"_build_update_ops",
"(",
"self",
",",
"mean",
",",
"variance",
",",
"is_training",
")",
":",
"def",
"build_update_ops",
"(",
")",
":",
"\"\"\"Builds the exponential moving average update ops.\"\"\"",
"update_mean_op",
"=",
"moving_averages",
".",
"assign_moving_av... | Builds the moving average update ops when using moving variance.
Args:
mean: The mean value to update with.
variance: The variance value to update with.
is_training: Boolean Tensor to indicate if we're currently in
training mode.
Returns:
Tuple of `(update_mean_op, update_varia... | [
"Builds",
"the",
"moving",
"average",
"update",
"ops",
"when",
"using",
"moving",
"variance",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/batch_norm_v2.py#L287-L334 | train |
deepmind/sonnet | sonnet/python/modules/batch_norm_v2.py | BatchNormV2._fused_batch_norm_op | def _fused_batch_norm_op(self, input_batch, mean, variance, use_batch_stats):
"""Creates a fused batch normalization op."""
# Store the original shape of the mean and variance.
mean_shape = mean.get_shape()
variance_shape = variance.get_shape()
# The fused batch norm expects the mean, variance, gamm... | python | def _fused_batch_norm_op(self, input_batch, mean, variance, use_batch_stats):
"""Creates a fused batch normalization op."""
# Store the original shape of the mean and variance.
mean_shape = mean.get_shape()
variance_shape = variance.get_shape()
# The fused batch norm expects the mean, variance, gamm... | [
"def",
"_fused_batch_norm_op",
"(",
"self",
",",
"input_batch",
",",
"mean",
",",
"variance",
",",
"use_batch_stats",
")",
":",
"# Store the original shape of the mean and variance.",
"mean_shape",
"=",
"mean",
".",
"get_shape",
"(",
")",
"variance_shape",
"=",
"varia... | Creates a fused batch normalization op. | [
"Creates",
"a",
"fused",
"batch",
"normalization",
"op",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/batch_norm_v2.py#L336-L404 | train |
deepmind/sonnet | sonnet/python/modules/batch_norm_v2.py | BatchNormV2._build | def _build(self,
input_batch,
is_training,
test_local_stats=False):
"""Connects the BatchNormV2 module into the graph.
Args:
input_batch: A Tensor of the same dimension as `len(data_format)`.
is_training: A boolean to indicate if the module should be connected... | python | def _build(self,
input_batch,
is_training,
test_local_stats=False):
"""Connects the BatchNormV2 module into the graph.
Args:
input_batch: A Tensor of the same dimension as `len(data_format)`.
is_training: A boolean to indicate if the module should be connected... | [
"def",
"_build",
"(",
"self",
",",
"input_batch",
",",
"is_training",
",",
"test_local_stats",
"=",
"False",
")",
":",
"input_shape",
"=",
"input_batch",
".",
"get_shape",
"(",
")",
"if",
"not",
"self",
".",
"_data_format",
":",
"if",
"len",
"(",
"input_sh... | Connects the BatchNormV2 module into the graph.
Args:
input_batch: A Tensor of the same dimension as `len(data_format)`.
is_training: A boolean to indicate if the module should be connected in
training mode, meaning the moving averages are updated. Can be a Tensor.
test_local_stats: A boo... | [
"Connects",
"the",
"BatchNormV2",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/batch_norm_v2.py#L496-L583 | train |
deepmind/sonnet | sonnet/python/modules/nets/dilation.py | _range_along_dimension | def _range_along_dimension(range_dim, shape):
"""Construct a Tensor whose values are the index along a dimension.
Construct a Tensor that counts the distance along a single dimension. This is
useful, for example, when constructing an identity matrix,
>>> x = _range_along_dimension(0, [2, 2]).eval()
>>> ... | python | def _range_along_dimension(range_dim, shape):
"""Construct a Tensor whose values are the index along a dimension.
Construct a Tensor that counts the distance along a single dimension. This is
useful, for example, when constructing an identity matrix,
>>> x = _range_along_dimension(0, [2, 2]).eval()
>>> ... | [
"def",
"_range_along_dimension",
"(",
"range_dim",
",",
"shape",
")",
":",
"rank",
"=",
"len",
"(",
"shape",
")",
"if",
"range_dim",
">=",
"rank",
":",
"raise",
"ValueError",
"(",
"\"Cannot calculate range along non-existent index.\"",
")",
"indices",
"=",
"tf",
... | Construct a Tensor whose values are the index along a dimension.
Construct a Tensor that counts the distance along a single dimension. This is
useful, for example, when constructing an identity matrix,
>>> x = _range_along_dimension(0, [2, 2]).eval()
>>> x
array([[0, 0],
[1, 1]], dtype=int3... | [
"Construct",
"a",
"Tensor",
"whose",
"values",
"are",
"the",
"index",
"along",
"a",
"dimension",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/dilation.py#L32-L70 | train |
deepmind/sonnet | sonnet/python/modules/nets/dilation.py | identity_kernel_initializer | def identity_kernel_initializer(shape, dtype=tf.float32, partition_info=None):
"""An initializer for constructing identity convolution kernels.
Constructs a convolution kernel such that applying it is the same as an
identity operation on the input. Formally, the kernel has entry [i, j, in,
out] = 1 if in equal... | python | def identity_kernel_initializer(shape, dtype=tf.float32, partition_info=None):
"""An initializer for constructing identity convolution kernels.
Constructs a convolution kernel such that applying it is the same as an
identity operation on the input. Formally, the kernel has entry [i, j, in,
out] = 1 if in equal... | [
"def",
"identity_kernel_initializer",
"(",
"shape",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"partition_info",
"=",
"None",
")",
":",
"if",
"len",
"(",
"shape",
")",
"!=",
"4",
":",
"raise",
"ValueError",
"(",
"\"Convolution kernels must be rank 4.\"",
")... | An initializer for constructing identity convolution kernels.
Constructs a convolution kernel such that applying it is the same as an
identity operation on the input. Formally, the kernel has entry [i, j, in,
out] = 1 if in equals out and i and j are the middle of the kernel and 0
otherwise.
Args:
shape... | [
"An",
"initializer",
"for",
"constructing",
"identity",
"convolution",
"kernels",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/dilation.py#L74-L118 | train |
deepmind/sonnet | sonnet/python/modules/nets/dilation.py | noisy_identity_kernel_initializer | def noisy_identity_kernel_initializer(base_num_channels, stddev=1e-8):
"""Build an initializer for constructing near-identity convolution kernels.
Construct a convolution kernel where in_channels and out_channels are
multiples of base_num_channels, but need not be equal. This initializer is
essentially the sam... | python | def noisy_identity_kernel_initializer(base_num_channels, stddev=1e-8):
"""Build an initializer for constructing near-identity convolution kernels.
Construct a convolution kernel where in_channels and out_channels are
multiples of base_num_channels, but need not be equal. This initializer is
essentially the sam... | [
"def",
"noisy_identity_kernel_initializer",
"(",
"base_num_channels",
",",
"stddev",
"=",
"1e-8",
")",
":",
"# pylint: disable=unused-argument",
"def",
"_noisy_identity_kernel_initializer",
"(",
"shape",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"partition_info",
"=... | Build an initializer for constructing near-identity convolution kernels.
Construct a convolution kernel where in_channels and out_channels are
multiples of base_num_channels, but need not be equal. This initializer is
essentially the same as identity_kernel_initializer, except that magnitude
is "spread out" ac... | [
"Build",
"an",
"initializer",
"for",
"constructing",
"near",
"-",
"identity",
"convolution",
"kernels",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/dilation.py#L121-L194 | train |
deepmind/sonnet | sonnet/python/modules/nets/dilation.py | Dilation._build | def _build(self, images):
"""Build dilation module.
Args:
images: Tensor of shape [batch_size, height, width, depth]
and dtype float32. Represents a set of images with an arbitrary depth.
Note that when using the default initializer, depth must equal
num_output_classes.
Retur... | python | def _build(self, images):
"""Build dilation module.
Args:
images: Tensor of shape [batch_size, height, width, depth]
and dtype float32. Represents a set of images with an arbitrary depth.
Note that when using the default initializer, depth must equal
num_output_classes.
Retur... | [
"def",
"_build",
"(",
"self",
",",
"images",
")",
":",
"num_classes",
"=",
"self",
".",
"_num_output_classes",
"if",
"len",
"(",
"images",
".",
"get_shape",
"(",
")",
")",
"!=",
"4",
":",
"raise",
"base",
".",
"IncompatibleShapeError",
"(",
"\"'images' mus... | Build dilation module.
Args:
images: Tensor of shape [batch_size, height, width, depth]
and dtype float32. Represents a set of images with an arbitrary depth.
Note that when using the default initializer, depth must equal
num_output_classes.
Returns:
Tensor of shape [batch_... | [
"Build",
"dilation",
"module",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/dilation.py#L257-L319 | train |
deepmind/sonnet | sonnet/python/modules/nets/dilation.py | Dilation._dilated_conv_layer | def _dilated_conv_layer(self, output_channels, dilation_rate, apply_relu,
name):
"""Create a dilated convolution layer.
Args:
output_channels: int. Number of output channels for each pixel.
dilation_rate: int. Represents how many pixels each stride offset will
move... | python | def _dilated_conv_layer(self, output_channels, dilation_rate, apply_relu,
name):
"""Create a dilated convolution layer.
Args:
output_channels: int. Number of output channels for each pixel.
dilation_rate: int. Represents how many pixels each stride offset will
move... | [
"def",
"_dilated_conv_layer",
"(",
"self",
",",
"output_channels",
",",
"dilation_rate",
",",
"apply_relu",
",",
"name",
")",
":",
"layer_components",
"=",
"[",
"conv",
".",
"Conv2D",
"(",
"output_channels",
",",
"[",
"3",
",",
"3",
"]",
",",
"initializers",... | Create a dilated convolution layer.
Args:
output_channels: int. Number of output channels for each pixel.
dilation_rate: int. Represents how many pixels each stride offset will
move. A value of 1 indicates a standard convolution.
apply_relu: bool. If True, a ReLU non-linearlity is added.
... | [
"Create",
"a",
"dilated",
"convolution",
"layer",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/dilation.py#L321-L345 | train |
deepmind/sonnet | sonnet/python/ops/nest.py | with_deprecation_warning | def with_deprecation_warning(fn, extra_message=''):
"""Wraps the function and prints a warn-once (per `extra_message`) warning."""
def new_fn(*args, **kwargs):
if extra_message not in _DONE_WARN:
tf.logging.warning(
'Sonnet nest is deprecated. Please use '
'tf.contrib.framework.nest in... | python | def with_deprecation_warning(fn, extra_message=''):
"""Wraps the function and prints a warn-once (per `extra_message`) warning."""
def new_fn(*args, **kwargs):
if extra_message not in _DONE_WARN:
tf.logging.warning(
'Sonnet nest is deprecated. Please use '
'tf.contrib.framework.nest in... | [
"def",
"with_deprecation_warning",
"(",
"fn",
",",
"extra_message",
"=",
"''",
")",
":",
"def",
"new_fn",
"(",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"if",
"extra_message",
"not",
"in",
"_DONE_WARN",
":",
"tf",
".",
"logging",
".",
"warning",
... | Wraps the function and prints a warn-once (per `extra_message`) warning. | [
"Wraps",
"the",
"function",
"and",
"prints",
"a",
"warn",
"-",
"once",
"(",
"per",
"extra_message",
")",
"warning",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/ops/nest.py#L33-L44 | train |
deepmind/sonnet | sonnet/examples/module_with_build_args.py | custom_build | def custom_build(inputs, is_training, keep_prob):
"""A custom build method to wrap into a sonnet Module."""
outputs = snt.Conv2D(output_channels=32, kernel_shape=4, stride=2)(inputs)
outputs = snt.BatchNorm()(outputs, is_training=is_training)
outputs = tf.nn.relu(outputs)
outputs = snt.Conv2D(output_channels=... | python | def custom_build(inputs, is_training, keep_prob):
"""A custom build method to wrap into a sonnet Module."""
outputs = snt.Conv2D(output_channels=32, kernel_shape=4, stride=2)(inputs)
outputs = snt.BatchNorm()(outputs, is_training=is_training)
outputs = tf.nn.relu(outputs)
outputs = snt.Conv2D(output_channels=... | [
"def",
"custom_build",
"(",
"inputs",
",",
"is_training",
",",
"keep_prob",
")",
":",
"outputs",
"=",
"snt",
".",
"Conv2D",
"(",
"output_channels",
"=",
"32",
",",
"kernel_shape",
"=",
"4",
",",
"stride",
"=",
"2",
")",
"(",
"inputs",
")",
"outputs",
"... | A custom build method to wrap into a sonnet Module. | [
"A",
"custom",
"build",
"method",
"to",
"wrap",
"into",
"a",
"sonnet",
"Module",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/module_with_build_args.py#L41-L52 | train |
deepmind/sonnet | sonnet/python/modules/base.py | _get_or_create_stack | def _get_or_create_stack(name):
"""Returns a thread local stack uniquified by the given name."""
stack = getattr(_LOCAL_STACKS, name, None)
if stack is None:
stack = []
setattr(_LOCAL_STACKS, name, stack)
return stack | python | def _get_or_create_stack(name):
"""Returns a thread local stack uniquified by the given name."""
stack = getattr(_LOCAL_STACKS, name, None)
if stack is None:
stack = []
setattr(_LOCAL_STACKS, name, stack)
return stack | [
"def",
"_get_or_create_stack",
"(",
"name",
")",
":",
"stack",
"=",
"getattr",
"(",
"_LOCAL_STACKS",
",",
"name",
",",
"None",
")",
"if",
"stack",
"is",
"None",
":",
"stack",
"=",
"[",
"]",
"setattr",
"(",
"_LOCAL_STACKS",
",",
"name",
",",
"stack",
")... | Returns a thread local stack uniquified by the given name. | [
"Returns",
"a",
"thread",
"local",
"stack",
"uniquified",
"by",
"the",
"given",
"name",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/base.py#L60-L66 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | diagonal_gaussian_posterior_builder | def diagonal_gaussian_posterior_builder(
getter, name, shape=None, *args, **kwargs):
"""A pre-canned builder for diagonal gaussian posterior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a diagonal posterior over a variable of t... | python | def diagonal_gaussian_posterior_builder(
getter, name, shape=None, *args, **kwargs):
"""A pre-canned builder for diagonal gaussian posterior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a diagonal posterior over a variable of t... | [
"def",
"diagonal_gaussian_posterior_builder",
"(",
"getter",
",",
"name",
",",
"shape",
"=",
"None",
",",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"# Please see the documentation for",
"# `tfp.distributions.param_static_shapes`.",
"parameter_shapes",
"=",
"tfp",
... | A pre-canned builder for diagonal gaussian posterior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a diagonal posterior over a variable of the
requisite shape.
Args:
getter: The `getter` passed to a `custom_getter`. Please se... | [
"A",
"pre",
"-",
"canned",
"builder",
"for",
"diagonal",
"gaussian",
"posterior",
"distributions",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L148-L183 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | fixed_gaussian_prior_builder | def fixed_gaussian_prior_builder(
getter, name, dtype=None, *args, **kwargs):
"""A pre-canned builder for fixed gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued fixed gaussian prior which
will be br... | python | def fixed_gaussian_prior_builder(
getter, name, dtype=None, *args, **kwargs):
"""A pre-canned builder for fixed gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued fixed gaussian prior which
will be br... | [
"def",
"fixed_gaussian_prior_builder",
"(",
"getter",
",",
"name",
",",
"dtype",
"=",
"None",
",",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"del",
"getter",
"# Unused.",
"del",
"args",
"# Unused.",
"del",
"kwargs",
"# Unused.",
"loc",
"=",
"tf",
"... | A pre-canned builder for fixed gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued fixed gaussian prior which
will be broadcast over a variable of the requisite shape.
Args:
getter: The `getter` passe... | [
"A",
"pre",
"-",
"canned",
"builder",
"for",
"fixed",
"gaussian",
"prior",
"distributions",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L188-L214 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | adaptive_gaussian_prior_builder | def adaptive_gaussian_prior_builder(
getter, name, *args, **kwargs):
"""A pre-canned builder for adaptive scalar gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued adaptive gaussian prior
which will b... | python | def adaptive_gaussian_prior_builder(
getter, name, *args, **kwargs):
"""A pre-canned builder for adaptive scalar gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued adaptive gaussian prior
which will b... | [
"def",
"adaptive_gaussian_prior_builder",
"(",
"getter",
",",
"name",
",",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"kwargs",
"[",
"\"shape\"",
"]",
"=",
"(",
")",
"loc_var",
"=",
"getter",
"(",
"name",
"+",
"\"_prior_loc\"",
",",
"*",
"args",
"... | A pre-canned builder for adaptive scalar gaussian prior distributions.
Given a true `getter` function and arguments forwarded from `tf.get_variable`,
return a distribution object for a scalar-valued adaptive gaussian prior
which will be broadcast over a variable of the requisite shape. This prior's
parameters ... | [
"A",
"pre",
"-",
"canned",
"builder",
"for",
"adaptive",
"scalar",
"gaussian",
"prior",
"distributions",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L218-L247 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | stochastic_kl_builder | def stochastic_kl_builder(posterior, prior, sample):
"""A pre-canned builder for a ubiquitous stochastic KL estimator."""
return tf.subtract(
tf.reduce_sum(posterior.log_prob(sample)),
tf.reduce_sum(prior.log_prob(sample))) | python | def stochastic_kl_builder(posterior, prior, sample):
"""A pre-canned builder for a ubiquitous stochastic KL estimator."""
return tf.subtract(
tf.reduce_sum(posterior.log_prob(sample)),
tf.reduce_sum(prior.log_prob(sample))) | [
"def",
"stochastic_kl_builder",
"(",
"posterior",
",",
"prior",
",",
"sample",
")",
":",
"return",
"tf",
".",
"subtract",
"(",
"tf",
".",
"reduce_sum",
"(",
"posterior",
".",
"log_prob",
"(",
"sample",
")",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"prior"... | A pre-canned builder for a ubiquitous stochastic KL estimator. | [
"A",
"pre",
"-",
"canned",
"builder",
"for",
"a",
"ubiquitous",
"stochastic",
"KL",
"estimator",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L250-L254 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | analytic_kl_builder | def analytic_kl_builder(posterior, prior, sample):
"""A pre-canned builder for the analytic kl divergence."""
del sample
return tf.reduce_sum(tfp.distributions.kl_divergence(posterior, prior)) | python | def analytic_kl_builder(posterior, prior, sample):
"""A pre-canned builder for the analytic kl divergence."""
del sample
return tf.reduce_sum(tfp.distributions.kl_divergence(posterior, prior)) | [
"def",
"analytic_kl_builder",
"(",
"posterior",
",",
"prior",
",",
"sample",
")",
":",
"del",
"sample",
"return",
"tf",
".",
"reduce_sum",
"(",
"tfp",
".",
"distributions",
".",
"kl_divergence",
"(",
"posterior",
",",
"prior",
")",
")"
] | A pre-canned builder for the analytic kl divergence. | [
"A",
"pre",
"-",
"canned",
"builder",
"for",
"the",
"analytic",
"kl",
"divergence",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L257-L260 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | bayes_by_backprop_getter | def bayes_by_backprop_getter(
posterior_builder=diagonal_gaussian_posterior_builder,
prior_builder=fixed_gaussian_prior_builder,
kl_builder=stochastic_kl_builder,
sampling_mode_tensor=None,
fresh_noise_per_connection=True,
keep_control_dependencies=False):
"""Creates a custom getter which does... | python | def bayes_by_backprop_getter(
posterior_builder=diagonal_gaussian_posterior_builder,
prior_builder=fixed_gaussian_prior_builder,
kl_builder=stochastic_kl_builder,
sampling_mode_tensor=None,
fresh_noise_per_connection=True,
keep_control_dependencies=False):
"""Creates a custom getter which does... | [
"def",
"bayes_by_backprop_getter",
"(",
"posterior_builder",
"=",
"diagonal_gaussian_posterior_builder",
",",
"prior_builder",
"=",
"fixed_gaussian_prior_builder",
",",
"kl_builder",
"=",
"stochastic_kl_builder",
",",
"sampling_mode_tensor",
"=",
"None",
",",
"fresh_noise_per_c... | Creates a custom getter which does Bayes by Backprop.
Please see `tf.get_variable` for general documentation on custom getters.
All arguments are optional. If nothing is configued, then a diagonal gaussian
posterior will be used, and a fixed N(0, 0.01) prior will be used. Please
see the default `posterior_bui... | [
"Creates",
"a",
"custom",
"getter",
"which",
"does",
"Bayes",
"by",
"Backprop",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L263-L452 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | _produce_posterior_estimate | def _produce_posterior_estimate(posterior_dist, posterior_estimate_mode,
raw_var_name):
"""Create tensor representing estimate of posterior.
Args:
posterior_dist: An instance of `tfp.distributions.Distribution`.
The variational posterior from which to produce an estimate... | python | def _produce_posterior_estimate(posterior_dist, posterior_estimate_mode,
raw_var_name):
"""Create tensor representing estimate of posterior.
Args:
posterior_dist: An instance of `tfp.distributions.Distribution`.
The variational posterior from which to produce an estimate... | [
"def",
"_produce_posterior_estimate",
"(",
"posterior_dist",
",",
"posterior_estimate_mode",
",",
"raw_var_name",
")",
":",
"conds",
"=",
"[",
"tf",
".",
"equal",
"(",
"posterior_estimate_mode",
",",
"tf",
".",
"constant",
"(",
"EstimatorModes",
".",
"sample",
")"... | Create tensor representing estimate of posterior.
Args:
posterior_dist: An instance of `tfp.distributions.Distribution`.
The variational posterior from which to produce an estimate of the
variable in question.
posterior_estimate_mode: A `Tensor` of dtype `tf.string`, which
determines ... | [
"Create",
"tensor",
"representing",
"estimate",
"of",
"posterior",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L455-L505 | train |
deepmind/sonnet | sonnet/python/custom_getters/bayes_by_backprop.py | get_total_kl_cost | def get_total_kl_cost(name="total_kl_cost", filter_by_name_substring=None):
"""Get the total cost for all (or a subset of) the stochastic variables.
Args:
name: A name for the tensor representing the total kl cost.
filter_by_name_substring: A string used to filter which variables count
toward the tot... | python | def get_total_kl_cost(name="total_kl_cost", filter_by_name_substring=None):
"""Get the total cost for all (or a subset of) the stochastic variables.
Args:
name: A name for the tensor representing the total kl cost.
filter_by_name_substring: A string used to filter which variables count
toward the tot... | [
"def",
"get_total_kl_cost",
"(",
"name",
"=",
"\"total_kl_cost\"",
",",
"filter_by_name_substring",
"=",
"None",
")",
":",
"all_variable_metadata",
"=",
"get_variable_metadata",
"(",
"filter_by_name_substring",
")",
"if",
"not",
"all_variable_metadata",
":",
"tf",
".",
... | Get the total cost for all (or a subset of) the stochastic variables.
Args:
name: A name for the tensor representing the total kl cost.
filter_by_name_substring: A string used to filter which variables count
toward the total KL cost. By default, this argument is `None`, and all
variables trained ... | [
"Get",
"the",
"total",
"cost",
"for",
"all",
"(",
"or",
"a",
"subset",
"of",
")",
"the",
"stochastic",
"variables",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/custom_getters/bayes_by_backprop.py#L508-L527 | train |
deepmind/sonnet | sonnet/python/modules/block_matrix.py | BlockTriangularMatrix.output_shape | def output_shape(self):
"""The shape of the output matrix."""
return (self._block_shape[0] * self._block_rows,
self._block_shape[1] * self._block_rows) | python | def output_shape(self):
"""The shape of the output matrix."""
return (self._block_shape[0] * self._block_rows,
self._block_shape[1] * self._block_rows) | [
"def",
"output_shape",
"(",
"self",
")",
":",
"return",
"(",
"self",
".",
"_block_shape",
"[",
"0",
"]",
"*",
"self",
".",
"_block_rows",
",",
"self",
".",
"_block_shape",
"[",
"1",
"]",
"*",
"self",
".",
"_block_rows",
")"
] | The shape of the output matrix. | [
"The",
"shape",
"of",
"the",
"output",
"matrix",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/block_matrix.py#L106-L109 | train |
deepmind/sonnet | sonnet/python/modules/block_matrix.py | BlockTriangularMatrix._left_zero_blocks | def _left_zero_blocks(self, r):
"""Number of blocks with zeros from the left in block row `r`."""
if not self._include_off_diagonal:
return r
elif not self._upper:
return 0
elif self._include_diagonal:
return r
else:
return r + 1 | python | def _left_zero_blocks(self, r):
"""Number of blocks with zeros from the left in block row `r`."""
if not self._include_off_diagonal:
return r
elif not self._upper:
return 0
elif self._include_diagonal:
return r
else:
return r + 1 | [
"def",
"_left_zero_blocks",
"(",
"self",
",",
"r",
")",
":",
"if",
"not",
"self",
".",
"_include_off_diagonal",
":",
"return",
"r",
"elif",
"not",
"self",
".",
"_upper",
":",
"return",
"0",
"elif",
"self",
".",
"_include_diagonal",
":",
"return",
"r",
"e... | Number of blocks with zeros from the left in block row `r`. | [
"Number",
"of",
"blocks",
"with",
"zeros",
"from",
"the",
"left",
"in",
"block",
"row",
"r",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/block_matrix.py#L160-L169 | train |
deepmind/sonnet | sonnet/python/modules/block_matrix.py | BlockTriangularMatrix._right_zero_blocks | def _right_zero_blocks(self, r):
"""Number of blocks with zeros from the right in block row `r`."""
if not self._include_off_diagonal:
return self._block_rows - r - 1
elif self._upper:
return 0
elif self._include_diagonal:
return self._block_rows - r - 1
else:
return self._bl... | python | def _right_zero_blocks(self, r):
"""Number of blocks with zeros from the right in block row `r`."""
if not self._include_off_diagonal:
return self._block_rows - r - 1
elif self._upper:
return 0
elif self._include_diagonal:
return self._block_rows - r - 1
else:
return self._bl... | [
"def",
"_right_zero_blocks",
"(",
"self",
",",
"r",
")",
":",
"if",
"not",
"self",
".",
"_include_off_diagonal",
":",
"return",
"self",
".",
"_block_rows",
"-",
"r",
"-",
"1",
"elif",
"self",
".",
"_upper",
":",
"return",
"0",
"elif",
"self",
".",
"_in... | Number of blocks with zeros from the right in block row `r`. | [
"Number",
"of",
"blocks",
"with",
"zeros",
"from",
"the",
"right",
"in",
"block",
"row",
"r",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/block_matrix.py#L171-L180 | train |
deepmind/sonnet | sonnet/python/modules/block_matrix.py | BlockTriangularMatrix._content_blocks | def _content_blocks(self, r):
"""Number of content blocks in block row `r`."""
return (self._block_rows - self._left_zero_blocks(r)
- self._right_zero_blocks(r)) | python | def _content_blocks(self, r):
"""Number of content blocks in block row `r`."""
return (self._block_rows - self._left_zero_blocks(r)
- self._right_zero_blocks(r)) | [
"def",
"_content_blocks",
"(",
"self",
",",
"r",
")",
":",
"return",
"(",
"self",
".",
"_block_rows",
"-",
"self",
".",
"_left_zero_blocks",
"(",
"r",
")",
"-",
"self",
".",
"_right_zero_blocks",
"(",
"r",
")",
")"
] | Number of content blocks in block row `r`. | [
"Number",
"of",
"content",
"blocks",
"in",
"block",
"row",
"r",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/block_matrix.py#L182-L185 | train |
deepmind/sonnet | sonnet/examples/rmc_nth_farthest.py | build_and_train | def build_and_train(iterations, log_stride, test=False):
"""Construct the data, model, loss and optimizer then train."""
# Test mode settings.
batch_size = 2 if test else FLAGS.batch_size
num_mems = 2 if test else FLAGS.num_mems
num_heads = 1 if test else FLAGS.num_mems
num_blocks = 1 if test else FLAGS.nu... | python | def build_and_train(iterations, log_stride, test=False):
"""Construct the data, model, loss and optimizer then train."""
# Test mode settings.
batch_size = 2 if test else FLAGS.batch_size
num_mems = 2 if test else FLAGS.num_mems
num_heads = 1 if test else FLAGS.num_mems
num_blocks = 1 if test else FLAGS.nu... | [
"def",
"build_and_train",
"(",
"iterations",
",",
"log_stride",
",",
"test",
"=",
"False",
")",
":",
"# Test mode settings.",
"batch_size",
"=",
"2",
"if",
"test",
"else",
"FLAGS",
".",
"batch_size",
"num_mems",
"=",
"2",
"if",
"test",
"else",
"FLAGS",
".",
... | Construct the data, model, loss and optimizer then train. | [
"Construct",
"the",
"data",
"model",
"loss",
"and",
"optimizer",
"then",
"train",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/rmc_nth_farthest.py#L93-L202 | train |
deepmind/sonnet | sonnet/examples/rmc_nth_farthest.py | SequenceModel._build | def _build(self, inputs):
"""Dynamic unroll across input objects.
Args:
inputs: tensor (batch x num_objects x feature). Objects to sort.
Returns:
Tensor (batch x num_objects); logits indicating the reference objects.
"""
batch_size = inputs.get_shape()[0]
output_sequence, _ = tf.nn... | python | def _build(self, inputs):
"""Dynamic unroll across input objects.
Args:
inputs: tensor (batch x num_objects x feature). Objects to sort.
Returns:
Tensor (batch x num_objects); logits indicating the reference objects.
"""
batch_size = inputs.get_shape()[0]
output_sequence, _ = tf.nn... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
")",
":",
"batch_size",
"=",
"inputs",
".",
"get_shape",
"(",
")",
"[",
"0",
"]",
"output_sequence",
",",
"_",
"=",
"tf",
".",
"nn",
".",
"dynamic_rnn",
"(",
"cell",
"=",
"self",
".",
"_core",
",",
"inpu... | Dynamic unroll across input objects.
Args:
inputs: tensor (batch x num_objects x feature). Objects to sort.
Returns:
Tensor (batch x num_objects); logits indicating the reference objects. | [
"Dynamic",
"unroll",
"across",
"input",
"objects",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/rmc_nth_farthest.py#L70-L90 | train |
deepmind/sonnet | sonnet/python/modules/clip_gradient.py | _clip_gradient_op | def _clip_gradient_op(dtype):
"""Create an op that clips gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these op
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is used.
... | python | def _clip_gradient_op(dtype):
"""Create an op that clips gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these op
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is used.
... | [
"def",
"_clip_gradient_op",
"(",
"dtype",
")",
":",
"def",
"clip_gradient_backward",
"(",
"op",
",",
"grad",
")",
":",
"clip_value_min",
"=",
"op",
".",
"inputs",
"[",
"1",
"]",
"clip_value_max",
"=",
"op",
".",
"inputs",
"[",
"2",
"]",
"clipped_grad",
"... | Create an op that clips gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these op
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is used.
This method produces a new op th... | [
"Create",
"an",
"op",
"that",
"clips",
"gradients",
"using",
"a",
"Defun",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/clip_gradient.py#L25-L59 | train |
deepmind/sonnet | sonnet/python/modules/clip_gradient.py | clip_gradient | def clip_gradient(net, clip_value_min, clip_value_max, name=None):
"""Clips respective gradients of a given tensor.
Acts as identity for the forward pass, but clips gradient tensor element-wise
by value during the backward pass. Any gradient values less than
`clip_value_min` or greater than `clip_values_max` a... | python | def clip_gradient(net, clip_value_min, clip_value_max, name=None):
"""Clips respective gradients of a given tensor.
Acts as identity for the forward pass, but clips gradient tensor element-wise
by value during the backward pass. Any gradient values less than
`clip_value_min` or greater than `clip_values_max` a... | [
"def",
"clip_gradient",
"(",
"net",
",",
"clip_value_min",
",",
"clip_value_max",
",",
"name",
"=",
"None",
")",
":",
"if",
"not",
"net",
".",
"dtype",
".",
"is_floating",
":",
"raise",
"ValueError",
"(",
"\"clip_gradient does not support non-float `net` inputs.\"",... | Clips respective gradients of a given tensor.
Acts as identity for the forward pass, but clips gradient tensor element-wise
by value during the backward pass. Any gradient values less than
`clip_value_min` or greater than `clip_values_max` are set to the respective
limit values.
Args:
net: A `tf.Tensor`... | [
"Clips",
"respective",
"gradients",
"of",
"a",
"given",
"tensor",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/clip_gradient.py#L62-L94 | train |
deepmind/sonnet | sonnet/examples/ptb_reader.py | ptb_raw_data | def ptb_raw_data(data_path):
"""Load PTB raw data from data directory "data_path".
Reads PTB text files, converts strings to integer ids,
and performs mini-batching of the inputs.
The PTB dataset comes from Tomas Mikolov's webpage:
http://www.fit.vutbr.cz/~imikolov/rnnlm/simple-examples.tgz
Args:
da... | python | def ptb_raw_data(data_path):
"""Load PTB raw data from data directory "data_path".
Reads PTB text files, converts strings to integer ids,
and performs mini-batching of the inputs.
The PTB dataset comes from Tomas Mikolov's webpage:
http://www.fit.vutbr.cz/~imikolov/rnnlm/simple-examples.tgz
Args:
da... | [
"def",
"ptb_raw_data",
"(",
"data_path",
")",
":",
"train_path",
"=",
"os",
".",
"path",
".",
"join",
"(",
"data_path",
",",
"\"ptb.train.txt\"",
")",
"valid_path",
"=",
"os",
".",
"path",
".",
"join",
"(",
"data_path",
",",
"\"ptb.valid.txt\"",
")",
"test... | Load PTB raw data from data directory "data_path".
Reads PTB text files, converts strings to integer ids,
and performs mini-batching of the inputs.
The PTB dataset comes from Tomas Mikolov's webpage:
http://www.fit.vutbr.cz/~imikolov/rnnlm/simple-examples.tgz
Args:
data_path: string path to the direct... | [
"Load",
"PTB",
"raw",
"data",
"from",
"data",
"directory",
"data_path",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/ptb_reader.py#L55-L82 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D._check_and_assign_normalization_members | def _check_and_assign_normalization_members(self, normalization_ctor,
normalization_kwargs):
"""Checks that the normalization constructor is callable."""
if isinstance(normalization_ctor, six.string_types):
normalization_ctor = util.parse_string_to_constructor... | python | def _check_and_assign_normalization_members(self, normalization_ctor,
normalization_kwargs):
"""Checks that the normalization constructor is callable."""
if isinstance(normalization_ctor, six.string_types):
normalization_ctor = util.parse_string_to_constructor... | [
"def",
"_check_and_assign_normalization_members",
"(",
"self",
",",
"normalization_ctor",
",",
"normalization_kwargs",
")",
":",
"if",
"isinstance",
"(",
"normalization_ctor",
",",
"six",
".",
"string_types",
")",
":",
"normalization_ctor",
"=",
"util",
".",
"parse_st... | Checks that the normalization constructor is callable. | [
"Checks",
"that",
"the",
"normalization",
"constructor",
"is",
"callable",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L252-L262 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D._parse_normalization_kwargs | def _parse_normalization_kwargs(self, use_batch_norm, batch_norm_config,
normalization_ctor, normalization_kwargs):
"""Sets up normalization, checking old and new flags."""
if use_batch_norm is not None:
# Delete this whole block when deprecation is done.
util.depre... | python | def _parse_normalization_kwargs(self, use_batch_norm, batch_norm_config,
normalization_ctor, normalization_kwargs):
"""Sets up normalization, checking old and new flags."""
if use_batch_norm is not None:
# Delete this whole block when deprecation is done.
util.depre... | [
"def",
"_parse_normalization_kwargs",
"(",
"self",
",",
"use_batch_norm",
",",
"batch_norm_config",
",",
"normalization_ctor",
",",
"normalization_kwargs",
")",
":",
"if",
"use_batch_norm",
"is",
"not",
"None",
":",
"# Delete this whole block when deprecation is done.",
"ut... | Sets up normalization, checking old and new flags. | [
"Sets",
"up",
"normalization",
"checking",
"old",
"and",
"new",
"flags",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L264-L286 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D._instantiate_layers | def _instantiate_layers(self):
"""Instantiates all the convolutional modules used in the network."""
# Here we are entering the module's variable scope to name our submodules
# correctly (not to create variables). As such it's safe to not check
# whether we're in the same graph. This is important if we... | python | def _instantiate_layers(self):
"""Instantiates all the convolutional modules used in the network."""
# Here we are entering the module's variable scope to name our submodules
# correctly (not to create variables). As such it's safe to not check
# whether we're in the same graph. This is important if we... | [
"def",
"_instantiate_layers",
"(",
"self",
")",
":",
"# Here we are entering the module's variable scope to name our submodules",
"# correctly (not to create variables). As such it's safe to not check",
"# whether we're in the same graph. This is important if we're constructing",
"# the module in ... | Instantiates all the convolutional modules used in the network. | [
"Instantiates",
"all",
"the",
"convolutional",
"modules",
"used",
"in",
"the",
"network",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L288-L309 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D._build | def _build(self, inputs, **normalization_build_kwargs):
"""Assembles the `ConvNet2D` and connects it to the graph.
Args:
inputs: A 4D Tensor of shape `[batch_size, input_height, input_width,
input_channels]`.
**normalization_build_kwargs: kwargs passed to the normalization module
at... | python | def _build(self, inputs, **normalization_build_kwargs):
"""Assembles the `ConvNet2D` and connects it to the graph.
Args:
inputs: A 4D Tensor of shape `[batch_size, input_height, input_width,
input_channels]`.
**normalization_build_kwargs: kwargs passed to the normalization module
at... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"*",
"*",
"normalization_build_kwargs",
")",
":",
"if",
"(",
"self",
".",
"_normalization_ctor",
"in",
"{",
"batch_norm",
".",
"BatchNorm",
",",
"batch_norm_v2",
".",
"BatchNormV2",
"}",
"and",
"\"is_training\"... | Assembles the `ConvNet2D` and connects it to the graph.
Args:
inputs: A 4D Tensor of shape `[batch_size, input_height, input_width,
input_channels]`.
**normalization_build_kwargs: kwargs passed to the normalization module
at _build time.
Returns:
A 4D Tensor of shape `[batch_... | [
"Assembles",
"the",
"ConvNet2D",
"and",
"connects",
"it",
"to",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L311-L361 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D._transpose | def _transpose(self,
transpose_constructor,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
nor... | python | def _transpose(self,
transpose_constructor,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
nor... | [
"def",
"_transpose",
"(",
"self",
",",
"transpose_constructor",
",",
"name",
"=",
"None",
",",
"output_channels",
"=",
"None",
",",
"kernel_shapes",
"=",
"None",
",",
"strides",
"=",
"None",
",",
"paddings",
"=",
"None",
",",
"activation",
"=",
"None",
","... | Returns transposed version of this network.
Args:
transpose_constructor: A method that creates an instance of the transposed
network type. The method must accept the same kwargs as this methods
with the exception of the `transpose_constructor` argument.
name: Optional string specifying ... | [
"Returns",
"transposed",
"version",
"of",
"this",
"network",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L445-L585 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2D.transpose | def transpose(self,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
normalization_ctor=None,
normalizati... | python | def transpose(self,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
normalization_ctor=None,
normalizati... | [
"def",
"transpose",
"(",
"self",
",",
"name",
"=",
"None",
",",
"output_channels",
"=",
"None",
",",
"kernel_shapes",
"=",
"None",
",",
"strides",
"=",
"None",
",",
"paddings",
"=",
"None",
",",
"activation",
"=",
"None",
",",
"activate_final",
"=",
"Non... | Returns transposed version of this network.
Args:
name: Optional string specifying the name of the transposed module. The
default name is constructed by appending "_transpose"
to `self.module_name`.
output_channels: Optional iterable of numbers of output channels.
kernel_shapes: O... | [
"Returns",
"transposed",
"version",
"of",
"this",
"network",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L588-L712 | train |
deepmind/sonnet | sonnet/python/modules/nets/convnet.py | ConvNet2DTranspose.transpose | def transpose(self,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
normalization_ctor=None,
normalizati... | python | def transpose(self,
name=None,
output_channels=None,
kernel_shapes=None,
strides=None,
paddings=None,
activation=None,
activate_final=None,
normalization_ctor=None,
normalizati... | [
"def",
"transpose",
"(",
"self",
",",
"name",
"=",
"None",
",",
"output_channels",
"=",
"None",
",",
"kernel_shapes",
"=",
"None",
",",
"strides",
"=",
"None",
",",
"paddings",
"=",
"None",
",",
"activation",
"=",
"None",
",",
"activate_final",
"=",
"Non... | Returns transposed version of this network.
Args:
name: Optional string specifying the name of the transposed module. The
default name is constructed by appending "_transpose"
to `self.module_name`.
output_channels: Optional iterable of numbers of output channels.
kernel_shapes: O... | [
"Returns",
"transposed",
"version",
"of",
"this",
"network",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/convnet.py#L880-L994 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | _get_raw_data | def _get_raw_data(subset):
"""Loads the data or reads it from cache."""
raw_data = _LOADED.get(subset)
if raw_data is not None:
return raw_data, _LOADED["vocab"]
else:
train_data, valid_data, test_data, vocab = ptb_reader.ptb_raw_data(
FLAGS.data_path)
_LOADED.update({
"train": np.ar... | python | def _get_raw_data(subset):
"""Loads the data or reads it from cache."""
raw_data = _LOADED.get(subset)
if raw_data is not None:
return raw_data, _LOADED["vocab"]
else:
train_data, valid_data, test_data, vocab = ptb_reader.ptb_raw_data(
FLAGS.data_path)
_LOADED.update({
"train": np.ar... | [
"def",
"_get_raw_data",
"(",
"subset",
")",
":",
"raw_data",
"=",
"_LOADED",
".",
"get",
"(",
"subset",
")",
"if",
"raw_data",
"is",
"not",
"None",
":",
"return",
"raw_data",
",",
"_LOADED",
"[",
"\"vocab\"",
"]",
"else",
":",
"train_data",
",",
"valid_d... | Loads the data or reads it from cache. | [
"Loads",
"the",
"data",
"or",
"reads",
"it",
"from",
"cache",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L89-L103 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | custom_scale_mixture_prior_builder | def custom_scale_mixture_prior_builder(getter, name, *args, **kwargs):
"""A builder for the gaussian scale-mixture prior of Fortunato et al.
Please see https://arxiv.org/abs/1704.02798, section 7.1
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variabl... | python | def custom_scale_mixture_prior_builder(getter, name, *args, **kwargs):
"""A builder for the gaussian scale-mixture prior of Fortunato et al.
Please see https://arxiv.org/abs/1704.02798, section 7.1
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variabl... | [
"def",
"custom_scale_mixture_prior_builder",
"(",
"getter",
",",
"name",
",",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"# This specific prior formulation doesn't need any of the arguments forwarded",
"# from `get_variable`.",
"del",
"getter",
"del",
"name",
"del",
... | A builder for the gaussian scale-mixture prior of Fortunato et al.
Please see https://arxiv.org/abs/1704.02798, section 7.1
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variable`.
name: The `name` argument passed to `tf.get_variable`.
*args: Po... | [
"A",
"builder",
"for",
"the",
"gaussian",
"scale",
"-",
"mixture",
"prior",
"of",
"Fortunato",
"et",
"al",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L181-L204 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | lstm_posterior_builder | def lstm_posterior_builder(getter, name, *args, **kwargs):
"""A builder for a particular diagonal gaussian posterior.
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variable`.
name: The `name` argument passed to `tf.get_variable`.
*args: Positiona... | python | def lstm_posterior_builder(getter, name, *args, **kwargs):
"""A builder for a particular diagonal gaussian posterior.
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variable`.
name: The `name` argument passed to `tf.get_variable`.
*args: Positiona... | [
"def",
"lstm_posterior_builder",
"(",
"getter",
",",
"name",
",",
"*",
"args",
",",
"*",
"*",
"kwargs",
")",
":",
"del",
"args",
"parameter_shapes",
"=",
"tfp",
".",
"distributions",
".",
"Normal",
".",
"param_static_shapes",
"(",
"kwargs",
"[",
"\"shape\"",... | A builder for a particular diagonal gaussian posterior.
Args:
getter: The `getter` passed to a `custom_getter`. Please see the
documentation for `tf.get_variable`.
name: The `name` argument passed to `tf.get_variable`.
*args: Positional arguments forwarded by `tf.get_variable`.
**kwargs: Keywor... | [
"A",
"builder",
"for",
"a",
"particular",
"diagonal",
"gaussian",
"posterior",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L207-L245 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | build_modules | def build_modules(is_training, vocab_size):
"""Construct the modules used in the graph."""
# Construct the custom getter which implements Bayes by Backprop.
if is_training:
estimator_mode = tf.constant(bbb.EstimatorModes.sample)
else:
estimator_mode = tf.constant(bbb.EstimatorModes.mean)
lstm_bbb_cus... | python | def build_modules(is_training, vocab_size):
"""Construct the modules used in the graph."""
# Construct the custom getter which implements Bayes by Backprop.
if is_training:
estimator_mode = tf.constant(bbb.EstimatorModes.sample)
else:
estimator_mode = tf.constant(bbb.EstimatorModes.mean)
lstm_bbb_cus... | [
"def",
"build_modules",
"(",
"is_training",
",",
"vocab_size",
")",
":",
"# Construct the custom getter which implements Bayes by Backprop.",
"if",
"is_training",
":",
"estimator_mode",
"=",
"tf",
".",
"constant",
"(",
"bbb",
".",
"EstimatorModes",
".",
"sample",
")",
... | Construct the modules used in the graph. | [
"Construct",
"the",
"modules",
"used",
"in",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L289-L330 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | build_logits | def build_logits(data_ops, embed_layer, rnn_core, output_linear, name_prefix):
"""This is the core model logic.
Unrolls a Bayesian RNN over the given sequence.
Args:
data_ops: A `sequence_data.SequenceDataOps` namedtuple.
embed_layer: A `snt.Embed` instance.
rnn_core: A `snt.RNNCore` instance.
o... | python | def build_logits(data_ops, embed_layer, rnn_core, output_linear, name_prefix):
"""This is the core model logic.
Unrolls a Bayesian RNN over the given sequence.
Args:
data_ops: A `sequence_data.SequenceDataOps` namedtuple.
embed_layer: A `snt.Embed` instance.
rnn_core: A `snt.RNNCore` instance.
o... | [
"def",
"build_logits",
"(",
"data_ops",
",",
"embed_layer",
",",
"rnn_core",
",",
"output_linear",
",",
"name_prefix",
")",
":",
"# Embed the input index sequence.",
"embedded_input_seq",
"=",
"snt",
".",
"BatchApply",
"(",
"embed_layer",
",",
"name",
"=",
"\"input_... | This is the core model logic.
Unrolls a Bayesian RNN over the given sequence.
Args:
data_ops: A `sequence_data.SequenceDataOps` namedtuple.
embed_layer: A `snt.Embed` instance.
rnn_core: A `snt.RNNCore` instance.
output_linear: A `snt.Linear` instance.
name_prefix: A string to use to prefix lo... | [
"This",
"is",
"the",
"core",
"model",
"logic",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L333-L376 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | build_loss | def build_loss(model_logits, sparse_targets):
"""Compute the log loss given predictions and targets."""
time_major_shape = [FLAGS.unroll_steps, FLAGS.batch_size]
flat_batch_shape = [FLAGS.unroll_steps * FLAGS.batch_size, -1]
xent = tf.nn.sparse_softmax_cross_entropy_with_logits(
logits=tf.reshape(model_lo... | python | def build_loss(model_logits, sparse_targets):
"""Compute the log loss given predictions and targets."""
time_major_shape = [FLAGS.unroll_steps, FLAGS.batch_size]
flat_batch_shape = [FLAGS.unroll_steps * FLAGS.batch_size, -1]
xent = tf.nn.sparse_softmax_cross_entropy_with_logits(
logits=tf.reshape(model_lo... | [
"def",
"build_loss",
"(",
"model_logits",
",",
"sparse_targets",
")",
":",
"time_major_shape",
"=",
"[",
"FLAGS",
".",
"unroll_steps",
",",
"FLAGS",
".",
"batch_size",
"]",
"flat_batch_shape",
"=",
"[",
"FLAGS",
".",
"unroll_steps",
"*",
"FLAGS",
".",
"batch_s... | Compute the log loss given predictions and targets. | [
"Compute",
"the",
"log",
"loss",
"given",
"predictions",
"and",
"targets",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L379-L390 | train |
deepmind/sonnet | sonnet/examples/brnn_ptb.py | train | def train(logdir):
"""Run a network on the PTB training set, checkpointing the weights."""
ptb_train = PTB(
name="ptb_train",
subset="train",
seq_len=FLAGS.unroll_steps,
batch_size=FLAGS.batch_size)
# Connect to training set.
data_ops = ptb_train()
embed_layer, rnn_core, output_linea... | python | def train(logdir):
"""Run a network on the PTB training set, checkpointing the weights."""
ptb_train = PTB(
name="ptb_train",
subset="train",
seq_len=FLAGS.unroll_steps,
batch_size=FLAGS.batch_size)
# Connect to training set.
data_ops = ptb_train()
embed_layer, rnn_core, output_linea... | [
"def",
"train",
"(",
"logdir",
")",
":",
"ptb_train",
"=",
"PTB",
"(",
"name",
"=",
"\"ptb_train\"",
",",
"subset",
"=",
"\"train\"",
",",
"seq_len",
"=",
"FLAGS",
".",
"unroll_steps",
",",
"batch_size",
"=",
"FLAGS",
".",
"batch_size",
")",
"# Connect to ... | Run a network on the PTB training set, checkpointing the weights. | [
"Run",
"a",
"network",
"on",
"the",
"PTB",
"training",
"set",
"checkpointing",
"the",
"weights",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/brnn_ptb.py#L393-L541 | train |
deepmind/sonnet | sonnet/__init__.py | _ensure_dependency_available_at_version | def _ensure_dependency_available_at_version(package_name, min_version):
"""Throw helpful error if required dependencies not available."""
try:
pkg = importlib.import_module(package_name)
except ImportError:
pip_name = package_name.replace('_', '-')
raise SystemError(
'Sonnet requires %s (mini... | python | def _ensure_dependency_available_at_version(package_name, min_version):
"""Throw helpful error if required dependencies not available."""
try:
pkg = importlib.import_module(package_name)
except ImportError:
pip_name = package_name.replace('_', '-')
raise SystemError(
'Sonnet requires %s (mini... | [
"def",
"_ensure_dependency_available_at_version",
"(",
"package_name",
",",
"min_version",
")",
":",
"try",
":",
"pkg",
"=",
"importlib",
".",
"import_module",
"(",
"package_name",
")",
"except",
"ImportError",
":",
"pip_name",
"=",
"package_name",
".",
"replace",
... | Throw helpful error if required dependencies not available. | [
"Throw",
"helpful",
"error",
"if",
"required",
"dependencies",
"not",
"available",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/__init__.py#L43-L61 | train |
deepmind/sonnet | sonnet/python/modules/scale_gradient.py | _scale_gradient_op | def _scale_gradient_op(dtype):
"""Create an op that scales gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these ops
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is use... | python | def _scale_gradient_op(dtype):
"""Create an op that scales gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these ops
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is use... | [
"def",
"_scale_gradient_op",
"(",
"dtype",
")",
":",
"def",
"scale_gradient_backward",
"(",
"op",
",",
"grad",
")",
":",
"scale",
"=",
"op",
".",
"inputs",
"[",
"1",
"]",
"scaled_grad",
"=",
"grad",
"*",
"scale",
"return",
"scaled_grad",
",",
"None",
"# ... | Create an op that scales gradients using a Defun.
The tensorflow Defun decorator creates an op and tensorflow caches these ops
automatically according to `func_name`. Using a Defun decorator twice with the
same `func_name` does not create a new op, instead the cached op is used.
This method produces a new op ... | [
"Create",
"an",
"op",
"that",
"scales",
"gradients",
"using",
"a",
"Defun",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/scale_gradient.py#L27-L61 | train |
deepmind/sonnet | sonnet/python/modules/scale_gradient.py | scale_gradient | def scale_gradient(net, scale, name="scale_gradient"):
"""Scales gradients for the backwards pass.
This might be used to, for example, allow one part of a model to learn at a
lower rate than the rest.
WARNING: Think carefully about how your optimizer works. If, for example, you
use rmsprop, the gradient is ... | python | def scale_gradient(net, scale, name="scale_gradient"):
"""Scales gradients for the backwards pass.
This might be used to, for example, allow one part of a model to learn at a
lower rate than the rest.
WARNING: Think carefully about how your optimizer works. If, for example, you
use rmsprop, the gradient is ... | [
"def",
"scale_gradient",
"(",
"net",
",",
"scale",
",",
"name",
"=",
"\"scale_gradient\"",
")",
":",
"if",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"if",
"not",
"callable",
"(",
"net",
")",
":",
"raise",
"ValueError",
"(",
"\"In eager mode `net` must b... | Scales gradients for the backwards pass.
This might be used to, for example, allow one part of a model to learn at a
lower rate than the rest.
WARNING: Think carefully about how your optimizer works. If, for example, you
use rmsprop, the gradient is always rescaled (with some additional epsilon)
towards uni... | [
"Scales",
"gradients",
"for",
"the",
"backwards",
"pass",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/scale_gradient.py#L64-L113 | train |
deepmind/sonnet | sonnet/examples/rmc_learn_to_execute.py | build_and_train | def build_and_train(iterations, log_stride, test=False):
"""Construct the data, model, loss and optimizer then train."""
# Test mode settings.
batch_size = 2 if test else FLAGS.batch_size
num_mems = 2 if test else FLAGS.num_mems
num_heads = 1 if test else FLAGS.num_mems
num_blocks = 1 if test else FLAGS.nu... | python | def build_and_train(iterations, log_stride, test=False):
"""Construct the data, model, loss and optimizer then train."""
# Test mode settings.
batch_size = 2 if test else FLAGS.batch_size
num_mems = 2 if test else FLAGS.num_mems
num_heads = 1 if test else FLAGS.num_mems
num_blocks = 1 if test else FLAGS.nu... | [
"def",
"build_and_train",
"(",
"iterations",
",",
"log_stride",
",",
"test",
"=",
"False",
")",
":",
"# Test mode settings.",
"batch_size",
"=",
"2",
"if",
"test",
"else",
"FLAGS",
".",
"batch_size",
"num_mems",
"=",
"2",
"if",
"test",
"else",
"FLAGS",
".",
... | Construct the data, model, loss and optimizer then train. | [
"Construct",
"the",
"data",
"model",
"loss",
"and",
"optimizer",
"then",
"train",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/rmc_learn_to_execute.py#L110-L227 | train |
deepmind/sonnet | sonnet/examples/rmc_learn_to_execute.py | SequenceModel._build | def _build(
self, inputs, targets, input_sequence_length, output_sequence_length):
"""Dynamic unroll across input objects.
Args:
inputs: tensor (input_sequence_length x batch x feature_size). Encoder
sequence.
targets: tensor (output_sequence_length x batch x feature_size). Decoder
... | python | def _build(
self, inputs, targets, input_sequence_length, output_sequence_length):
"""Dynamic unroll across input objects.
Args:
inputs: tensor (input_sequence_length x batch x feature_size). Encoder
sequence.
targets: tensor (output_sequence_length x batch x feature_size). Decoder
... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"targets",
",",
"input_sequence_length",
",",
"output_sequence_length",
")",
":",
"# Connect decoding steps.",
"batch_size",
"=",
"inputs",
".",
"get_shape",
"(",
")",
"[",
"1",
"]",
"initial_state",
"=",
"self",... | Dynamic unroll across input objects.
Args:
inputs: tensor (input_sequence_length x batch x feature_size). Encoder
sequence.
targets: tensor (output_sequence_length x batch x feature_size). Decoder
sequence.
input_sequence_length: tensor (batch). Size of each batched input
... | [
"Dynamic",
"unroll",
"across",
"input",
"objects",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/rmc_learn_to_execute.py#L69-L107 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | lstm_with_recurrent_dropout | def lstm_with_recurrent_dropout(hidden_size, keep_prob=0.5, **kwargs):
"""LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob: the probability to keep an entry when applying dropout.
**kwargs: Extra keyword arguments to pass to the LSTM.
Returns:
A tuple (train_lstm, ... | python | def lstm_with_recurrent_dropout(hidden_size, keep_prob=0.5, **kwargs):
"""LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob: the probability to keep an entry when applying dropout.
**kwargs: Extra keyword arguments to pass to the LSTM.
Returns:
A tuple (train_lstm, ... | [
"def",
"lstm_with_recurrent_dropout",
"(",
"hidden_size",
",",
"keep_prob",
"=",
"0.5",
",",
"*",
"*",
"kwargs",
")",
":",
"lstm",
"=",
"LSTM",
"(",
"hidden_size",
",",
"*",
"*",
"kwargs",
")",
"return",
"RecurrentDropoutWrapper",
"(",
"lstm",
",",
"LSTMStat... | LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob: the probability to keep an entry when applying dropout.
**kwargs: Extra keyword arguments to pass to the LSTM.
Returns:
A tuple (train_lstm, test_lstm) where train_lstm is an LSTM with
recurrent dropout enabled to... | [
"LSTM",
"with",
"recurrent",
"dropout",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L433-L448 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | lstm_with_zoneout | def lstm_with_zoneout(hidden_size, keep_prob_c=0.5, keep_prob_h=0.95, **kwargs):
"""LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob_c: the probability to use the new value of the cell state rather
than freezing it.
keep_prob_h: the probability to use the new value ... | python | def lstm_with_zoneout(hidden_size, keep_prob_c=0.5, keep_prob_h=0.95, **kwargs):
"""LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob_c: the probability to use the new value of the cell state rather
than freezing it.
keep_prob_h: the probability to use the new value ... | [
"def",
"lstm_with_zoneout",
"(",
"hidden_size",
",",
"keep_prob_c",
"=",
"0.5",
",",
"keep_prob_h",
"=",
"0.95",
",",
"*",
"*",
"kwargs",
")",
":",
"lstm",
"=",
"LSTM",
"(",
"hidden_size",
",",
"*",
"*",
"kwargs",
")",
"keep_probs",
"=",
"LSTMState",
"("... | LSTM with recurrent dropout.
Args:
hidden_size: the LSTM hidden size.
keep_prob_c: the probability to use the new value of the cell state rather
than freezing it.
keep_prob_h: the probability to use the new value of the hidden state
rather than freezing it.
**kwargs: Extra keyword argumen... | [
"LSTM",
"with",
"recurrent",
"dropout",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L519-L540 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | highway_core_with_recurrent_dropout | def highway_core_with_recurrent_dropout(
hidden_size,
num_layers,
keep_prob=0.5,
**kwargs):
"""Highway core with recurrent dropout.
Args:
hidden_size: (int) Hidden size dimensionality.
num_layers: (int) Number of highway layers.
keep_prob: the probability to keep an entry when applying ... | python | def highway_core_with_recurrent_dropout(
hidden_size,
num_layers,
keep_prob=0.5,
**kwargs):
"""Highway core with recurrent dropout.
Args:
hidden_size: (int) Hidden size dimensionality.
num_layers: (int) Number of highway layers.
keep_prob: the probability to keep an entry when applying ... | [
"def",
"highway_core_with_recurrent_dropout",
"(",
"hidden_size",
",",
"num_layers",
",",
"keep_prob",
"=",
"0.5",
",",
"*",
"*",
"kwargs",
")",
":",
"core",
"=",
"HighwayCore",
"(",
"hidden_size",
",",
"num_layers",
",",
"*",
"*",
"kwargs",
")",
"return",
"... | Highway core with recurrent dropout.
Args:
hidden_size: (int) Hidden size dimensionality.
num_layers: (int) Number of highway layers.
keep_prob: the probability to keep an entry when applying dropout.
**kwargs: Extra keyword arguments to pass to the highway core.
Returns:
A tuple (train_core, ... | [
"Highway",
"core",
"with",
"recurrent",
"dropout",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1748-L1768 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | LSTM.get_possible_initializer_keys | def get_possible_initializer_keys(cls, use_peepholes=False,
use_projection=False):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates: bias of gates
w_f_... | python | def get_possible_initializer_keys(cls, use_peepholes=False,
use_projection=False):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates: bias of gates
w_f_... | [
"def",
"get_possible_initializer_keys",
"(",
"cls",
",",
"use_peepholes",
"=",
"False",
",",
"use_projection",
"=",
"False",
")",
":",
"possible_keys",
"=",
"cls",
".",
"POSSIBLE_INITIALIZER_KEYS",
".",
"copy",
"(",
")",
"if",
"not",
"use_peepholes",
":",
"possi... | Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates: bias of gates
w_f_diag: weight for prev_cell -> forget gate peephole
w_i_diag: weight for prev_cell -> input gate peephole
w_o_diag:... | [
"Returns",
"the",
"keys",
"the",
"dictionary",
"of",
"variable",
"initializers",
"may",
"contain",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L178-L207 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | LSTM._build | def _build(self, inputs, prev_state):
"""Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
the... | python | def _build(self, inputs, prev_state):
"""Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
the... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
")",
":",
"prev_hidden",
",",
"prev_cell",
"=",
"prev_state",
"# pylint: disable=invalid-unary-operand-type",
"if",
"self",
".",
"_hidden_clip_value",
"is",
"not",
"None",
":",
"prev_hidden",
"=",
"tf... | Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The batch s... | [
"Connects",
"the",
"LSTM",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L209-L276 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | LSTM._create_gate_variables | def _create_gate_variables(self, input_shape, dtype):
"""Initialize the variables used for the gates."""
if len(input_shape) != 2:
raise ValueError(
"Rank of shape must be {} not: {}".format(2, len(input_shape)))
equiv_input_size = self._hidden_state_size + input_shape.dims[1].value
ini... | python | def _create_gate_variables(self, input_shape, dtype):
"""Initialize the variables used for the gates."""
if len(input_shape) != 2:
raise ValueError(
"Rank of shape must be {} not: {}".format(2, len(input_shape)))
equiv_input_size = self._hidden_state_size + input_shape.dims[1].value
ini... | [
"def",
"_create_gate_variables",
"(",
"self",
",",
"input_shape",
",",
"dtype",
")",
":",
"if",
"len",
"(",
"input_shape",
")",
"!=",
"2",
":",
"raise",
"ValueError",
"(",
"\"Rank of shape must be {} not: {}\"",
".",
"format",
"(",
"2",
",",
"len",
"(",
"inp... | Initialize the variables used for the gates. | [
"Initialize",
"the",
"variables",
"used",
"for",
"the",
"gates",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L278-L310 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | LSTM._create_peephole_variables | def _create_peephole_variables(self, dtype):
"""Initialize the variables used for the peephole connections."""
self._w_f_diag = tf.get_variable(
self.W_F_DIAG,
shape=[self._hidden_size],
dtype=dtype,
initializer=self._initializers.get(self.W_F_DIAG),
partitioner=self._par... | python | def _create_peephole_variables(self, dtype):
"""Initialize the variables used for the peephole connections."""
self._w_f_diag = tf.get_variable(
self.W_F_DIAG,
shape=[self._hidden_size],
dtype=dtype,
initializer=self._initializers.get(self.W_F_DIAG),
partitioner=self._par... | [
"def",
"_create_peephole_variables",
"(",
"self",
",",
"dtype",
")",
":",
"self",
".",
"_w_f_diag",
"=",
"tf",
".",
"get_variable",
"(",
"self",
".",
"W_F_DIAG",
",",
"shape",
"=",
"[",
"self",
".",
"_hidden_size",
"]",
",",
"dtype",
"=",
"dtype",
",",
... | Initialize the variables used for the peephole connections. | [
"Initialize",
"the",
"variables",
"used",
"for",
"the",
"peephole",
"connections",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L312-L334 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | LSTM.state_size | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
return LSTMState(tf.TensorShape([self._hidden_state_size]),
tf.TensorShape([self._hidden_size])) | python | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
return LSTMState(tf.TensorShape([self._hidden_state_size]),
tf.TensorShape([self._hidden_size])) | [
"def",
"state_size",
"(",
"self",
")",
":",
"return",
"LSTMState",
"(",
"tf",
".",
"TensorShape",
"(",
"[",
"self",
".",
"_hidden_state_size",
"]",
")",
",",
"tf",
".",
"TensorShape",
"(",
"[",
"self",
".",
"_hidden_size",
"]",
")",
")"
] | Tuple of `tf.TensorShape`s indicating the size of state tensors. | [
"Tuple",
"of",
"tf",
".",
"TensorShape",
"s",
"indicating",
"the",
"size",
"of",
"state",
"tensors",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L337-L340 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | RecurrentDropoutWrapper.initial_state | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros."""
core_initial_state = self._core.initial_state(
batch_size, dtype=dtyp... | python | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros."""
core_initial_state = self._core.initial_state(
batch_size, dtype=dtyp... | [
"def",
"initial_state",
"(",
"self",
",",
"batch_size",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"trainable",
"=",
"False",
",",
"trainable_initializers",
"=",
"None",
",",
"trainable_regularizers",
"=",
"None",
",",
"name",
"=",
"None",
")",
":",
"co... | Builds the default start state tensor of zeros. | [
"Builds",
"the",
"default",
"start",
"state",
"tensor",
"of",
"zeros",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L404-L422 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | ZoneoutWrapper.initial_state | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros."""
return self._core.initial_state(
batch_size, dtype=dtype, trainable=t... | python | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros."""
return self._core.initial_state(
batch_size, dtype=dtype, trainable=t... | [
"def",
"initial_state",
"(",
"self",
",",
"batch_size",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"trainable",
"=",
"False",
",",
"trainable_initializers",
"=",
"None",
",",
"trainable_regularizers",
"=",
"None",
",",
"name",
"=",
"None",
")",
":",
"re... | Builds the default start state tensor of zeros. | [
"Builds",
"the",
"default",
"start",
"state",
"tensor",
"of",
"zeros",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L501-L508 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM.with_batch_norm_control | def with_batch_norm_control(self, is_training, test_local_stats=True):
"""Wraps this RNNCore with the additional control input to the `BatchNorm`s.
Example usage:
lstm = snt.BatchNormLSTM(4)
is_training = tf.placeholder(tf.bool)
rnn_input = ...
my_rnn = rnn.rnn(lstm.with_batch_norm_con... | python | def with_batch_norm_control(self, is_training, test_local_stats=True):
"""Wraps this RNNCore with the additional control input to the `BatchNorm`s.
Example usage:
lstm = snt.BatchNormLSTM(4)
is_training = tf.placeholder(tf.bool)
rnn_input = ...
my_rnn = rnn.rnn(lstm.with_batch_norm_con... | [
"def",
"with_batch_norm_control",
"(",
"self",
",",
"is_training",
",",
"test_local_stats",
"=",
"True",
")",
":",
"return",
"BatchNormLSTM",
".",
"CoreWithExtraBuildArgs",
"(",
"self",
",",
"is_training",
"=",
"is_training",
",",
"test_local_stats",
"=",
"test_loca... | Wraps this RNNCore with the additional control input to the `BatchNorm`s.
Example usage:
lstm = snt.BatchNormLSTM(4)
is_training = tf.placeholder(tf.bool)
rnn_input = ...
my_rnn = rnn.rnn(lstm.with_batch_norm_control(is_training), rnn_input)
Args:
is_training: Boolean that indic... | [
"Wraps",
"this",
"RNNCore",
"with",
"the",
"additional",
"control",
"input",
"to",
"the",
"BatchNorm",
"s",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L742-L767 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM.get_possible_initializer_keys | def get_possible_initializer_keys(
cls, use_peepholes=False, use_batch_norm_h=True, use_batch_norm_x=False,
use_batch_norm_c=False):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates:... | python | def get_possible_initializer_keys(
cls, use_peepholes=False, use_batch_norm_h=True, use_batch_norm_x=False,
use_batch_norm_c=False):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates:... | [
"def",
"get_possible_initializer_keys",
"(",
"cls",
",",
"use_peepholes",
"=",
"False",
",",
"use_batch_norm_h",
"=",
"True",
",",
"use_batch_norm_x",
"=",
"False",
",",
"use_batch_norm_c",
"=",
"False",
")",
":",
"possible_keys",
"=",
"cls",
".",
"POSSIBLE_INITIA... | Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
w_gates: weight for gates
b_gates: bias of gates
w_f_diag: weight for prev_cell -> forget gate peephole
w_i_diag: weight for prev_cell -> input gate peephole
w_o_diag:... | [
"Returns",
"the",
"keys",
"the",
"dictionary",
"of",
"variable",
"initializers",
"may",
"contain",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L770-L814 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM._build | def _build(self, inputs, prev_state, is_training=None, test_local_stats=True):
"""Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing var... | python | def _build(self, inputs, prev_state, is_training=None, test_local_stats=True):
"""Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing var... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
",",
"is_training",
"=",
"None",
",",
"test_local_stats",
"=",
"True",
")",
":",
"if",
"is_training",
"is",
"None",
":",
"raise",
"ValueError",
"(",
"\"Boolean is_training flag must be explicitly spec... | Connects the LSTM module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The batch s... | [
"Connects",
"the",
"LSTM",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L816-L920 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM._create_batch_norm_variables | def _create_batch_norm_variables(self, dtype):
"""Initialize the variables used for the `BatchNorm`s (if any)."""
# The paper recommends a value of 0.1 for good gradient flow through the
# tanh nonlinearity (although doesn't say whether this is for all gammas,
# or just some).
gamma_initializer = tf... | python | def _create_batch_norm_variables(self, dtype):
"""Initialize the variables used for the `BatchNorm`s (if any)."""
# The paper recommends a value of 0.1 for good gradient flow through the
# tanh nonlinearity (although doesn't say whether this is for all gammas,
# or just some).
gamma_initializer = tf... | [
"def",
"_create_batch_norm_variables",
"(",
"self",
",",
"dtype",
")",
":",
"# The paper recommends a value of 0.1 for good gradient flow through the",
"# tanh nonlinearity (although doesn't say whether this is for all gammas,",
"# or just some).",
"gamma_initializer",
"=",
"tf",
".",
... | Initialize the variables used for the `BatchNorm`s (if any). | [
"Initialize",
"the",
"variables",
"used",
"for",
"the",
"BatchNorm",
"s",
"(",
"if",
"any",
")",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L922-L959 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM._create_gate_variables | def _create_gate_variables(self, input_shape, dtype):
"""Initialize the variables used for the gates."""
if len(input_shape) != 2:
raise ValueError(
"Rank of shape must be {} not: {}".format(2, len(input_shape)))
input_size = input_shape.dims[1].value
b_shape = [4 * self._hidden_size]
... | python | def _create_gate_variables(self, input_shape, dtype):
"""Initialize the variables used for the gates."""
if len(input_shape) != 2:
raise ValueError(
"Rank of shape must be {} not: {}".format(2, len(input_shape)))
input_size = input_shape.dims[1].value
b_shape = [4 * self._hidden_size]
... | [
"def",
"_create_gate_variables",
"(",
"self",
",",
"input_shape",
",",
"dtype",
")",
":",
"if",
"len",
"(",
"input_shape",
")",
"!=",
"2",
":",
"raise",
"ValueError",
"(",
"\"Rank of shape must be {} not: {}\"",
".",
"format",
"(",
"2",
",",
"len",
"(",
"inp... | Initialize the variables used for the gates. | [
"Initialize",
"the",
"variables",
"used",
"for",
"the",
"gates",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L961-L1002 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM.initial_state | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros.
Args:
batch_size: An int, float or scalar Tensor representing the batch s... | python | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state tensor of zeros.
Args:
batch_size: An int, float or scalar Tensor representing the batch s... | [
"def",
"initial_state",
"(",
"self",
",",
"batch_size",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"trainable",
"=",
"False",
",",
"trainable_initializers",
"=",
"None",
",",
"trainable_regularizers",
"=",
"None",
",",
"name",
"=",
"None",
")",
":",
"if... | Builds the default start state tensor of zeros.
Args:
batch_size: An int, float or scalar Tensor representing the batch size.
dtype: The data type to use for the state.
trainable: Boolean that indicates whether to learn the initial state.
trainable_initializers: An optional pair of initiali... | [
"Builds",
"the",
"default",
"start",
"state",
"tensor",
"of",
"zeros",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1028-L1074 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | BatchNormLSTM.state_size | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
if self._max_unique_stats == 1:
return (tf.TensorShape([self._hidden_size]),
tf.TensorShape([self._hidden_size]))
else:
return (tf.TensorShape([self._hidden_size]),
tf.TensorS... | python | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
if self._max_unique_stats == 1:
return (tf.TensorShape([self._hidden_size]),
tf.TensorShape([self._hidden_size]))
else:
return (tf.TensorShape([self._hidden_size]),
tf.TensorS... | [
"def",
"state_size",
"(",
"self",
")",
":",
"if",
"self",
".",
"_max_unique_stats",
"==",
"1",
":",
"return",
"(",
"tf",
".",
"TensorShape",
"(",
"[",
"self",
".",
"_hidden_size",
"]",
")",
",",
"tf",
".",
"TensorShape",
"(",
"[",
"self",
".",
"_hidd... | Tuple of `tf.TensorShape`s indicating the size of state tensors. | [
"Tuple",
"of",
"tf",
".",
"TensorShape",
"s",
"indicating",
"the",
"size",
"of",
"state",
"tensors",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1077-L1085 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | ConvLSTM._new_convolution | def _new_convolution(self, use_bias):
"""Returns new convolution.
Args:
use_bias: Use bias in convolutions. If False, clean_dict removes bias
entries from initializers, partitioners and regularizers passed to
the constructor of the convolution.
"""
def clean_dict(input_dict):
... | python | def _new_convolution(self, use_bias):
"""Returns new convolution.
Args:
use_bias: Use bias in convolutions. If False, clean_dict removes bias
entries from initializers, partitioners and regularizers passed to
the constructor of the convolution.
"""
def clean_dict(input_dict):
... | [
"def",
"_new_convolution",
"(",
"self",
",",
"use_bias",
")",
":",
"def",
"clean_dict",
"(",
"input_dict",
")",
":",
"if",
"input_dict",
"and",
"not",
"use_bias",
":",
"cleaned_dict",
"=",
"input_dict",
".",
"copy",
"(",
")",
"cleaned_dict",
".",
"pop",
"(... | Returns new convolution.
Args:
use_bias: Use bias in convolutions. If False, clean_dict removes bias
entries from initializers, partitioners and regularizers passed to
the constructor of the convolution. | [
"Returns",
"new",
"convolution",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1321-L1345 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | ConvLSTM.state_size | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
hidden_size = tf.TensorShape(
self._input_shape[:-1] + (self._output_channels,))
return (hidden_size, hidden_size) | python | def state_size(self):
"""Tuple of `tf.TensorShape`s indicating the size of state tensors."""
hidden_size = tf.TensorShape(
self._input_shape[:-1] + (self._output_channels,))
return (hidden_size, hidden_size) | [
"def",
"state_size",
"(",
"self",
")",
":",
"hidden_size",
"=",
"tf",
".",
"TensorShape",
"(",
"self",
".",
"_input_shape",
"[",
":",
"-",
"1",
"]",
"+",
"(",
"self",
".",
"_output_channels",
",",
")",
")",
"return",
"(",
"hidden_size",
",",
"hidden_si... | Tuple of `tf.TensorShape`s indicating the size of state tensors. | [
"Tuple",
"of",
"tf",
".",
"TensorShape",
"s",
"indicating",
"the",
"size",
"of",
"state",
"tensors",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1352-L1356 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | GRU._build | def _build(self, inputs, prev_state):
"""Connects the GRU module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
thei... | python | def _build(self, inputs, prev_state):
"""Connects the GRU module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
thei... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
")",
":",
"input_size",
"=",
"inputs",
".",
"get_shape",
"(",
")",
"[",
"1",
"]",
"weight_shape",
"=",
"(",
"input_size",
",",
"self",
".",
"_hidden_size",
")",
"u_shape",
"=",
"(",
"self"... | Connects the GRU module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as inputs and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The batch si... | [
"Connects",
"the",
"GRU",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1509-L1583 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | HighwayCore.get_possible_initializer_keys | def get_possible_initializer_keys(cls, num_layers):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
wt: weight for input -> T gate
wh: weight for input -> H gate
wtL: weight for prev state -> T gate for layer L (indexed fr... | python | def get_possible_initializer_keys(cls, num_layers):
"""Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
wt: weight for input -> T gate
wh: weight for input -> H gate
wtL: weight for prev state -> T gate for layer L (indexed fr... | [
"def",
"get_possible_initializer_keys",
"(",
"cls",
",",
"num_layers",
")",
":",
"keys",
"=",
"[",
"cls",
".",
"WT",
",",
"cls",
".",
"WH",
"]",
"for",
"layer_index",
"in",
"xrange",
"(",
"num_layers",
")",
":",
"layer_str",
"=",
"str",
"(",
"layer_index... | Returns the keys the dictionary of variable initializers may contain.
The set of all possible initializer keys are:
wt: weight for input -> T gate
wh: weight for input -> H gate
wtL: weight for prev state -> T gate for layer L (indexed from 0)
whL: weight for prev state -> H gate for layer ... | [
"Returns",
"the",
"keys",
"the",
"dictionary",
"of",
"variable",
"initializers",
"may",
"contain",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1661-L1686 | train |
deepmind/sonnet | sonnet/python/modules/gated_rnn.py | HighwayCore._build | def _build(self, inputs, prev_state):
"""Connects the highway core module into the graph.
Args:
inputs: Tensor of size `[batch_size, input_size]`.
prev_state: Tensor of size `[batch_size, hidden_size]`.
Returns:
A tuple (output, next_state) where `output` is a Tensor of size
`[batc... | python | def _build(self, inputs, prev_state):
"""Connects the highway core module into the graph.
Args:
inputs: Tensor of size `[batch_size, input_size]`.
prev_state: Tensor of size `[batch_size, hidden_size]`.
Returns:
A tuple (output, next_state) where `output` is a Tensor of size
`[batc... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
")",
":",
"input_size",
"=",
"inputs",
".",
"get_shape",
"(",
")",
"[",
"1",
"]",
"weight_shape",
"=",
"(",
"input_size",
",",
"self",
".",
"_hidden_size",
")",
"u_shape",
"=",
"(",
"self"... | Connects the highway core module into the graph.
Args:
inputs: Tensor of size `[batch_size, input_size]`.
prev_state: Tensor of size `[batch_size, hidden_size]`.
Returns:
A tuple (output, next_state) where `output` is a Tensor of size
`[batch_size, hidden_size]` and `next_state` is a T... | [
"Connects",
"the",
"highway",
"core",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/gated_rnn.py#L1688-L1737 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | _get_flat_core_sizes | def _get_flat_core_sizes(cores):
"""Obtains the list flattened output sizes of a list of cores.
Args:
cores: list of cores to get the shapes from.
Returns:
List of lists that, for each core, contains the list of its output
dimensions.
"""
core_sizes_lists = []
for core in cores:
flat_out... | python | def _get_flat_core_sizes(cores):
"""Obtains the list flattened output sizes of a list of cores.
Args:
cores: list of cores to get the shapes from.
Returns:
List of lists that, for each core, contains the list of its output
dimensions.
"""
core_sizes_lists = []
for core in cores:
flat_out... | [
"def",
"_get_flat_core_sizes",
"(",
"cores",
")",
":",
"core_sizes_lists",
"=",
"[",
"]",
"for",
"core",
"in",
"cores",
":",
"flat_output_size",
"=",
"nest",
".",
"flatten",
"(",
"core",
".",
"output_size",
")",
"core_sizes_lists",
".",
"append",
"(",
"[",
... | Obtains the list flattened output sizes of a list of cores.
Args:
cores: list of cores to get the shapes from.
Returns:
List of lists that, for each core, contains the list of its output
dimensions. | [
"Obtains",
"the",
"list",
"flattened",
"output",
"sizes",
"of",
"a",
"list",
"of",
"cores",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L39-L54 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | _get_shape_without_batch_dimension | def _get_shape_without_batch_dimension(tensor_nest):
"""Converts Tensor nest to a TensorShape nest, removing batch dimension."""
def _strip_batch_and_convert_to_shape(tensor):
return tensor.get_shape()[1:]
return nest.map_structure(_strip_batch_and_convert_to_shape, tensor_nest) | python | def _get_shape_without_batch_dimension(tensor_nest):
"""Converts Tensor nest to a TensorShape nest, removing batch dimension."""
def _strip_batch_and_convert_to_shape(tensor):
return tensor.get_shape()[1:]
return nest.map_structure(_strip_batch_and_convert_to_shape, tensor_nest) | [
"def",
"_get_shape_without_batch_dimension",
"(",
"tensor_nest",
")",
":",
"def",
"_strip_batch_and_convert_to_shape",
"(",
"tensor",
")",
":",
"return",
"tensor",
".",
"get_shape",
"(",
")",
"[",
"1",
":",
"]",
"return",
"nest",
".",
"map_structure",
"(",
"_str... | Converts Tensor nest to a TensorShape nest, removing batch dimension. | [
"Converts",
"Tensor",
"nest",
"to",
"a",
"TensorShape",
"nest",
"removing",
"batch",
"dimension",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L57-L61 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | VanillaRNN._build | def _build(self, input_, prev_state):
"""Connects the VanillaRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
... | python | def _build(self, input_, prev_state):
"""Connects the VanillaRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
... | [
"def",
"_build",
"(",
"self",
",",
"input_",
",",
"prev_state",
")",
":",
"self",
".",
"_in_to_hidden_linear",
"=",
"basic",
".",
"Linear",
"(",
"self",
".",
"_hidden_size",
",",
"name",
"=",
"\"in_to_hidden\"",
",",
"initializers",
"=",
"self",
".",
"_ini... | Connects the VanillaRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The b... | [
"Connects",
"the",
"VanillaRNN",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L110-L149 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | DeepRNN._check_cores_output_sizes | def _check_cores_output_sizes(self):
"""Checks the output_sizes of the cores of the DeepRNN module.
Raises:
ValueError: if the outputs of the cores cannot be concatenated along their
first dimension.
"""
for core_sizes in zip(*tuple(_get_flat_core_sizes(self._cores))):
first_core_li... | python | def _check_cores_output_sizes(self):
"""Checks the output_sizes of the cores of the DeepRNN module.
Raises:
ValueError: if the outputs of the cores cannot be concatenated along their
first dimension.
"""
for core_sizes in zip(*tuple(_get_flat_core_sizes(self._cores))):
first_core_li... | [
"def",
"_check_cores_output_sizes",
"(",
"self",
")",
":",
"for",
"core_sizes",
"in",
"zip",
"(",
"*",
"tuple",
"(",
"_get_flat_core_sizes",
"(",
"self",
".",
"_cores",
")",
")",
")",
":",
"first_core_list",
"=",
"core_sizes",
"[",
"0",
"]",
"[",
"1",
":... | Checks the output_sizes of the cores of the DeepRNN module.
Raises:
ValueError: if the outputs of the cores cannot be concatenated along their
first dimension. | [
"Checks",
"the",
"output_sizes",
"of",
"the",
"cores",
"of",
"the",
"DeepRNN",
"module",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L294-L309 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | DeepRNN._build | def _build(self, inputs, prev_state, **kwargs):
"""Connects the DeepRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct siz... | python | def _build(self, inputs, prev_state, **kwargs):
"""Connects the DeepRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct siz... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
",",
"*",
"*",
"kwargs",
")",
":",
"current_input",
"=",
"inputs",
"next_states",
"=",
"[",
"]",
"outputs",
"=",
"[",
"]",
"recurrent_idx",
"=",
"0",
"concatenate",
"=",
"lambda",
"*",
"ar... | Connects the DeepRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The batc... | [
"Connects",
"the",
"DeepRNN",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L311-L370 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | DeepRNN.initial_state | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state for a DeepRNN.
Args:
batch_size: An int, float or scalar Tensor representing the batch siz... | python | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state for a DeepRNN.
Args:
batch_size: An int, float or scalar Tensor representing the batch siz... | [
"def",
"initial_state",
"(",
"self",
",",
"batch_size",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"trainable",
"=",
"False",
",",
"trainable_initializers",
"=",
"None",
",",
"trainable_regularizers",
"=",
"None",
",",
"name",
"=",
"None",
")",
":",
"in... | Builds the default start state for a DeepRNN.
Args:
batch_size: An int, float or scalar Tensor representing the batch size.
dtype: The data type to use for the state.
trainable: Boolean that indicates whether to learn the initial state.
trainable_initializers: An initializer function or nes... | [
"Builds",
"the",
"default",
"start",
"state",
"for",
"a",
"DeepRNN",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L372-L426 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | ModelRNN._build | def _build(self, inputs, prev_state):
"""Connects the ModelRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
... | python | def _build(self, inputs, prev_state):
"""Connects the ModelRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"prev_state",
")",
":",
"next_state",
"=",
"self",
".",
"_model",
"(",
"prev_state",
")",
"# For ModelRNN, the next state of the RNN is the same as the output",
"return",
"next_state",
",",
"next_state"
] | Connects the ModelRNN module into the graph.
If this is not the first time the module has been connected to the graph,
the Tensors provided as input_ and state must have the same final
dimension, in order for the existing variables to be the correct size for
their corresponding multiplications. The bat... | [
"Connects",
"the",
"ModelRNN",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L517-L537 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | BidirectionalRNN._build | def _build(self, input_sequence, state):
"""Connects the BidirectionalRNN module into the graph.
Args:
input_sequence: tensor (time, batch, [feature_1, ..]). It must be
time_major.
state: tuple of states for the forward and backward cores.
Returns:
A dict with forward/backard s... | python | def _build(self, input_sequence, state):
"""Connects the BidirectionalRNN module into the graph.
Args:
input_sequence: tensor (time, batch, [feature_1, ..]). It must be
time_major.
state: tuple of states for the forward and backward cores.
Returns:
A dict with forward/backard s... | [
"def",
"_build",
"(",
"self",
",",
"input_sequence",
",",
"state",
")",
":",
"input_shape",
"=",
"input_sequence",
".",
"get_shape",
"(",
")",
"if",
"input_shape",
"[",
"0",
"]",
"is",
"None",
":",
"raise",
"ValueError",
"(",
"\"Time dimension of input (dim 0)... | Connects the BidirectionalRNN module into the graph.
Args:
input_sequence: tensor (time, batch, [feature_1, ..]). It must be
time_major.
state: tuple of states for the forward and backward cores.
Returns:
A dict with forward/backard states and output sequences:
"outputs":{... | [
"Connects",
"the",
"BidirectionalRNN",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L583-L649 | train |
deepmind/sonnet | sonnet/python/modules/basic_rnn.py | BidirectionalRNN.initial_state | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state for a BidirectionalRNN.
The Bidirectional RNN flattens the states of its forward and backward co... | python | def initial_state(self, batch_size, dtype=tf.float32, trainable=False,
trainable_initializers=None, trainable_regularizers=None,
name=None):
"""Builds the default start state for a BidirectionalRNN.
The Bidirectional RNN flattens the states of its forward and backward co... | [
"def",
"initial_state",
"(",
"self",
",",
"batch_size",
",",
"dtype",
"=",
"tf",
".",
"float32",
",",
"trainable",
"=",
"False",
",",
"trainable_initializers",
"=",
"None",
",",
"trainable_regularizers",
"=",
"None",
",",
"name",
"=",
"None",
")",
":",
"na... | Builds the default start state for a BidirectionalRNN.
The Bidirectional RNN flattens the states of its forward and backward cores
and concatentates them.
Args:
batch_size: An int, float or scalar Tensor representing the batch size.
dtype: The data type to use for the state.
trainable: B... | [
"Builds",
"the",
"default",
"start",
"state",
"for",
"a",
"BidirectionalRNN",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/basic_rnn.py#L651-L686 | train |
deepmind/sonnet | sonnet/python/modules/nets/mlp.py | MLP._instantiate_layers | def _instantiate_layers(self):
"""Instantiates all the linear modules used in the network.
Layers are instantiated in the constructor, as opposed to the build
function, because MLP implements the Transposable interface, and the
transpose function can be called before the module is actually connected
... | python | def _instantiate_layers(self):
"""Instantiates all the linear modules used in the network.
Layers are instantiated in the constructor, as opposed to the build
function, because MLP implements the Transposable interface, and the
transpose function can be called before the module is actually connected
... | [
"def",
"_instantiate_layers",
"(",
"self",
")",
":",
"# Here we are entering the module's variable scope to name our submodules",
"# correctly (not to create variables). As such it's safe to not check",
"# whether we're in the same graph. This is important if we're constructing",
"# the module in ... | Instantiates all the linear modules used in the network.
Layers are instantiated in the constructor, as opposed to the build
function, because MLP implements the Transposable interface, and the
transpose function can be called before the module is actually connected
to the graph and build is called.
... | [
"Instantiates",
"all",
"the",
"linear",
"modules",
"used",
"in",
"the",
"network",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/mlp.py#L112-L139 | train |
deepmind/sonnet | sonnet/python/modules/nets/mlp.py | MLP._build | def _build(self, inputs, is_training=True, dropout_keep_prob=0.5):
"""Assembles the `MLP` and connects it to the graph.
Args:
inputs: A 2D Tensor of size `[batch_size, input_size]`.
is_training: A bool or tf.Bool Tensor. Indicates whether we are
currently training. Defaults to `True`.
... | python | def _build(self, inputs, is_training=True, dropout_keep_prob=0.5):
"""Assembles the `MLP` and connects it to the graph.
Args:
inputs: A 2D Tensor of size `[batch_size, input_size]`.
is_training: A bool or tf.Bool Tensor. Indicates whether we are
currently training. Defaults to `True`.
... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"is_training",
"=",
"True",
",",
"dropout_keep_prob",
"=",
"0.5",
")",
":",
"self",
".",
"_input_shape",
"=",
"tuple",
"(",
"inputs",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
")",
"net",
... | Assembles the `MLP` and connects it to the graph.
Args:
inputs: A 2D Tensor of size `[batch_size, input_size]`.
is_training: A bool or tf.Bool Tensor. Indicates whether we are
currently training. Defaults to `True`.
dropout_keep_prob: The probability that each element is kept when
... | [
"Assembles",
"the",
"MLP",
"and",
"connects",
"it",
"to",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/mlp.py#L145-L174 | train |
deepmind/sonnet | sonnet/python/modules/nets/mlp.py | MLP.output_sizes | def output_sizes(self):
"""Returns a tuple of all output sizes of all the layers."""
return tuple([l() if callable(l) else l for l in self._output_sizes]) | python | def output_sizes(self):
"""Returns a tuple of all output sizes of all the layers."""
return tuple([l() if callable(l) else l for l in self._output_sizes]) | [
"def",
"output_sizes",
"(",
"self",
")",
":",
"return",
"tuple",
"(",
"[",
"l",
"(",
")",
"if",
"callable",
"(",
"l",
")",
"else",
"l",
"for",
"l",
"in",
"self",
".",
"_output_sizes",
"]",
")"
] | Returns a tuple of all output sizes of all the layers. | [
"Returns",
"a",
"tuple",
"of",
"all",
"output",
"sizes",
"of",
"all",
"the",
"layers",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/mlp.py#L182-L184 | train |
deepmind/sonnet | sonnet/python/modules/nets/mlp.py | MLP.transpose | def transpose(self, name=None, activate_final=None):
"""Returns transposed `MLP`.
Args:
name: Optional string specifying the name of the transposed module. The
default name is constructed by appending "_transpose"
to `self.module_name`.
activate_final: Optional boolean determining i... | python | def transpose(self, name=None, activate_final=None):
"""Returns transposed `MLP`.
Args:
name: Optional string specifying the name of the transposed module. The
default name is constructed by appending "_transpose"
to `self.module_name`.
activate_final: Optional boolean determining i... | [
"def",
"transpose",
"(",
"self",
",",
"name",
"=",
"None",
",",
"activate_final",
"=",
"None",
")",
":",
"if",
"name",
"is",
"None",
":",
"name",
"=",
"self",
".",
"module_name",
"+",
"\"_transpose\"",
"if",
"activate_final",
"is",
"None",
":",
"activate... | Returns transposed `MLP`.
Args:
name: Optional string specifying the name of the transposed module. The
default name is constructed by appending "_transpose"
to `self.module_name`.
activate_final: Optional boolean determining if the activation and batch
normalization, if turned ... | [
"Returns",
"transposed",
"MLP",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/mlp.py#L237-L265 | train |
deepmind/sonnet | sonnet/python/modules/nets/mlp.py | MLP.clone | def clone(self, name=None):
"""Creates a new MLP with the same structure.
Args:
name: Optional string specifying the name of the new module. The default
name is constructed by appending "_clone" to the original name.
Returns:
A cloned `MLP` module.
"""
if name is None:
n... | python | def clone(self, name=None):
"""Creates a new MLP with the same structure.
Args:
name: Optional string specifying the name of the new module. The default
name is constructed by appending "_clone" to the original name.
Returns:
A cloned `MLP` module.
"""
if name is None:
n... | [
"def",
"clone",
"(",
"self",
",",
"name",
"=",
"None",
")",
":",
"if",
"name",
"is",
"None",
":",
"name",
"=",
"self",
".",
"module_name",
"+",
"\"_clone\"",
"return",
"MLP",
"(",
"name",
"=",
"name",
",",
"output_sizes",
"=",
"self",
".",
"output_si... | Creates a new MLP with the same structure.
Args:
name: Optional string specifying the name of the new module. The default
name is constructed by appending "_clone" to the original name.
Returns:
A cloned `MLP` module. | [
"Creates",
"a",
"new",
"MLP",
"with",
"the",
"same",
"structure",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/mlp.py#L267-L289 | train |
deepmind/sonnet | sonnet/python/modules/nets/alexnet.py | AlexNet._calc_min_size | def _calc_min_size(self, conv_layers):
"""Calculates the minimum size of the input layer.
Given a set of convolutional layers, calculate the minimum value of
the `input_height` and `input_width`, i.e. such that the output has
size 1x1. Assumes snt.VALID padding.
Args:
conv_layers: List of tu... | python | def _calc_min_size(self, conv_layers):
"""Calculates the minimum size of the input layer.
Given a set of convolutional layers, calculate the minimum value of
the `input_height` and `input_width`, i.e. such that the output has
size 1x1. Assumes snt.VALID padding.
Args:
conv_layers: List of tu... | [
"def",
"_calc_min_size",
"(",
"self",
",",
"conv_layers",
")",
":",
"input_size",
"=",
"1",
"for",
"_",
",",
"conv_params",
",",
"max_pooling",
"in",
"reversed",
"(",
"conv_layers",
")",
":",
"if",
"max_pooling",
"is",
"not",
"None",
":",
"kernel_size",
",... | Calculates the minimum size of the input layer.
Given a set of convolutional layers, calculate the minimum value of
the `input_height` and `input_width`, i.e. such that the output has
size 1x1. Assumes snt.VALID padding.
Args:
conv_layers: List of tuples `(output_channels, (kernel_size, stride),... | [
"Calculates",
"the",
"minimum",
"size",
"of",
"the",
"input",
"layer",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/alexnet.py#L154-L179 | train |
deepmind/sonnet | sonnet/python/modules/nets/alexnet.py | AlexNet._build | def _build(self, inputs, keep_prob=None, is_training=None,
test_local_stats=True):
"""Connects the AlexNet module into the graph.
The is_training flag only controls the batch norm settings, if `False` it
does not force no dropout by overriding any input `keep_prob`. To avoid any
confusion ... | python | def _build(self, inputs, keep_prob=None, is_training=None,
test_local_stats=True):
"""Connects the AlexNet module into the graph.
The is_training flag only controls the batch norm settings, if `False` it
does not force no dropout by overriding any input `keep_prob`. To avoid any
confusion ... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
",",
"keep_prob",
"=",
"None",
",",
"is_training",
"=",
"None",
",",
"test_local_stats",
"=",
"True",
")",
":",
"# Check input shape",
"if",
"(",
"self",
".",
"_use_batch_norm",
"or",
"keep_prob",
"is",
"not",
"... | Connects the AlexNet module into the graph.
The is_training flag only controls the batch norm settings, if `False` it
does not force no dropout by overriding any input `keep_prob`. To avoid any
confusion this may cause, if `is_training=False` and `keep_prob` would cause
dropout to be applied, an error ... | [
"Connects",
"the",
"AlexNet",
"module",
"into",
"the",
"graph",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/nets/alexnet.py#L181-L291 | train |
deepmind/sonnet | sonnet/examples/dataset_nth_farthest.py | NthFarthest._get_single_set | def _get_single_set(self, num_objects, num_features):
"""Generate one input sequence and output label.
Each sequences of objects has a feature that consists of the feature vector
for that object plus the encoding for its ID, the reference vector ID and
the n-th value relative ID for a total feature siz... | python | def _get_single_set(self, num_objects, num_features):
"""Generate one input sequence and output label.
Each sequences of objects has a feature that consists of the feature vector
for that object plus the encoding for its ID, the reference vector ID and
the n-th value relative ID for a total feature siz... | [
"def",
"_get_single_set",
"(",
"self",
",",
"num_objects",
",",
"num_features",
")",
":",
"# Generate random binary vectors",
"data",
"=",
"np",
".",
"random",
".",
"uniform",
"(",
"-",
"1",
",",
"1",
",",
"size",
"=",
"(",
"num_objects",
",",
"num_features"... | Generate one input sequence and output label.
Each sequences of objects has a feature that consists of the feature vector
for that object plus the encoding for its ID, the reference vector ID and
the n-th value relative ID for a total feature size of:
`num_objects` * 3 + `num_features`
Args:
... | [
"Generate",
"one",
"input",
"sequence",
"and",
"output",
"label",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/dataset_nth_farthest.py#L51-L97 | train |
deepmind/sonnet | sonnet/examples/dataset_nth_farthest.py | NthFarthest._get_batch_data | def _get_batch_data(self, batch_size, num_objects, num_features):
"""Assembles a batch of input tensors and output labels.
Args:
batch_size: int. number of sequence batches.
num_objects: int. number of objects in the sequence.
num_features: int. feature size of each object.
Returns:
... | python | def _get_batch_data(self, batch_size, num_objects, num_features):
"""Assembles a batch of input tensors and output labels.
Args:
batch_size: int. number of sequence batches.
num_objects: int. number of objects in the sequence.
num_features: int. feature size of each object.
Returns:
... | [
"def",
"_get_batch_data",
"(",
"self",
",",
"batch_size",
",",
"num_objects",
",",
"num_features",
")",
":",
"all_inputs",
"=",
"[",
"]",
"all_labels",
"=",
"[",
"]",
"for",
"_",
"in",
"six",
".",
"moves",
".",
"range",
"(",
"batch_size",
")",
":",
"in... | Assembles a batch of input tensors and output labels.
Args:
batch_size: int. number of sequence batches.
num_objects: int. number of objects in the sequence.
num_features: int. feature size of each object.
Returns:
1. np.ndarray (`batch_size`, `num_objects`,
(`num_... | [
"Assembles",
"a",
"batch",
"of",
"input",
"tensors",
"and",
"output",
"labels",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/dataset_nth_farthest.py#L99-L120 | train |
deepmind/sonnet | sonnet/examples/dataset_nth_farthest.py | NthFarthest.get_batch | def get_batch(self):
"""Returns set of nth-farthest input tensors and labels.
Returns:
1. tf.Tensor (`batch_size`, `num_objects`,
(`num_features` + 3 * `num_objects`)).
2. tf.Tensor (`batch_size`). Output object reference label.
"""
params = [self._batch_size, self._num... | python | def get_batch(self):
"""Returns set of nth-farthest input tensors and labels.
Returns:
1. tf.Tensor (`batch_size`, `num_objects`,
(`num_features` + 3 * `num_objects`)).
2. tf.Tensor (`batch_size`). Output object reference label.
"""
params = [self._batch_size, self._num... | [
"def",
"get_batch",
"(",
"self",
")",
":",
"params",
"=",
"[",
"self",
".",
"_batch_size",
",",
"self",
".",
"_num_objects",
",",
"self",
".",
"_num_features",
"]",
"inputs",
",",
"labels",
"=",
"tf",
".",
"py_func",
"(",
"self",
".",
"_get_batch_data",
... | Returns set of nth-farthest input tensors and labels.
Returns:
1. tf.Tensor (`batch_size`, `num_objects`,
(`num_features` + 3 * `num_objects`)).
2. tf.Tensor (`batch_size`). Output object reference label. | [
"Returns",
"set",
"of",
"nth",
"-",
"farthest",
"input",
"tensors",
"and",
"labels",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/examples/dataset_nth_farthest.py#L122-L136 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _default_transpose_size | def _default_transpose_size(input_shape, stride, kernel_shape=None,
padding=SAME):
"""Returns default (maximal) output shape for a transpose convolution.
In general, there are multiple possible output shapes that a transpose
convolution with a given `input_shape` can map to. This func... | python | def _default_transpose_size(input_shape, stride, kernel_shape=None,
padding=SAME):
"""Returns default (maximal) output shape for a transpose convolution.
In general, there are multiple possible output shapes that a transpose
convolution with a given `input_shape` can map to. This func... | [
"def",
"_default_transpose_size",
"(",
"input_shape",
",",
"stride",
",",
"kernel_shape",
"=",
"None",
",",
"padding",
"=",
"SAME",
")",
":",
"if",
"not",
"input_shape",
":",
"raise",
"TypeError",
"(",
"\"input_shape is None; if using Sonnet, are you sure you \"",
"\"... | Returns default (maximal) output shape for a transpose convolution.
In general, there are multiple possible output shapes that a transpose
convolution with a given `input_shape` can map to. This function returns the
output shape which evenly divides the stride to produce the input shape in
a forward convolutio... | [
"Returns",
"default",
"(",
"maximal",
")",
"output",
"shape",
"for",
"a",
"transpose",
"convolution",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L63-L107 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _fill_shape | def _fill_shape(x, n):
"""Converts a dimension to a tuple of dimensions of a given size.
This is used to allow shorthand notation for various configuration parameters.
A user can provide either, for example, `2` or `[2, 2]` as a kernel shape, and
this function returns `(2, 2)` in both cases. Passing `[1, 2]` w... | python | def _fill_shape(x, n):
"""Converts a dimension to a tuple of dimensions of a given size.
This is used to allow shorthand notation for various configuration parameters.
A user can provide either, for example, `2` or `[2, 2]` as a kernel shape, and
this function returns `(2, 2)` in both cases. Passing `[1, 2]` w... | [
"def",
"_fill_shape",
"(",
"x",
",",
"n",
")",
":",
"if",
"not",
"isinstance",
"(",
"n",
",",
"numbers",
".",
"Integral",
")",
"or",
"n",
"<",
"1",
":",
"raise",
"TypeError",
"(",
"\"n must be a positive integer\"",
")",
"if",
"(",
"isinstance",
"(",
"... | Converts a dimension to a tuple of dimensions of a given size.
This is used to allow shorthand notation for various configuration parameters.
A user can provide either, for example, `2` or `[2, 2]` as a kernel shape, and
this function returns `(2, 2)` in both cases. Passing `[1, 2]` will return
`(1, 2)`.
Ar... | [
"Converts",
"a",
"dimension",
"to",
"a",
"tuple",
"of",
"dimensions",
"of",
"a",
"given",
"size",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L110-L145 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _fill_and_verify_parameter_shape | def _fill_and_verify_parameter_shape(x, n, parameter_label):
"""Expands x if necessary into a `n`-D kernel shape and reports errors."""
try:
return _fill_shape(x, n)
except TypeError as e:
raise base.IncompatibleShapeError("Invalid " + parameter_label + " shape: "
"{}... | python | def _fill_and_verify_parameter_shape(x, n, parameter_label):
"""Expands x if necessary into a `n`-D kernel shape and reports errors."""
try:
return _fill_shape(x, n)
except TypeError as e:
raise base.IncompatibleShapeError("Invalid " + parameter_label + " shape: "
"{}... | [
"def",
"_fill_and_verify_parameter_shape",
"(",
"x",
",",
"n",
",",
"parameter_label",
")",
":",
"try",
":",
"return",
"_fill_shape",
"(",
"x",
",",
"n",
")",
"except",
"TypeError",
"as",
"e",
":",
"raise",
"base",
".",
"IncompatibleShapeError",
"(",
"\"Inva... | Expands x if necessary into a `n`-D kernel shape and reports errors. | [
"Expands",
"x",
"if",
"necessary",
"into",
"a",
"n",
"-",
"D",
"kernel",
"shape",
"and",
"reports",
"errors",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L148-L154 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _fill_and_verify_padding | def _fill_and_verify_padding(padding, n):
"""Verifies that the provided padding is supported and expands to size n.
Args:
padding: One of ALLOWED_PADDINGS, or an iterable of them.
n: An integer, the size of the desired output list.
Returns:
If `padding` is one of ALLOWED_PADDINGS, a tuple of size `n... | python | def _fill_and_verify_padding(padding, n):
"""Verifies that the provided padding is supported and expands to size n.
Args:
padding: One of ALLOWED_PADDINGS, or an iterable of them.
n: An integer, the size of the desired output list.
Returns:
If `padding` is one of ALLOWED_PADDINGS, a tuple of size `n... | [
"def",
"_fill_and_verify_padding",
"(",
"padding",
",",
"n",
")",
":",
"if",
"not",
"isinstance",
"(",
"n",
",",
"numbers",
".",
"Integral",
")",
"or",
"n",
"<",
"1",
":",
"raise",
"TypeError",
"(",
"\"n must be a positive integer\"",
")",
"if",
"isinstance"... | Verifies that the provided padding is supported and expands to size n.
Args:
padding: One of ALLOWED_PADDINGS, or an iterable of them.
n: An integer, the size of the desired output list.
Returns:
If `padding` is one of ALLOWED_PADDINGS, a tuple of size `n` containing `n`
copies of `padding`.
I... | [
"Verifies",
"that",
"the",
"provided",
"padding",
"is",
"supported",
"and",
"expands",
"to",
"size",
"n",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L176-L206 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _padding_to_conv_op_padding | def _padding_to_conv_op_padding(padding):
"""Whether to use SAME or VALID for the underlying convolution op.
Args:
padding: A tuple of members of ALLOWED_PADDINGS, e.g. as returned from
`_fill_and_verify_padding`.
Returns:
One of CONV_OP_ALLOWED_PADDINGS, the padding method to use for the
unde... | python | def _padding_to_conv_op_padding(padding):
"""Whether to use SAME or VALID for the underlying convolution op.
Args:
padding: A tuple of members of ALLOWED_PADDINGS, e.g. as returned from
`_fill_and_verify_padding`.
Returns:
One of CONV_OP_ALLOWED_PADDINGS, the padding method to use for the
unde... | [
"def",
"_padding_to_conv_op_padding",
"(",
"padding",
")",
":",
"if",
"not",
"isinstance",
"(",
"padding",
",",
"tuple",
")",
":",
"raise",
"ValueError",
"(",
"\"padding should be a tuple.\"",
")",
"if",
"all",
"(",
"p",
"==",
"SAME",
"for",
"p",
"in",
"padd... | Whether to use SAME or VALID for the underlying convolution op.
Args:
padding: A tuple of members of ALLOWED_PADDINGS, e.g. as returned from
`_fill_and_verify_padding`.
Returns:
One of CONV_OP_ALLOWED_PADDINGS, the padding method to use for the
underlying convolution op.
Raises:
ValueErro... | [
"Whether",
"to",
"use",
"SAME",
"or",
"VALID",
"for",
"the",
"underlying",
"convolution",
"op",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L209-L233 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _fill_and_one_pad_stride | def _fill_and_one_pad_stride(stride, n, data_format=DATA_FORMAT_NHWC):
"""Expands the provided stride to size n and pads it with 1s."""
if isinstance(stride, numbers.Integral) or (
isinstance(stride, collections.Iterable) and len(stride) <= n):
if data_format.startswith("NC"):
return (1, 1,) + _fill... | python | def _fill_and_one_pad_stride(stride, n, data_format=DATA_FORMAT_NHWC):
"""Expands the provided stride to size n and pads it with 1s."""
if isinstance(stride, numbers.Integral) or (
isinstance(stride, collections.Iterable) and len(stride) <= n):
if data_format.startswith("NC"):
return (1, 1,) + _fill... | [
"def",
"_fill_and_one_pad_stride",
"(",
"stride",
",",
"n",
",",
"data_format",
"=",
"DATA_FORMAT_NHWC",
")",
":",
"if",
"isinstance",
"(",
"stride",
",",
"numbers",
".",
"Integral",
")",
"or",
"(",
"isinstance",
"(",
"stride",
",",
"collections",
".",
"Iter... | Expands the provided stride to size n and pads it with 1s. | [
"Expands",
"the",
"provided",
"stride",
"to",
"size",
"n",
"and",
"pads",
"it",
"with",
"1s",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L236-L253 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _verify_inputs | def _verify_inputs(inputs, channel_index, data_format):
"""Verifies `inputs` is semantically correct.
Args:
inputs: An input tensor provided by the user.
channel_index: The index of the channel dimension.
data_format: The format of the data in `inputs`.
Raises:
base.IncompatibleShapeError: If th... | python | def _verify_inputs(inputs, channel_index, data_format):
"""Verifies `inputs` is semantically correct.
Args:
inputs: An input tensor provided by the user.
channel_index: The index of the channel dimension.
data_format: The format of the data in `inputs`.
Raises:
base.IncompatibleShapeError: If th... | [
"def",
"_verify_inputs",
"(",
"inputs",
",",
"channel_index",
",",
"data_format",
")",
":",
"# Check shape.",
"input_shape",
"=",
"tuple",
"(",
"inputs",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
")",
"if",
"len",
"(",
"input_shape",
")",
"!=",... | Verifies `inputs` is semantically correct.
Args:
inputs: An input tensor provided by the user.
channel_index: The index of the channel dimension.
data_format: The format of the data in `inputs`.
Raises:
base.IncompatibleShapeError: If the shape of `inputs` doesn't match
`data_format`.
ba... | [
"Verifies",
"inputs",
"is",
"semantically",
"correct",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L256-L292 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | create_weight_initializer | def create_weight_initializer(fan_in_shape, dtype=tf.float32):
"""Returns a default initializer for the weights of a convolutional module."""
stddev = 1 / math.sqrt(np.prod(fan_in_shape))
return tf.truncated_normal_initializer(stddev=stddev, dtype=dtype) | python | def create_weight_initializer(fan_in_shape, dtype=tf.float32):
"""Returns a default initializer for the weights of a convolutional module."""
stddev = 1 / math.sqrt(np.prod(fan_in_shape))
return tf.truncated_normal_initializer(stddev=stddev, dtype=dtype) | [
"def",
"create_weight_initializer",
"(",
"fan_in_shape",
",",
"dtype",
"=",
"tf",
".",
"float32",
")",
":",
"stddev",
"=",
"1",
"/",
"math",
".",
"sqrt",
"(",
"np",
".",
"prod",
"(",
"fan_in_shape",
")",
")",
"return",
"tf",
".",
"truncated_normal_initiali... | Returns a default initializer for the weights of a convolutional module. | [
"Returns",
"a",
"default",
"initializer",
"for",
"the",
"weights",
"of",
"a",
"convolutional",
"module",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L295-L298 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _find_channel_index | def _find_channel_index(data_format):
"""Returns the index of the channel dimension.
Args:
data_format: A string of characters corresponding to Tensor dimensionality.
Returns:
channel_index: An integer indicating the channel dimension.
Raises:
ValueError: If no channel dimension was found.
"""
... | python | def _find_channel_index(data_format):
"""Returns the index of the channel dimension.
Args:
data_format: A string of characters corresponding to Tensor dimensionality.
Returns:
channel_index: An integer indicating the channel dimension.
Raises:
ValueError: If no channel dimension was found.
"""
... | [
"def",
"_find_channel_index",
"(",
"data_format",
")",
":",
"for",
"i",
",",
"c",
"in",
"enumerate",
"(",
"data_format",
")",
":",
"if",
"c",
"==",
"\"C\"",
":",
"return",
"i",
"raise",
"ValueError",
"(",
"\"data_format requires a channel dimension. Got: {}\"",
... | Returns the index of the channel dimension.
Args:
data_format: A string of characters corresponding to Tensor dimensionality.
Returns:
channel_index: An integer indicating the channel dimension.
Raises:
ValueError: If no channel dimension was found. | [
"Returns",
"the",
"index",
"of",
"the",
"channel",
"dimension",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L306-L322 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _apply_bias | def _apply_bias(inputs, outputs, channel_index, data_format, output_channels,
initializers, partitioners, regularizers):
"""Initialize and apply a bias to the outputs.
Figures out the shape of the bias vector, initialize it, and applies it.
Args:
inputs: A Tensor of shape `data_format`.
... | python | def _apply_bias(inputs, outputs, channel_index, data_format, output_channels,
initializers, partitioners, regularizers):
"""Initialize and apply a bias to the outputs.
Figures out the shape of the bias vector, initialize it, and applies it.
Args:
inputs: A Tensor of shape `data_format`.
... | [
"def",
"_apply_bias",
"(",
"inputs",
",",
"outputs",
",",
"channel_index",
",",
"data_format",
",",
"output_channels",
",",
"initializers",
",",
"partitioners",
",",
"regularizers",
")",
":",
"bias_shape",
"=",
"(",
"output_channels",
",",
")",
"if",
"\"b\"",
... | Initialize and apply a bias to the outputs.
Figures out the shape of the bias vector, initialize it, and applies it.
Args:
inputs: A Tensor of shape `data_format`.
outputs: A Tensor of shape `data_format`.
channel_index: The index of the channel dimension in `inputs`.
data_format: Format of `input... | [
"Initialize",
"and",
"apply",
"a",
"bias",
"to",
"the",
"outputs",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L325-L369 | train |
deepmind/sonnet | sonnet/python/modules/conv.py | _ConvND._build | def _build(self, inputs):
"""Connects the _ConvND module into the graph, with input Tensor `inputs`.
If this is not the first time the module has been connected to the graph,
the input Tensor provided here must have the same number of channels, in
order for the existing variables to be the correct size... | python | def _build(self, inputs):
"""Connects the _ConvND module into the graph, with input Tensor `inputs`.
If this is not the first time the module has been connected to the graph,
the input Tensor provided here must have the same number of channels, in
order for the existing variables to be the correct size... | [
"def",
"_build",
"(",
"self",
",",
"inputs",
")",
":",
"_verify_inputs",
"(",
"inputs",
",",
"self",
".",
"_channel_index",
",",
"self",
".",
"_data_format",
")",
"self",
".",
"_input_shape",
"=",
"tuple",
"(",
"inputs",
".",
"get_shape",
"(",
")",
".",
... | Connects the _ConvND module into the graph, with input Tensor `inputs`.
If this is not the first time the module has been connected to the graph,
the input Tensor provided here must have the same number of channels, in
order for the existing variables to be the correct size for the
multiplication; the ... | [
"Connects",
"the",
"_ConvND",
"module",
"into",
"the",
"graph",
"with",
"input",
"Tensor",
"inputs",
"."
] | 00612ca3178964d86b556e062694d808ff81fcca | https://github.com/deepmind/sonnet/blob/00612ca3178964d86b556e062694d808ff81fcca/sonnet/python/modules/conv.py#L520-L570 | train |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.