diff options
author | A. Unique TensorFlower <gardener@tensorflow.org> | 2017-02-12 16:34:17 -0800 |
---|---|---|
committer | TensorFlower Gardener <gardener@tensorflow.org> | 2017-02-12 16:52:39 -0800 |
commit | 90eaf9539ced75d868336d2ae29476e01a0b5086 (patch) | |
tree | fe9f25e6ca02c246fe42f7d03c5ad97ccdf050e3 /tensorflow/g3doc | |
parent | b4d475bf0966ac148f61d7f42bd9b46155bb04f6 (diff) |
Update generated Python Op docs.
Change: 147297663
Diffstat (limited to 'tensorflow/g3doc')
62 files changed, 1811 insertions, 1857 deletions
diff --git a/tensorflow/g3doc/api_docs/python/contrib.distributions.bijector.md b/tensorflow/g3doc/api_docs/python/contrib.distributions.bijector.md index 61ba4b2521..eef1e88ac5 100644 --- a/tensorflow/g3doc/api_docs/python/contrib.distributions.bijector.md +++ b/tensorflow/g3doc/api_docs/python/contrib.distributions.bijector.md @@ -11,7 +11,7 @@ An API for invertible, differentiable transformations of random variables. Differentiable, bijective transformations of continuous random variables alter the calculations made in the cumulative/probability distribution functions and -sample function. This module provides a standard interface for making these +sample function. This module provides a standard interface for making these manipulations. For more details and examples, see the `Bijector` docstring. @@ -109,34 +109,34 @@ specified then `scale += IdentityMatrix`. Otherwise specifying a ##### Args: -* <b>`shift`</b>: Numeric `Tensor`. If this is set to `None`, no shift is applied. +* <b>`shift`</b>: Floating-point `Tensor`. If this is set to `None`, no shift is + applied. * <b>`scale_identity_multiplier`</b>: floating point rank 0 `Tensor` representing a scaling done to the identity matrix. When `scale_identity_multiplier = scale_diag = scale_tril = None` then `scale += IdentityMatrix`. Otherwise no scaled-identity-matrix is added to `scale`. -* <b>`scale_diag`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_diag` has shape [N1, N2, ... k], which represents a k x k +* <b>`scale_diag`</b>: Floating-point `Tensor` representing the diagonal matrix. + `scale_diag` has shape [N1, N2, ... k], which represents a k x k diagonal matrix. When `None` no diagonal term is added to `scale`. -* <b>`scale_tril`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_diag` has shape [N1, N2, ... k, k], which represents a k x k +* <b>`scale_tril`</b>: Floating-point `Tensor` representing the diagonal matrix. + `scale_diag` has shape [N1, N2, ... k, k], which represents a k x k lower triangular matrix. When `None` no `scale_tril` term is added to `scale`. The upper triangular elements above the diagonal are ignored. -* <b>`scale_perturb_factor`</b>: Numeric `Tensor` representing factor matrix with - last two dimensions of shape `(k, r)`. - When `None`, no rank-r update is added to `scale`. -* <b>`scale_perturb_diag`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_perturb_diag` has shape [N1, N2, ... r], which represents an - r x r Diagonal matrix. - When `None` low rank updates will take the form `scale_perturb_factor * - scale_perturb_factor.T`. +* <b>`scale_perturb_factor`</b>: Floating-point `Tensor` representing factor matrix + with last two dimensions of shape `(k, r)`. When `None`, no rank-r + update is added to `scale`. +* <b>`scale_perturb_diag`</b>: Floating-point `Tensor` representing the diagonal + matrix. `scale_perturb_diag` has shape [N1, N2, ... r], which + represents an `r x r` diagonal matrix. When `None` low rank updates will + take the form `scale_perturb_factor * scale_perturb_factor.T`. * <b>`event_ndims`</b>: Scalar `int32` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 1. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -391,7 +391,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -435,7 +436,7 @@ If `X` is a scalar then the forward transformation is: `scale * X + shift` where `*` denotes the scalar product. Note: we don't always simply transpose `X` (but write it this way for -brevity). Actually the input `X` undergoes the following transformation +brevity). Actually the input `X` undergoes the following transformation before being premultiplied by `scale`: 1. If there are no sample dims, we call `X = tf.expand_dims(X, 0)`, i.e., @@ -450,8 +451,8 @@ before being premultiplied by `scale`: (For more details see `shape.make_batch_of_event_sample_matrices`.) The result of the above transformation is that `X` can be regarded as a batch -of matrices where each column is a draw from the distribution. After -premultiplying by `scale`, we take the inverse of this procedure. The input +of matrices where each column is a draw from the distribution. After +premultiplying by `scale`, we take the inverse of this procedure. The input `Y` also undergoes the same transformation before/after premultiplying by `inv(scale)`. @@ -489,14 +490,14 @@ Instantiates the `AffineLinearOperator` bijector. ##### Args: -* <b>`shift`</b>: Numeric `Tensor`. -* <b>`scale`</b>: Subclass of `LinearOperator`. Represents the (batch) positive +* <b>`shift`</b>: Floating-point `Tensor`. +* <b>`scale`</b>: Subclass of `LinearOperator`. Represents the (batch) positive definite matrix `M` in `R^{k x k}`. * <b>`event_ndims`</b>: Scalar `integer` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 1. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -753,7 +754,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -795,7 +797,7 @@ A `Bijector` implements a [diffeomorphism](https://en.wikipedia.org/wiki/Diffeomorphism), i.e., a bijective, differentiable function. A `Bijector` is used by `TransformedDistribution` but can be generally used for transforming a -`Distribution` generated `Tensor`. A `Bijector` is characterized by three +`Distribution` generated `Tensor`. A `Bijector` is characterized by three operations: 1. Forward Evaluation @@ -813,7 +815,7 @@ operations: "The log of the determinant of the matrix of all first-order partial derivatives of the inverse function." Useful for inverting a transformation to compute one probability in terms - of another. Geometrically, the det(Jacobian) is the volume of the + of another. Geometrically, the det(Jacobian) is the volume of the transformation and is used to scale the probability. By convention, transformations of random variables are named in terms of the @@ -825,7 +827,7 @@ Example Use: - Basic properties: ```python - x = ... # A tensor. + x = ... # A tensor. # Evaluate forward transformation. fwd_x = my_bijector.forward(x) x == my_bijector.inverse(fwd_x) @@ -882,7 +884,7 @@ Example transformations: if self.event_ndims is None: raise ValueError("Jacobian requires known event_ndims.") event_dims = array_ops.shape(x)[-self.event_ndims:] - return math_ops.reduce_sum(x, reduction_indices=event_dims) + return math_ops.reduce_sum(x, axis=event_dims) ``` - "Affine" @@ -907,8 +909,8 @@ Example of why a `Bijector` needs to understand sample, batch, event partitioning: - Consider the `Exp` `Bijector` applied to a `Tensor` which has sample, batch, - and event (S, B, E) shape semantics. Suppose - the `Tensor`'s partitioned-shape is `(S=[4], B=[2], E=[3, 3])`. + and event (S, B, E) shape semantics. Suppose the `Tensor`'s + partitioned-shape is `(S=[4], B=[2], E=[3, 3])`. For `Exp`, the shape of the `Tensor` returned by `forward` and `inverse` is unchanged, i.e., `[4, 2, 3, 3]`. However the shape returned by @@ -923,7 +925,7 @@ Subclass Requirements: - If the `Bijector`'s use is limited to `TransformedDistribution` (or friends like `QuantizedDistribution`) then depending on your use, you may not need - to implement all of `_forward` and `_inverse` functions. Examples: + to implement all of `_forward` and `_inverse` functions. Examples: 1. Sampling (e.g., `sample`) only requires `_forward`. 2. Probability functions (e.g., `prob`, `cdf`, `survival`) only require `_inverse` (and related). @@ -931,7 +933,7 @@ Subclass Requirements: `_inverse` can be implemented as a cache lookup. See `Example Use` [above] which shows how these functions are used to - transform a distribution. (Note: `_forward` could theoretically be + transform a distribution. (Note: `_forward` could theoretically be implemented as a cache lookup but this would require controlling the underlying sample generation mechanism.) @@ -949,7 +951,7 @@ Subclass Requirements: - Subclasses should implement `_forward_event_shape`, `_forward_event_shape_tensor` (and `inverse` counterparts) if the - transformation is shape-changing. By default the event-shape is assumed + transformation is shape-changing. By default the event-shape is assumed unchanged from input. Tips for implementing `_inverse` and `_inverse_log_det_jacobian`: @@ -958,14 +960,14 @@ Tips for implementing `_inverse` and `_inverse_log_det_jacobian`: can be implemented as a cache lookup. - The inverse `log o det o Jacobian` can be implemented as the negative of the - forward `log o det o Jacobian`. This is useful if the `inverse` is + forward `log o det o Jacobian`. This is useful if the `inverse` is implemented as a cache or the inverse Jacobian is computationally more expensive (e.g., `CholeskyOuterProduct` `Bijector`). The following demonstrates the suggested implementation. ```python def _inverse_and_log_det_jacobian(self, y): - x = # ... implement inverse, possibly via cache. + x = ... # implement inverse, possibly via cache. return x, -self._forward_log_det_jac(x) # Note negation. ``` @@ -1024,10 +1026,10 @@ See `Bijector` subclass docstring for more details and specific examples. * <b>`event_ndims`</b>: number of dimensions associated with event coordinates. * <b>`graph_parents`</b>: Python list of graph prerequisites of this `Bijector`. -* <b>`is_constant_jacobian`</b>: `Boolean` indicating that the Jacobian is not a +* <b>`is_constant_jacobian`</b>: Python `bool` indicating that the Jacobian is not a function of the input. -* <b>`validate_args`</b>: `Boolean`, default `False`. Whether to validate input with - asserts. If `validate_args` is `False`, and the inputs are invalid, +* <b>`validate_args`</b>: Python `bool`, default `False`. Whether to validate input + with asserts. If `validate_args` is `False`, and the inputs are invalid, correct behavior is not guaranteed. * <b>`dtype`</b>: `tf.dtype` supported by this `Bijector`. `None` means dtype is not enforced. @@ -1280,7 +1282,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -1344,10 +1347,10 @@ Instantiates `Chain` bijector. * <b>`bijectors`</b>: Python list of bijector instances. An empty list makes this bijector equivalent to the `Identity` bijector. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. Default: E.g., - `Chain([Exp(), Softplus()]).name == "chain_of_exp_of_softplus"`. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. Default: + E.g., `Chain([Exp(), Softplus()]).name == "chain_of_exp_of_softplus"`. ##### Raises: @@ -1608,7 +1611,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -1657,9 +1661,9 @@ Instantiates the `CholeskyOuterProduct` bijector. * <b>`event_ndims`</b>: `constant` `int32` scalar `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 2. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -1913,7 +1917,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -1964,9 +1969,9 @@ Instantiates the `Exp` bijector. * <b>`event_ndims`</b>: Scalar `int32` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. - - - @@ -2215,7 +2220,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -2509,7 +2515,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -2540,7 +2547,7 @@ exp = Inline( forward_fn=tf.exp, inverse_fn=tf.log, inverse_log_det_jacobian_fn=( - lambda y: -tf.reduce_sum(tf.log(y), reduction_indices=-1)), + lambda y: -tf.reduce_sum(tf.log(y), axis=-1)), name="exp") ``` @@ -2568,11 +2575,11 @@ Creates a `Bijector` from callables. static event shape changes. Default: shape is assumed unchanged. * <b>`inverse_event_shape_tensor_fn`</b>: Python callable implementing non-identical event shape changes. Default: shape is assumed unchanged. -* <b>`is_constant_jacobian`</b>: `Boolean` indicating that the Jacobian is constant - for all input arguments. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. +* <b>`is_constant_jacobian`</b>: Python `bool` indicating that the Jacobian is + constant for all input arguments. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. - - - @@ -2821,7 +2828,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -2874,9 +2882,9 @@ return -self.inverse_log_det_jacobian(y, **kwargs) * <b>`bijector`</b>: Bijector instance. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. - - - @@ -3132,7 +3140,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -3174,9 +3183,9 @@ Instantiates the `PowerTransform` bijector. `Y = g(X) = (1 + X * c)**(1 / c)` where `c` is the `power`. * <b>`event_ndims`</b>: Python scalar indicating the number of dimensions associated with a particular draw from the distribution. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -3430,7 +3439,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -3717,7 +3727,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -3743,7 +3754,7 @@ Bijector which computes `Y = g(X) = exp([X 0]) / sum(exp([X 0]))`. To implement [softmax](https://en.wikipedia.org/wiki/Softmax_function) as a bijection, the forward transformation appends a value to the input and the -inverse removes this coordinate. The appended coordinate represents a pivot, +inverse removes this coordinate. The appended coordinate represents a pivot, e.g., `softmax(x) = exp(x-c) / sum(exp(x-c))` where `c` is the implicit last coordinate. @@ -3764,7 +3775,7 @@ bijector.SoftmaxCentered(event_ndims=1).inverse([0.2, 0.3, 0.4, 0.1]) At first blush it may seem like the [Invariance of domain]( https://en.wikipedia.org/wiki/Invariance_of_domain) theorem implies this -implementation is not a bijection. However, the appended dimension +implementation is not a bijection. However, the appended dimension makes the (forward) image non-open and the theorem does not directly apply. - - - @@ -4019,7 +4030,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - @@ -4318,7 +4330,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/contrib.distributions.md b/tensorflow/g3doc/api_docs/python/contrib.distributions.md index e1f4518456..85a18c8c50 100644 --- a/tensorflow/g3doc/api_docs/python/contrib.distributions.md +++ b/tensorflow/g3doc/api_docs/python/contrib.distributions.md @@ -26,7 +26,7 @@ one of two possible properties for samples from a distribution: `NOT_REPARAMETERIZED`: Samples from the distribution are not fully reparameterized, and straight-through gradients are either partially - unsupported or are not supported at all. In this case, for purposes of + unsupported or are not supported at all. In this case, for purposes of e.g. RL or variational inference, it is generally safest to wrap the sample results in a `stop_gradients` call and instead use policy gradients / surrogate loss instead. @@ -76,8 +76,8 @@ A generic probability distribution base class. ### Subclassing Subclasses are expected to implement a leading-underscore version of the -same-named function. The argument signature should be identical except for -the omission of `name="..."`. For example, to enable `log_prob(value, +same-named function. The argument signature should be identical except for +the omission of `name="..."`. For example, to enable `log_prob(value, name="log_prob")` a subclass should implement `_log_prob(value)`. Subclasses can append to public-level docstrings by providing @@ -90,7 +90,7 @@ def _log_prob(self, value): ``` would add the string "Some other details." to the `log_prob` function -docstring. This is implemented as a simple decorator to avoid python +docstring. This is implemented as a simple decorator to avoid python linter complaining about missing Args/Returns/Raises sections in the partial docstrings. @@ -103,7 +103,7 @@ The shape of arguments to `__init__`, `cdf`, `log_cdf`, `prob`, and `log_prob` reflect this broadcasting, as does the return value of `sample` and `sample_n`. -`sample_n_shape = (n,) + batch_shape + event_shape`, where `sample_n_shape` is +`sample_n_shape = [n] + batch_shape + event_shape`, where `sample_n_shape` is the shape of the `Tensor` returned from `sample_n`, `n` is the number of samples, `batch_shape` defines how many independent distributions there are, and `event_shape` defines the shape of samples from each of those independent @@ -128,19 +128,19 @@ event_shape = u.event_shape # `event_shape_t` is a `Tensor` which will evaluate to []. event_shape_t = u.event_shape_tensor() -# Sampling returns a sample per distribution. `samples` has shape -# (5, 2, 2), which is (n,) + batch_shape + event_shape, where n=5, -# batch_shape=(2, 2), and event_shape=(). +# Sampling returns a sample per distribution. `samples` has shape +# [5, 2, 2], which is [n] + batch_shape + event_shape, where n=5, +# batch_shape=[2, 2], and event_shape=[]. samples = u.sample_n(5) # The broadcasting holds across methods. Here we use `cdf` as an example. The # same holds for `log_cdf` and the likelihood functions. -# `cum_prob` has shape (2, 2) as the `value` argument was broadcasted to the +# `cum_prob` has shape [2, 2] as the `value` argument was broadcasted to the # shape of the `Uniform` instance. cum_prob_broadcast = u.cdf(4.0) -# `cum_prob`'s shape is (2, 2), one per distribution. No broadcasting +# `cum_prob`'s shape is [2, 2], one per distribution. No broadcasting # occurred. cum_prob_per_dist = u.cdf([[4.0, 5.0], [6.0, 7.0]]) @@ -153,9 +153,9 @@ cum_prob_invalid = u.cdf([4.0, 5.0, 6.0]) ### Parameter values leading to undefined statistics or distributions. Some distributions do not have well-defined statistics for all initialization -parameter values. For example, the beta distribution is parameterized by -positive real numbers `a` and `b`, and does not have well-defined mode if -`a < 1` or `b < 1`. +parameter values. For example, the beta distribution is parameterized by +positive real numbers `concentration1` and `concentration0`, and does not have +well-defined mode if `concentration1 < 1` or `concentration0 < 1`. The user is given the option of raising an exception or returning `NaN`. @@ -192,25 +192,28 @@ Constructs the `Distribution`. * <b>`dtype`</b>: The type of the event samples. `None` implies no type-enforcement. -* <b>`is_continuous`</b>: Python boolean. If `True` this - `Distribution` is continuous over its supported domain. +* <b>`is_continuous`</b>: Python `bool`. If `True` this `Distribution` is continuous + over its supported domain. * <b>`reparameterization_type`</b>: Instance of `ReparameterizationType`. If `distributions.FULLY_REPARAMETERIZED`, this `Distribution` can be reparameterized in terms of some standard distribution with a function whose Jacobian is constant for the support - of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, + of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, then no such reparameterization is available. -* <b>`validate_args`</b>: Python boolean. Whether to validate input with asserts. - If `validate_args` is `False`, and the inputs are invalid, - correct behavior is not guaranteed. -* <b>`allow_nan_stats`</b>: Python boolean. If `False`, raise an - exception if a statistic (e.g., mean, mode) is undefined for any batch - member. If True, batch members with valid parameters leading to - undefined statistics will return `NaN` for this statistic. -* <b>`parameters`</b>: Python dictionary of parameters used to instantiate this +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution + parameters are checked for validity despite possibly degrading runtime + performance. When `False` invalid inputs may silently render incorrect + outputs. +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics + (e.g., mean, mode, variance) use the value "`NaN`" to indicate the + result is undefined. When `False`, an exception is raised if one or + more of the statistic's batch members are undefined. +* <b>`parameters`</b>: Python `dict` of parameters used to instantiate this + `Distribution`. +* <b>`graph_parents`</b>: Python `list` of graph prerequisites of this `Distribution`. -* <b>`graph_parents`</b>: Python list of graph prerequisites of this `Distribution`. -* <b>`name`</b>: A name for this distribution. Default: subclass name. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: + subclass name. ##### Raises: @@ -222,21 +225,20 @@ Constructs the `Distribution`. #### `tf.contrib.distributions.Distribution.allow_nan_stats` {#Distribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -434,7 +436,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -451,7 +453,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -584,8 +586,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -726,7 +728,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Distribution.validate_args` {#Distribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -838,7 +840,7 @@ Initialize a batch of Binomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor with shape broadcastable to `[N1,..., Nm]` with `m >= 0` and the same dtype as `probs` or - `logits`. Defines this as a batch of `N1 x ... x Nm` different Binomial + `logits`. Defines this as a batch of `N1 x ... x Nm` different Binomial distributions. Its components should be equal to integer values. * <b>`logits`</b>: Floating point tensor representing the log-odds of a positive event with shape broadcastable to `[N1,..., Nm]` `m >= 0`, and @@ -849,36 +851,35 @@ Initialize a batch of Binomial distributions. `[N1,..., Nm]` `m >= 0`, `probs in [0, 1]`. Each entry represents the probability of success for independent Binomial distributions. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Binomial.allow_nan_stats` {#Binomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -1076,7 +1077,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -1093,7 +1094,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -1209,7 +1210,7 @@ Mode. Additional documentation from `Binomial`: Note that when `(1 + total_count) * probs` is an integer, there are -actually two modes. Namely, `(1 + total_count) * probs` and +actually two modes. Namely, `(1 + total_count) * probs` and `(1 + total_count) * probs - 1` are both modes. Here we return only the larger of the two modes. @@ -1253,8 +1254,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -1422,7 +1423,7 @@ Number of trials. #### `tf.contrib.distributions.Binomial.validate_args` {#Binomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -1479,15 +1480,15 @@ Construct Bernoulli distributions. Bernoulli distribution. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples. Default: `int32`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -1499,21 +1500,20 @@ Construct Bernoulli distributions. #### `tf.contrib.distributions.Bernoulli.allow_nan_stats` {#Bernoulli.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -1711,7 +1711,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -1728,7 +1728,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -1872,8 +1872,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -2021,7 +2021,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Bernoulli.validate_args` {#Bernoulli.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -2068,21 +2068,20 @@ Bernoulli with `probs = nn.sigmoid(logits)`. #### `tf.contrib.distributions.BernoulliWithSigmoidProbs.allow_nan_stats` {#BernoulliWithSigmoidProbs.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -2280,7 +2279,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -2297,7 +2296,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -2441,8 +2440,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -2590,7 +2589,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.BernoulliWithSigmoidProbs.validate_args` {#BernoulliWithSigmoidProbs.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -2714,36 +2713,35 @@ Initialize a batch of Beta distributions. * <b>`concentration0`</b>: Positive floating-point `Tensor` indicating mean number of failures; aka "beta". Otherwise has same semantics as `concentration1`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Beta.allow_nan_stats` {#Beta.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -2961,7 +2959,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -2978,7 +2976,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -3087,7 +3085,7 @@ Additional documentation from `Beta`: Note: The mode is undefined when `concentration1 <= 1` or `concentration0 <= 1`. If `self.allow_nan_stats` is `True`, `NaN` -is used for undefined modes. If `self.allow_nan_stats` is `False` an +is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -3130,8 +3128,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -3285,7 +3283,7 @@ Sum of concentration parameters. #### `tf.contrib.distributions.Beta.validate_args` {#Beta.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -3332,21 +3330,20 @@ Beta with softplus transform of `concentration1` and `concentration0`. #### `tf.contrib.distributions.BetaWithSoftplusConcentration.allow_nan_stats` {#BetaWithSoftplusConcentration.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -3564,7 +3561,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -3581,7 +3578,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -3690,7 +3687,7 @@ Additional documentation from `Beta`: Note: The mode is undefined when `concentration1 <= 1` or `concentration0 <= 1`. If `self.allow_nan_stats` is `True`, `NaN` -is used for undefined modes. If `self.allow_nan_stats` is `False` an +is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -3733,8 +3730,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -3888,7 +3885,7 @@ Sum of concentration parameters. #### `tf.contrib.distributions.BetaWithSoftplusConcentration.validate_args` {#BetaWithSoftplusConcentration.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -3983,36 +3980,35 @@ Initialize Categorical distributions using class log-probabilities. represents a vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Categorical.allow_nan_stats` {#Categorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -4217,7 +4213,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -4234,7 +4230,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -4374,8 +4370,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -4523,7 +4519,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Categorical.validate_args` {#Categorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -4594,37 +4590,36 @@ Construct Chi2 distributions with parameter `df`. * <b>`df`</b>: Floating point tensor, the degrees of freedom of the - distribution(s). `df` must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution + distribution(s). `df` must contain only positive values. +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Chi2.allow_nan_stats` {#Chi2.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -4836,7 +4831,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -4853,7 +4848,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -4949,7 +4944,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -4992,8 +4987,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -5141,7 +5136,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Chi2.validate_args` {#Chi2.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -5188,21 +5183,20 @@ Chi2 with parameter transform `df = floor(abs(df))`. #### `tf.contrib.distributions.Chi2WithAbsDf.allow_nan_stats` {#Chi2WithAbsDf.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -5414,7 +5408,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -5431,7 +5425,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -5527,7 +5521,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -5570,8 +5564,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -5719,7 +5713,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Chi2WithAbsDf.validate_args` {#Chi2WithAbsDf.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -5794,36 +5788,35 @@ Construct Exponential distribution with parameter `rate`. * <b>`rate`</b>: Floating point tensor, equivalent to `1 / mean`. Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Exponential.allow_nan_stats` {#Exponential.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -6028,7 +6021,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -6045,7 +6038,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -6141,7 +6134,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -6184,8 +6177,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -6333,7 +6326,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Exponential.validate_args` {#Exponential.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -6380,21 +6373,20 @@ Exponential with softplus transform on `rate`. #### `tf.contrib.distributions.ExponentialWithSoftplusRate.allow_nan_stats` {#ExponentialWithSoftplusRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -6599,7 +6591,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -6616,7 +6608,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -6712,7 +6704,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -6755,8 +6747,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -6904,7 +6896,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.ExponentialWithSoftplusRate.validate_args` {#ExponentialWithSoftplusRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -7005,15 +6997,15 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). distribution(s). Must contain only positive values. * <b>`rate`</b>: Floating point tensor, the inverse scale params of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -7025,21 +7017,20 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). #### `tf.contrib.distributions.Gamma.allow_nan_stats` {#Gamma.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -7244,7 +7235,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -7261,7 +7252,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -7357,7 +7348,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -7400,8 +7391,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -7549,7 +7540,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Gamma.validate_args` {#Gamma.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -7596,21 +7587,20 @@ denotes expectation, and `Var.shape = batch_shape + event_shape`. #### `tf.contrib.distributions.GammaWithSoftplusConcentrationRate.allow_nan_stats` {#GammaWithSoftplusConcentrationRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -7815,7 +7805,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -7832,7 +7822,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -7928,7 +7918,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -7971,8 +7961,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -8120,7 +8110,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.GammaWithSoftplusConcentrationRate.validate_args` {#GammaWithSoftplusConcentrationRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -8221,15 +8211,15 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). distribution(s). Must contain only positive values. * <b>`rate`</b>: Floating point tensor, the inverse scale params of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -8242,21 +8232,20 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). #### `tf.contrib.distributions.InverseGamma.allow_nan_stats` {#InverseGamma.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -8461,7 +8450,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -8478,7 +8467,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -8568,7 +8557,7 @@ Additional documentation from `InverseGamma`: The mean of an inverse gamma distribution is `rate / (concentration - 1)`, when `concentration > 1`, and `NaN` -otherwise. If `self.allow_nan_stats` is `False`, an exception will be +otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN` @@ -8623,8 +8612,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -8772,7 +8761,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.InverseGamma.validate_args` {#InverseGamma.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -8826,21 +8815,20 @@ than returning `NaN`. #### `tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.allow_nan_stats` {#InverseGammaWithSoftplusConcentrationRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -9045,7 +9033,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -9062,7 +9050,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -9152,7 +9140,7 @@ Additional documentation from `InverseGamma`: The mean of an inverse gamma distribution is `rate / (concentration - 1)`, when `concentration > 1`, and `NaN` -otherwise. If `self.allow_nan_stats` is `False`, an exception will be +otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN` @@ -9207,8 +9195,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -9356,7 +9344,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.validate_args` {#InverseGammaWithSoftplusConcentrationRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -9438,15 +9426,15 @@ broadcasting (e.g., `loc / scale` is a valid operation). of the distribution. * <b>`scale`</b>: Positive floating point tensor which characterizes the spread of the distribution. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -9458,21 +9446,20 @@ broadcasting (e.g., `loc / scale` is a valid operation). #### `tf.contrib.distributions.Laplace.allow_nan_stats` {#Laplace.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -9670,7 +9657,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -9687,7 +9674,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -9827,8 +9814,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -9976,7 +9963,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Laplace.validate_args` {#Laplace.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -10023,21 +10010,20 @@ Laplace with softplus applied to `scale`. #### `tf.contrib.distributions.LaplaceWithSoftplusScale.allow_nan_stats` {#LaplaceWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -10235,7 +10221,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -10252,7 +10238,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -10392,8 +10378,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -10541,7 +10527,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.LaplaceWithSoftplusScale.validate_args` {#LaplaceWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -10646,13 +10632,13 @@ broadcasting (e.g. `loc + scale` is a valid operation). * <b>`loc`</b>: Floating point tensor, the means of the distribution(s). * <b>`scale`</b>: Floating point tensor, the scales of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. * <b>`name`</b>: The name to give Ops created by the initializer. @@ -10666,21 +10652,20 @@ broadcasting (e.g. `loc + scale` is a valid operation). #### `tf.contrib.distributions.Logistic.allow_nan_stats` {#Logistic.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -10878,7 +10863,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -10895,7 +10880,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -11035,8 +11020,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -11184,7 +11169,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Logistic.validate_args` {#Logistic.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -11291,15 +11276,15 @@ broadcasting (e.g. `loc + scale` is a valid operation). * <b>`loc`</b>: Floating point tensor; the means of the distribution(s). * <b>`scale`</b>: Floating point tensor; the stddevs of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -11311,21 +11296,20 @@ broadcasting (e.g. `loc + scale` is a valid operation). #### `tf.contrib.distributions.Normal.allow_nan_stats` {#Normal.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -11523,7 +11507,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -11540,7 +11524,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -11680,8 +11664,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -11829,7 +11813,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Normal.validate_args` {#Normal.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -11876,21 +11860,20 @@ Normal with softplus applied to `scale`. #### `tf.contrib.distributions.NormalWithSoftplusScale.allow_nan_stats` {#NormalWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -12088,7 +12071,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -12105,7 +12088,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -12245,8 +12228,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -12394,7 +12377,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.NormalWithSoftplusScale.validate_args` {#NormalWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -12454,36 +12437,35 @@ Initialize a batch of Poisson distributions. * <b>`rate`</b>: Floating point tensor, the rate parameter of the distribution(s). `rate` must be positive. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Poisson.allow_nan_stats` {#Poisson.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -12688,7 +12670,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -12705,7 +12687,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -12857,8 +12839,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -13013,7 +12995,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Poisson.validate_args` {#Poisson.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -13078,7 +13060,7 @@ Y = loc + scale * X ``` Notice that `scale` has semantics more similar to standard deviation than -variance. However it is not actually the std. deviation; the Student's +variance. However it is not actually the std. deviation; the Student's t-distribution std. dev. is `scale sqrt(df / (df - 2))` when `df > 2`. #### Examples @@ -13133,22 +13115,22 @@ supports broadcasting (e.g. `df + loc + scale` is a valid operation). ##### Args: -* <b>`df`</b>: Numeric `Tensor`. The degrees of freedom of the distribution(s). - `df` must contain only positive values. -* <b>`loc`</b>: Numeric `Tensor`. The mean(s) of the distribution(s). -* <b>`scale`</b>: Numeric `Tensor`. The scaling factor(s) for the distribution(s). - Note that `scale` is not technically the standard deviation of this - distribution but has semantics more similar to standard deviation than - variance. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`df`</b>: Floating-point `Tensor`. The degrees of freedom of the + distribution(s). `df` must contain only positive values. +* <b>`loc`</b>: Floating-point `Tensor`. The mean(s) of the distribution(s). +* <b>`scale`</b>: Floating-point `Tensor`. The scaling factor(s) for the + distribution(s). Note that `scale` is not technically the standard + deviation of this distribution but has semantics more similar to + standard deviation than variance. +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -13160,21 +13142,20 @@ supports broadcasting (e.g. `df + loc + scale` is a valid operation). #### `tf.contrib.distributions.StudentT.allow_nan_stats` {#StudentT.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -13379,7 +13360,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -13396,7 +13377,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -13492,7 +13473,7 @@ Mean. Additional documentation from `StudentT`: The mean of Student's T equals `loc` if `df > 1`, otherwise it is -`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised +`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised rather than returning `NaN`. @@ -13542,8 +13523,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -13691,7 +13672,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.StudentT.validate_args` {#StudentT.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -13749,21 +13730,20 @@ StudentT with `df = floor(abs(df))` and `scale = softplus(scale)`. #### `tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.allow_nan_stats` {#StudentTWithAbsDfSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -13968,7 +13948,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -13985,7 +13965,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -14081,7 +14061,7 @@ Mean. Additional documentation from `StudentT`: The mean of Student's T equals `loc` if `df > 1`, otherwise it is -`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised +`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised rather than returning `NaN`. @@ -14131,8 +14111,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -14280,7 +14260,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.validate_args` {#StudentTWithAbsDfSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -14377,15 +14357,15 @@ Initialize a batch of Uniform distributions. have `low < high`. * <b>`high`</b>: Floating point tensor, upper boundary of the output interval. Must have `low < high`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -14397,21 +14377,20 @@ Initialize a batch of Uniform distributions. #### `tf.contrib.distributions.Uniform.allow_nan_stats` {#Uniform.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -14616,7 +14595,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -14633,7 +14612,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -14773,8 +14752,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -14922,7 +14901,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Uniform.validate_args` {#Uniform.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -15106,15 +15085,15 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then `k x k` identity matrices added to `scale`. When both `scale_identity_multiplier` and `scale_diag` are `None` then `scale` is the `Identity`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -15126,21 +15105,20 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then #### `tf.contrib.distributions.MultivariateNormalDiag.allow_nan_stats` {#MultivariateNormalDiag.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -15359,7 +15337,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -15376,7 +15354,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -15539,8 +15517,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -15704,7 +15682,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiag.validate_args` {#MultivariateNormalDiag.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -15862,15 +15840,15 @@ Additional leading dimensions (if any) will index batches. * <b>`scale_tril`</b>: Floating-point, lower-triangular `Tensor` with non-zero diagonal elements. `scale_tril` has shape `[B1, ..., Bb, k, k]` where `b >= 0` and `k` is the event size. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -15882,21 +15860,20 @@ Additional leading dimensions (if any) will index batches. #### `tf.contrib.distributions.MultivariateNormalTriL.allow_nan_stats` {#MultivariateNormalTriL.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -16115,7 +16092,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -16132,7 +16109,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -16295,8 +16272,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -16460,7 +16437,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalTriL.validate_args` {#MultivariateNormalTriL.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -16657,18 +16634,18 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then * <b>`scale_perturb_diag`</b>: Floating-point `Tensor` representing a diagonal matrix inside the rank-`r` perturbation added to `scale`. May have shape `[B1, ..., Bb, r]`, `b >= 0`, and characterizes `b`-batches of `r x r` - diagonal matrices inside the perturbation added to `scale`. When + diagonal matrices inside the perturbation added to `scale`. When `None`, an identity matrix is used inside the perturbation. Can only be specified if `scale_perturb_factor` is also specified. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -16680,21 +16657,20 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then #### `tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.allow_nan_stats` {#MultivariateNormalDiagPlusLowRank.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -16913,7 +16889,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -16930,7 +16906,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -17093,8 +17069,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -17258,7 +17234,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.validate_args` {#MultivariateNormalDiagPlusLowRank.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -17305,21 +17281,20 @@ MultivariateNormalDiag with `diag_stddev = softplus(diag_stddev)`. #### `tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.allow_nan_stats` {#MultivariateNormalDiagWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -17538,7 +17513,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -17555,7 +17530,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -17718,8 +17693,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -17883,7 +17858,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.validate_args` {#MultivariateNormalDiagWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -18018,36 +17993,35 @@ Initialize a batch of Dirichlet distributions. `concentration.shape = [N1, N2, ..., Nm, k]` then `batch_shape = [N1, N2, ..., Nm]` and `event_shape = [k]`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Dirichlet.allow_nan_stats` {#Dirichlet.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -18252,7 +18226,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -18269,7 +18243,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -18373,7 +18347,7 @@ Mode. Additional documentation from `Dirichlet`: Note: The mode is undefined when any `concentration <= 1`. If -`self.allow_nan_stats` is `True`, `NaN` is used for undefined modes. If +`self.allow_nan_stats` is `True`, `NaN` is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -18417,8 +18391,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -18574,7 +18548,7 @@ Sum of last dim of concentration parameter. #### `tf.contrib.distributions.Dirichlet.validate_args` {#Dirichlet.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -18707,43 +18681,42 @@ Initialize a batch of DirichletMultinomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor, whose dtype is the same as `concentration`. The shape is broadcastable to `[N1,..., Nm]` with - `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different + `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different Dirichlet multinomial distributions. Its components should be equal to integer values. * <b>`concentration`</b>: Positive floating point tensor, whose dtype is the same as `n` with shape broadcastable to `[N1,..., Nm, k]` `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different `k` class Dirichlet multinomial distributions. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.DirichletMultinomial.allow_nan_stats` {#DirichletMultinomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -18968,7 +18941,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -18985,7 +18958,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -19027,10 +19000,10 @@ Log probability density/mass function (depending on `is_continuous`). Additional documentation from `DirichletMultinomial`: For each batch of counts, -`value = [n_0, ... ,n_{k-1}]`, `P[value]` is the probability that after sampling -`self.total_count` draws from this Dirichlet-Multinomial distribution, the -number of draws falling in class `j` is `n_j`. Since this definition is -[exchangeable]( https://en.wikipedia.org/wiki/Exchangeable_random_variables); +`value = [n_0, ..., n_{k-1}]`, `P[value]` is the probability that after +sampling `self.total_count` draws from this Dirichlet-Multinomial distribution, +the number of draws falling in class `j` is `n_j`. Since this definition is +[exchangeable](https://en.wikipedia.org/wiki/Exchangeable_random_variables); different sequences have the same counts so the probability includes a combinatorial coefficient. @@ -19134,8 +19107,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -19173,10 +19146,10 @@ Probability density/mass function (depending on `is_continuous`). Additional documentation from `DirichletMultinomial`: For each batch of counts, -`value = [n_0, ... ,n_{k-1}]`, `P[value]` is the probability that after sampling -`self.total_count` draws from this Dirichlet-Multinomial distribution, the -number of draws falling in class `j` is `n_j`. Since this definition is -[exchangeable]( https://en.wikipedia.org/wiki/Exchangeable_random_variables); +`value = [n_0, ..., n_{k-1}]`, `P[value]` is the probability that after +sampling `self.total_count` draws from this Dirichlet-Multinomial distribution, +the number of draws falling in class `j` is `n_j`. Since this definition is +[exchangeable](https://en.wikipedia.org/wiki/Exchangeable_random_variables); different sequences have the same counts so the probability includes a combinatorial coefficient. @@ -19306,7 +19279,7 @@ Number of trials used to construct a sample. #### `tf.contrib.distributions.DirichletMultinomial.validate_args` {#DirichletMultinomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -19425,7 +19398,7 @@ Initialize a batch of Multinomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor with shape broadcastable to `[N1,..., Nm]` with `m >= 0`. Defines this as a batch of - `N1 x ... x Nm` different Multinomial distributions. Its components + `N1 x ... x Nm` different Multinomial distributions. Its components should be equal to integer values. * <b>`logits`</b>: Floating point tensor representing the log-odds of a positive event with shape broadcastable to `[N1,..., Nm, k], m >= 0`, @@ -19433,40 +19406,39 @@ Initialize a batch of Multinomial distributions. `N1 x ... x Nm` different `k` class Multinomial distributions. Only one of `logits` or `probs` should be passed in. * <b>`probs`</b>: Positive floating point tensor with shape broadcastable to - `[N1,..., Nm, k]` `m >= 0` and same dtype as `total_count`. Defines + `[N1,..., Nm, k]` `m >= 0` and same dtype as `total_count`. Defines this as a batch of `N1 x ... x Nm` different `k` class Multinomial distributions. `probs`'s components in the last portion of its shape should sum to `1`. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Multinomial.allow_nan_stats` {#Multinomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -19664,7 +19636,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -19681,7 +19653,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -19837,8 +19809,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -20009,7 +19981,7 @@ Number of trials used to construct a sample. #### `tf.contrib.distributions.Multinomial.validate_args` {#Multinomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -20082,7 +20054,7 @@ chol_scale = tf.cholesky(...) # Shape is [3, 3]. dist = tf.contrib.distributions.WishartCholesky(df=df, scale=chol_scale) # Evaluate this on an observation in R^3, returning a scalar. -x = ... # A 3x3 positive definite matrix. +x = ... # A 3x3 positive definite matrix. dist.prob(x) # Shape is [], a scalar. # Evaluate this on a two observations, each in R^{3x3}, returning a length two @@ -20115,41 +20087,40 @@ Construct Wishart distributions. or equal to dimension of the scale matrix. * <b>`scale`</b>: `float` or `double` `Tensor`. The Cholesky factorization of the symmetric positive definite scale matrix of the distribution. -* <b>`cholesky_input_output_matrices`</b>: `Boolean`. Any function which whose input - or output is a matrix assumes the input is Cholesky and returns a +* <b>`cholesky_input_output_matrices`</b>: Python `bool`. Any function which whose + input or output is a matrix assumes the input is Cholesky and returns a Cholesky factored matrix. Example `log_prob` input takes a Cholesky and `sample_n` returns a Cholesky when `cholesky_input_output_matrices=True`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.WishartCholesky.allow_nan_stats` {#WishartCholesky.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -20368,7 +20339,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -20385,7 +20356,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -20532,8 +20503,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -20688,7 +20659,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.WishartCholesky.validate_args` {#WishartCholesky.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -20757,7 +20728,7 @@ scale = ... # Shape is [3, 3]; positive definite. dist = tf.contrib.distributions.WishartFull(df=df, scale=scale) # Evaluate this on an observation in R^3, returning a scalar. -x = ... # A 3x3 positive definite matrix. +x = ... # A 3x3 positive definite matrix. dist.prob(x) # Shape is [], a scalar. # Evaluate this on a two observations, each in R^{3x3}, returning a length two @@ -20790,41 +20761,40 @@ Construct Wishart distributions. or equal to dimension of the scale matrix. * <b>`scale`</b>: `float` or `double` `Tensor`. The symmetric positive definite scale matrix of the distribution. -* <b>`cholesky_input_output_matrices`</b>: `Boolean`. Any function which whose input - or output is a matrix assumes the input is Cholesky and returns a +* <b>`cholesky_input_output_matrices`</b>: Python `bool`. Any function which whose + input or output is a matrix assumes the input is Cholesky and returns a Cholesky factored matrix. Example `log_prob` input takes a Cholesky and `sample_n` returns a Cholesky when `cholesky_input_output_matrices=True`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.WishartFull.allow_nan_stats` {#WishartFull.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -21043,7 +21013,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -21060,7 +21030,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -21207,8 +21177,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -21363,7 +21333,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.WishartFull.validate_args` {#WishartFull.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -21410,7 +21380,7 @@ Create a trainable covariance defined by a Cholesky factor: matrix_values = tf.contrib.layers.fully_connected(activations, 4) matrix = tf.reshape(matrix_values, (batch_size, 2, 2)) -# Make the diagonal positive. If the upper triangle was zero, this would be a +# Make the diagonal positive. If the upper triangle was zero, this would be a # valid Cholesky factor. chol = matrix_diag_transform(matrix, transform=tf.nn.softplus) @@ -21432,7 +21402,7 @@ mu = tf.contrib.layers.fully_connected(activations, 2) # This is a fully trainable multivariate normal! dist = tf.contrib.distributions.MVNCholesky(mu, chol) -# Standard log loss. Minimizing this will "train" mu and chol, and then dist +# Standard log loss. Minimizing this will "train" mu and chol, and then dist # will be a distribution predicting labels as multivariate Gaussians. loss = -1 * tf.reduce_mean(dist.log_prob(labels)) ``` @@ -21442,9 +21412,9 @@ loss = -1 * tf.reduce_mean(dist.log_prob(labels)) * <b>`matrix`</b>: Rank `R` `Tensor`, `R >= 2`, where the last two dimensions are equal. -* <b>`transform`</b>: Element-wise function mapping `Tensors` to `Tensors`. To - be applied to the diagonal of `matrix`. If `None`, `matrix` is returned - unchanged. Defaults to `None`. +* <b>`transform`</b>: Element-wise function mapping `Tensors` to `Tensors`. To + be applied to the diagonal of `matrix`. If `None`, `matrix` is returned + unchanged. Defaults to `None`. * <b>`name`</b>: A name to give created ops. Defaults to "matrix_diag_transform". @@ -21479,7 +21449,7 @@ We now describe how a `TransformedDistribution` alters the input/outputs of a Write `cdf(Y=y)` for an absolutely continuous cumulative distribution function of random variable `Y`; write the probability density function `pdf(Y=y) := d^k / (dy_1,...,dy_k) cdf(Y=y)` for its derivative wrt to `Y` evaluated at -`y`. Assume that `Y = g(X)` where `g` is a deterministic diffeomorphism, +`y`. Assume that `Y = g(X)` where `g` is a deterministic diffeomorphism, i.e., a non-random, continuous, differentiable, and invertible function. Write the inverse of `g` as `X = g^{-1}(Y)` and `(J o g)(x)` for the Jacobian of `g` evaluated at `x`. @@ -21554,7 +21524,7 @@ log_normal = ds.TransformedDistribution( forward_fn=tf.exp, inverse_fn=tf.log, inverse_log_det_jacobian_fn=( - lambda y: -tf.reduce_sum(tf.log(y), reduction_indices=-1)), + lambda y: -tf.reduce_sum(tf.log(y), axis=-1)), name="LogNormalTransformedDistribution") ``` @@ -21570,7 +21540,7 @@ normal = ds.TransformedDistribution( A `TransformedDistribution`'s batch- and event-shape are implied by the base distribution unless explicitly overridden by `batch_shape` or `event_shape` -arguments. Specifying an overriding `batch_shape` (`event_shape`) is +arguments. Specifying an overriding `batch_shape` (`event_shape`) is permitted only if the base distribution has scalar batch-shape (event-shape). The bijector is applied to the distribution as if the distribution possessed the overridden shape(s). The following example demonstrates how to construct a @@ -21611,11 +21581,11 @@ Construct a Transformed Distribution. `batch_shape`; valid only if `distribution.is_scalar_batch()`. * <b>`event_shape`</b>: `integer` vector `Tensor` which overrides `distribution` `event_shape`; valid only if `distribution.is_scalar_event()`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. Default: +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: `bijector.name + distribution.name`. @@ -21623,21 +21593,20 @@ Construct a Transformed Distribution. #### `tf.contrib.distributions.TransformedDistribution.allow_nan_stats` {#TransformedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -21849,7 +21818,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -21866,7 +21835,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -21999,8 +21968,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -22141,7 +22110,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.TransformedDistribution.validate_args` {#TransformedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -22242,19 +22211,19 @@ the `distribution`. * <b>`distribution`</b>: The base distribution class to transform. Typically an instance of `Distribution`. * <b>`low`</b>: `Tensor` with same `dtype` as this distribution and shape - able to be added to samples. Should be a whole number. Default `None`. + able to be added to samples. Should be a whole number. Default `None`. If provided, base distribution's `prob` should be defined at `low`. * <b>`high`</b>: `Tensor` with same `dtype` as this distribution and shape - able to be added to samples. Should be a whole number. Default `None`. + able to be added to samples. Should be a whole number. Default `None`. If provided, base distribution's `prob` should be defined at `high - 1`. `high` must be strictly greater than `low`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -22268,21 +22237,20 @@ the `distribution`. #### `tf.contrib.distributions.QuantizedDistribution.allow_nan_stats` {#QuantizedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -22505,7 +22473,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -22522,7 +22490,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -22591,7 +22559,7 @@ P[Y = y] := P[X <= low], if y == low, ``` -The base distribution's `log_cdf` method must be defined on `y - 1`. If the +The base distribution's `log_cdf` method must be defined on `y - 1`. If the base distribution has a `log_survival_function` method results will be more accurate for large values of `y`, and in this case the `log_survival_function` must also be defined on `y - 1`. @@ -22709,8 +22677,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -22757,7 +22725,7 @@ P[Y = y] := P[X <= low], if y == low, ``` -The base distribution's `cdf` method must be defined on `y - 1`. If the +The base distribution's `cdf` method must be defined on `y - 1`. If the base distribution has a `survival_function` method, results will be more accurate for large values of `y`, and in this case the `survival_function` must also be defined on `y - 1`. @@ -22887,7 +22855,7 @@ The base distribution's `cdf` method must be defined on `y - 1`. #### `tf.contrib.distributions.QuantizedDistribution.validate_args` {#QuantizedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -22955,13 +22923,13 @@ time and match `len(components)`. * <b>`components`</b>: A list or tuple of `Distribution` instances. Each instance must have the same type, be defined on the same domain, and have matching `event_shape` and `batch_shape`. -* <b>`validate_args`</b>: `Boolean`, default `False`. If `True`, raise a runtime +* <b>`validate_args`</b>: Python `bool`, default `False`. If `True`, raise a runtime error if batch or event ranks are inconsistent between cat and any of - the distributions. This is only checked if the ranks cannot be + the distributions. This is only checked if the ranks cannot be determined statically at graph construction time. -* <b>`allow_nan_stats`</b>: Boolean, default `True`. If `False`, raise an +* <b>`allow_nan_stats`</b>: Boolean, default `True`. If `False`, raise an exception if a statistic (e.g. mean/mode/etc...) is undefined for any - batch member. If `True`, batch members with valid parameters leading to + batch member. If `True`, batch members with valid parameters leading to undefined statistics will return NaN for this statistic. * <b>`name`</b>: A name for this distribution (optional). @@ -22984,21 +22952,20 @@ time and match `len(components)`. #### `tf.contrib.distributions.Mixture.allow_nan_stats` {#Mixture.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -23175,7 +23142,7 @@ distribution: \\) where \\( p \\) is the prior distribution, \\( q \\) is the variational, -and \\( H[q] \\) is the entropy of \\( q \\). If there is a lower bound +and \\( H[q] \\) is the entropy of \\( q \\). If there is a lower bound \\( G[q] \\) such that \\( H[q] \geq G[q] \\) then it can be used in place of \\( H[q] \\). @@ -23256,7 +23223,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -23273,7 +23240,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -23413,8 +23380,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -23555,7 +23522,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Mixture.validate_args` {#Mixture.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -23602,7 +23569,7 @@ Posterior Normal distribution with conjugate prior on the mean. This model assumes that `n` observations (with sum `s`) come from a Normal with unknown mean `loc` (described by the Normal `prior`) -and known variance `scale^2`. The "known scale posterior" is +and known variance `scale**2`. The "known scale posterior" is the distribution of the unknown `loc`. Accepts a prior Normal distribution object, having parameters @@ -23612,12 +23579,12 @@ and statistical estimates `s` (the sum(s) of the observations) and `n` (the number(s) of observations). Returns a posterior (also Normal) distribution object, with parameters -`(loc', scale'^2)`, where: +`(loc', scale'**2)`, where: ``` -mu ~ N(mu', sigma'^2) -sigma'^2 = 1/(1/sigma0^2 + n/sigma^2), -mu' = (mu0/sigma0^2 + s/sigma^2) * sigma'^2. +mu ~ N(mu', sigma'**2) +sigma'**2 = 1/(1/sigma0**2 + n/sigma**2), +mu' = (mu0/sigma0**2 + s/sigma**2) * sigma'**2. ``` Distribution parameters from `prior`, as well as `scale`, `s`, and `n`. @@ -23630,8 +23597,8 @@ will broadcast in the case of multidimensional sets of parameters. the prior distribution having parameters `(loc0, scale0)`. * <b>`scale`</b>: tensor of type `dtype`, taking values `scale > 0`. The known stddev parameter(s). -* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. -* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. +* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. +* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. ##### Returns: @@ -23653,7 +23620,7 @@ Posterior predictive Normal distribution w. conjugate prior on the mean. This model assumes that `n` observations (with sum `s`) come from a Normal with unknown mean `loc` (described by the Normal `prior`) -and known variance `scale^2`. The "known scale predictive" +and known variance `scale**2`. The "known scale predictive" is the distribution of new observations, conditioned on the existing observations and our prior. @@ -23663,20 +23630,20 @@ distribution(s) (also assumed Normal), and statistical estimates `s` (the sum(s) of the observations) and `n` (the number(s) of observations). -Calculates the Normal distribution(s) `p(x | sigma^2)`: +Calculates the Normal distribution(s) `p(x | sigma**2)`: ``` -p(x | sigma^2) = int N(x | mu, sigma^2) N(mu | prior.loc, prior.scale**2) dmu - = N(x | prior.loc, 1/(sigma^2 + prior.scale**2)) +p(x | sigma**2) = int N(x | mu, sigma**2)N(mu | prior.loc, prior.scale**2) dmu + = N(x | prior.loc, 1 / (sigma**2 + prior.scale**2)) ``` Returns the predictive posterior distribution object, with parameters -`(loc', scale'^2)`, where: +`(loc', scale'**2)`, where: ``` -sigma_n^2 = 1/(1/sigma0^2 + n/sigma^2), -mu' = (mu0/sigma0^2 + s/sigma^2) * sigma_n^2. -sigma'^2 = sigma_n^2 + sigma^2, +sigma_n**2 = 1/(1/sigma0**2 + n/sigma**2), +mu' = (mu0/sigma0**2 + s/sigma**2) * sigma_n**2. +sigma'**2 = sigma_n**2 + sigma**2, ``` Distribution parameters from `prior`, as well as `scale`, `s`, and `n`. @@ -23689,8 +23656,8 @@ will broadcast in the case of multidimensional sets of parameters. the prior distribution having parameters `(loc0, scale0)`. * <b>`scale`</b>: tensor of type `dtype`, taking values `scale > 0`. The known stddev parameter(s). -* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. -* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. +* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. +* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. ##### Returns: @@ -23732,7 +23699,7 @@ identified in the search is used (favoring a shorter MRO distance to * <b>`dist_b`</b>: The second distribution. * <b>`allow_nan`</b>: If `False` (default), a runtime error is raised if the KL returns NaN values for any batch entry of the given - distributions. If `True`, the KL may return a NaN for the given entry. + distributions. If `True`, the KL may return a NaN for the given entry. * <b>`name`</b>: (optional) Name scope to use for created operations. ##### Returns: @@ -23939,36 +23906,35 @@ Initialize ExpRelaxedOneHotCategorical using class log-probabilities. the last dimension represents a vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.ExpRelaxedOneHotCategorical.allow_nan_stats` {#ExpRelaxedOneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -24173,7 +24139,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -24190,7 +24156,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -24330,8 +24296,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -24486,7 +24452,7 @@ Batchwise temperature tensor of a RelaxedCategorical. #### `tf.contrib.distributions.ExpRelaxedOneHotCategorical.validate_args` {#ExpRelaxedOneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -24532,11 +24498,11 @@ Categorical has event_dim=() while OneHotCategorical has event_dim=K, where K is the number of classes. This class provides methods to create indexed batches of OneHotCategorical -distributions. If the provided `logits` or `probs` is rank 2 or higher, for +distributions. If the provided `logits` or `probs` is rank 2 or higher, for every fixed set of leading dimensions, the last dimension represents one -single OneHotCategorical distribution. When calling distribution +single OneHotCategorical distribution. When calling distribution functions (e.g. `dist.prob(x)`), `logits` and `x` are broadcast to the -same shape (if possible). In all cases, the last dimension of `logits,x` +same shape (if possible). In all cases, the last dimension of `logits,x` represents single OneHotCategorical distributions. #### Examples @@ -24589,36 +24555,35 @@ Initialize OneHotCategorical distributions using class log-probabilities. vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.OneHotCategorical.allow_nan_stats` {#OneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -24823,7 +24788,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -24840,7 +24805,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -24980,8 +24945,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -25129,7 +25094,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.OneHotCategorical.validate_args` {#OneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -25282,15 +25247,15 @@ Construct RelaxedBernoulli distributions. * <b>`probs`</b>: An N-D `Tensor` representing the probability of a positive event. Each entry in the `Tensor` parameterizes an independent Bernoulli distribution. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -25302,21 +25267,20 @@ Construct RelaxedBernoulli distributions. #### `tf.contrib.distributions.RelaxedBernoulli.allow_nan_stats` {#RelaxedBernoulli.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -25528,7 +25492,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -25545,7 +25509,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -25685,8 +25649,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -25841,7 +25805,7 @@ Distribution parameter for the location. #### `tf.contrib.distributions.RelaxedBernoulli.validate_args` {#RelaxedBernoulli.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -25968,9 +25932,9 @@ Initialize RelaxedOneHotCategorical using class log-probabilities. of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). * <b>`validate_args`</b>: Unused in this distribution. -* <b>`allow_nan_stats`</b>: `Boolean`, default `True`. If `False`, raise an +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. If `False`, raise an exception if a statistic (e.g. mean/mode/etc...) is undefined for any - batch member. If `True`, batch members with valid parameters leading to + batch member. If `True`, batch members with valid parameters leading to undefined statistics will return NaN for this statistic. * <b>`name`</b>: A name for this distribution (optional). @@ -25979,21 +25943,20 @@ Initialize RelaxedOneHotCategorical using class log-probabilities. #### `tf.contrib.distributions.RelaxedOneHotCategorical.allow_nan_stats` {#RelaxedOneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -26205,7 +26168,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -26222,7 +26185,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -26355,8 +26318,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -26497,7 +26460,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.RelaxedOneHotCategorical.validate_args` {#RelaxedOneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -26550,25 +26513,28 @@ Constructs the `Distribution`. * <b>`dtype`</b>: The type of the event samples. `None` implies no type-enforcement. -* <b>`is_continuous`</b>: Python boolean. If `True` this - `Distribution` is continuous over its supported domain. +* <b>`is_continuous`</b>: Python `bool`. If `True` this `Distribution` is continuous + over its supported domain. * <b>`reparameterization_type`</b>: Instance of `ReparameterizationType`. If `distributions.FULLY_REPARAMETERIZED`, this `Distribution` can be reparameterized in terms of some standard distribution with a function whose Jacobian is constant for the support - of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, + of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, then no such reparameterization is available. -* <b>`validate_args`</b>: Python boolean. Whether to validate input with asserts. - If `validate_args` is `False`, and the inputs are invalid, - correct behavior is not guaranteed. -* <b>`allow_nan_stats`</b>: Python boolean. If `False`, raise an - exception if a statistic (e.g., mean, mode) is undefined for any batch - member. If True, batch members with valid parameters leading to - undefined statistics will return `NaN` for this statistic. -* <b>`parameters`</b>: Python dictionary of parameters used to instantiate this +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution + parameters are checked for validity despite possibly degrading runtime + performance. When `False` invalid inputs may silently render incorrect + outputs. +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics + (e.g., mean, mode, variance) use the value "`NaN`" to indicate the + result is undefined. When `False`, an exception is raised if one or + more of the statistic's batch members are undefined. +* <b>`parameters`</b>: Python `dict` of parameters used to instantiate this + `Distribution`. +* <b>`graph_parents`</b>: Python `list` of graph prerequisites of this `Distribution`. -* <b>`graph_parents`</b>: Python list of graph prerequisites of this `Distribution`. -* <b>`name`</b>: A name for this distribution. Default: subclass name. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: + subclass name. ##### Raises: @@ -26580,21 +26546,20 @@ Constructs the `Distribution`. #### `tf.contrib.distributions.ConditionalDistribution.allow_nan_stats` {#ConditionalDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -26776,7 +26741,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -26793,7 +26758,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -26876,8 +26841,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -26978,7 +26943,7 @@ denotes expectation, and `stddev.shape = batch_shape + event_shape`. #### `tf.contrib.distributions.ConditionalDistribution.validate_args` {#ConditionalDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - @@ -27031,11 +26996,11 @@ Construct a Transformed Distribution. `batch_shape`; valid only if `distribution.is_scalar_batch()`. * <b>`event_shape`</b>: `integer` vector `Tensor` which overrides `distribution` `event_shape`; valid only if `distribution.is_scalar_event()`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. Default: +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: `bijector.name + distribution.name`. @@ -27043,21 +27008,20 @@ Construct a Transformed Distribution. #### `tf.contrib.distributions.ConditionalTransformedDistribution.allow_nan_stats` {#ConditionalTransformedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -27256,7 +27220,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -27273,7 +27237,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -27365,8 +27329,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -27473,7 +27437,7 @@ Additional documentation from `ConditionalTransformedDistribution`: #### `tf.contrib.distributions.ConditionalTransformedDistribution.validate_args` {#ConditionalTransformedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Bernoulli.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Bernoulli.md index b51d9785fa..0a5bca5052 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Bernoulli.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Bernoulli.md @@ -20,15 +20,15 @@ Construct Bernoulli distributions. Bernoulli distribution. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples. Default: `int32`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -40,21 +40,20 @@ Construct Bernoulli distributions. #### `tf.contrib.distributions.Bernoulli.allow_nan_stats` {#Bernoulli.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -252,7 +251,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -269,7 +268,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -413,8 +412,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -562,7 +561,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Bernoulli.validate_args` {#Bernoulli.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Chi2WithAbsDf.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Chi2WithAbsDf.md index c845a63e8e..a6b66f8560 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Chi2WithAbsDf.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Chi2WithAbsDf.md @@ -10,21 +10,20 @@ Chi2 with parameter transform `df = floor(abs(df))`. #### `tf.contrib.distributions.Chi2WithAbsDf.allow_nan_stats` {#Chi2WithAbsDf.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -236,7 +235,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -253,7 +252,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -349,7 +348,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -392,8 +391,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -541,7 +540,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Chi2WithAbsDf.validate_args` {#Chi2WithAbsDf.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Dirichlet.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Dirichlet.md index 5b140c4213..3d6690ab9a 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Dirichlet.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Dirichlet.md @@ -95,36 +95,35 @@ Initialize a batch of Dirichlet distributions. `concentration.shape = [N1, N2, ..., Nm, k]` then `batch_shape = [N1, N2, ..., Nm]` and `event_shape = [k]`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Dirichlet.allow_nan_stats` {#Dirichlet.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -329,7 +328,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -346,7 +345,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -450,7 +449,7 @@ Mode. Additional documentation from `Dirichlet`: Note: The mode is undefined when any `concentration <= 1`. If -`self.allow_nan_stats` is `True`, `NaN` is used for undefined modes. If +`self.allow_nan_stats` is `True`, `NaN` is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -494,8 +493,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -651,7 +650,7 @@ Sum of last dim of concentration parameter. #### `tf.contrib.distributions.Dirichlet.validate_args` {#Dirichlet.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Distribution.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Distribution.md index a2658ef44b..0076cdc4ff 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Distribution.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.Distribution.md @@ -6,8 +6,8 @@ A generic probability distribution base class. ### Subclassing Subclasses are expected to implement a leading-underscore version of the -same-named function. The argument signature should be identical except for -the omission of `name="..."`. For example, to enable `log_prob(value, +same-named function. The argument signature should be identical except for +the omission of `name="..."`. For example, to enable `log_prob(value, name="log_prob")` a subclass should implement `_log_prob(value)`. Subclasses can append to public-level docstrings by providing @@ -20,7 +20,7 @@ def _log_prob(self, value): ``` would add the string "Some other details." to the `log_prob` function -docstring. This is implemented as a simple decorator to avoid python +docstring. This is implemented as a simple decorator to avoid python linter complaining about missing Args/Returns/Raises sections in the partial docstrings. @@ -33,7 +33,7 @@ The shape of arguments to `__init__`, `cdf`, `log_cdf`, `prob`, and `log_prob` reflect this broadcasting, as does the return value of `sample` and `sample_n`. -`sample_n_shape = (n,) + batch_shape + event_shape`, where `sample_n_shape` is +`sample_n_shape = [n] + batch_shape + event_shape`, where `sample_n_shape` is the shape of the `Tensor` returned from `sample_n`, `n` is the number of samples, `batch_shape` defines how many independent distributions there are, and `event_shape` defines the shape of samples from each of those independent @@ -58,19 +58,19 @@ event_shape = u.event_shape # `event_shape_t` is a `Tensor` which will evaluate to []. event_shape_t = u.event_shape_tensor() -# Sampling returns a sample per distribution. `samples` has shape -# (5, 2, 2), which is (n,) + batch_shape + event_shape, where n=5, -# batch_shape=(2, 2), and event_shape=(). +# Sampling returns a sample per distribution. `samples` has shape +# [5, 2, 2], which is [n] + batch_shape + event_shape, where n=5, +# batch_shape=[2, 2], and event_shape=[]. samples = u.sample_n(5) # The broadcasting holds across methods. Here we use `cdf` as an example. The # same holds for `log_cdf` and the likelihood functions. -# `cum_prob` has shape (2, 2) as the `value` argument was broadcasted to the +# `cum_prob` has shape [2, 2] as the `value` argument was broadcasted to the # shape of the `Uniform` instance. cum_prob_broadcast = u.cdf(4.0) -# `cum_prob`'s shape is (2, 2), one per distribution. No broadcasting +# `cum_prob`'s shape is [2, 2], one per distribution. No broadcasting # occurred. cum_prob_per_dist = u.cdf([[4.0, 5.0], [6.0, 7.0]]) @@ -83,9 +83,9 @@ cum_prob_invalid = u.cdf([4.0, 5.0, 6.0]) ### Parameter values leading to undefined statistics or distributions. Some distributions do not have well-defined statistics for all initialization -parameter values. For example, the beta distribution is parameterized by -positive real numbers `a` and `b`, and does not have well-defined mode if -`a < 1` or `b < 1`. +parameter values. For example, the beta distribution is parameterized by +positive real numbers `concentration1` and `concentration0`, and does not have +well-defined mode if `concentration1 < 1` or `concentration0 < 1`. The user is given the option of raising an exception or returning `NaN`. @@ -122,25 +122,28 @@ Constructs the `Distribution`. * <b>`dtype`</b>: The type of the event samples. `None` implies no type-enforcement. -* <b>`is_continuous`</b>: Python boolean. If `True` this - `Distribution` is continuous over its supported domain. +* <b>`is_continuous`</b>: Python `bool`. If `True` this `Distribution` is continuous + over its supported domain. * <b>`reparameterization_type`</b>: Instance of `ReparameterizationType`. If `distributions.FULLY_REPARAMETERIZED`, this `Distribution` can be reparameterized in terms of some standard distribution with a function whose Jacobian is constant for the support - of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, + of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, then no such reparameterization is available. -* <b>`validate_args`</b>: Python boolean. Whether to validate input with asserts. - If `validate_args` is `False`, and the inputs are invalid, - correct behavior is not guaranteed. -* <b>`allow_nan_stats`</b>: Python boolean. If `False`, raise an - exception if a statistic (e.g., mean, mode) is undefined for any batch - member. If True, batch members with valid parameters leading to - undefined statistics will return `NaN` for this statistic. -* <b>`parameters`</b>: Python dictionary of parameters used to instantiate this +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution + parameters are checked for validity despite possibly degrading runtime + performance. When `False` invalid inputs may silently render incorrect + outputs. +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics + (e.g., mean, mode, variance) use the value "`NaN`" to indicate the + result is undefined. When `False`, an exception is raised if one or + more of the statistic's batch members are undefined. +* <b>`parameters`</b>: Python `dict` of parameters used to instantiate this `Distribution`. -* <b>`graph_parents`</b>: Python list of graph prerequisites of this `Distribution`. -* <b>`name`</b>: A name for this distribution. Default: subclass name. +* <b>`graph_parents`</b>: Python `list` of graph prerequisites of this + `Distribution`. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: + subclass name. ##### Raises: @@ -152,21 +155,20 @@ Constructs the `Distribution`. #### `tf.contrib.distributions.Distribution.allow_nan_stats` {#Distribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -364,7 +366,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -381,7 +383,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -514,8 +516,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -656,7 +658,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Distribution.validate_args` {#Distribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.GammaWithSoftplusConcentrationRate.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.GammaWithSoftplusConcentrationRate.md index 298fe91fe5..059ab2b546 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.GammaWithSoftplusConcentrationRate.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.GammaWithSoftplusConcentrationRate.md @@ -10,21 +10,20 @@ #### `tf.contrib.distributions.GammaWithSoftplusConcentrationRate.allow_nan_stats` {#GammaWithSoftplusConcentrationRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -229,7 +228,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -246,7 +245,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -342,7 +341,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -385,8 +384,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -534,7 +533,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.GammaWithSoftplusConcentrationRate.validate_args` {#GammaWithSoftplusConcentrationRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.md index 0b19b2f069..e99645559c 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.md @@ -10,21 +10,20 @@ #### `tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.allow_nan_stats` {#InverseGammaWithSoftplusConcentrationRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -229,7 +228,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -246,7 +245,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -336,7 +335,7 @@ Additional documentation from `InverseGamma`: The mean of an inverse gamma distribution is `rate / (concentration - 1)`, when `concentration > 1`, and `NaN` -otherwise. If `self.allow_nan_stats` is `False`, an exception will be +otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN` @@ -391,8 +390,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -540,7 +539,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.InverseGammaWithSoftplusConcentrationRate.validate_args` {#InverseGammaWithSoftplusConcentrationRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.RelaxedOneHotCategorical.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.RelaxedOneHotCategorical.md index e8fa8e5e63..acd76a6fa6 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.RelaxedOneHotCategorical.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.RelaxedOneHotCategorical.md @@ -90,9 +90,9 @@ Initialize RelaxedOneHotCategorical using class log-probabilities. of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). * <b>`validate_args`</b>: Unused in this distribution. -* <b>`allow_nan_stats`</b>: `Boolean`, default `True`. If `False`, raise an +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. If `False`, raise an exception if a statistic (e.g. mean/mode/etc...) is undefined for any - batch member. If `True`, batch members with valid parameters leading to + batch member. If `True`, batch members with valid parameters leading to undefined statistics will return NaN for this statistic. * <b>`name`</b>: A name for this distribution (optional). @@ -101,21 +101,20 @@ Initialize RelaxedOneHotCategorical using class log-probabilities. #### `tf.contrib.distributions.RelaxedOneHotCategorical.allow_nan_stats` {#RelaxedOneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -327,7 +326,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -344,7 +343,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -477,8 +476,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -619,7 +618,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.RelaxedOneHotCategorical.validate_args` {#RelaxedOneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.CholeskyOuterProduct.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.CholeskyOuterProduct.md index c3ca4f5131..78a9703bba 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.CholeskyOuterProduct.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.CholeskyOuterProduct.md @@ -25,9 +25,9 @@ Instantiates the `CholeskyOuterProduct` bijector. * <b>`event_ndims`</b>: `constant` `int32` scalar `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 2. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -281,7 +281,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.SigmoidCentered.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.SigmoidCentered.md index de0ab4bb9c..0d85d1e9e4 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.SigmoidCentered.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard0/tf.contrib.distributions.bijector.SigmoidCentered.md @@ -256,7 +256,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiag.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiag.md index 6724502112..c0aec27082 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiag.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiag.md @@ -142,15 +142,15 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then `k x k` identity matrices added to `scale`. When both `scale_identity_multiplier` and `scale_diag` are `None` then `scale` is the `Identity`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -162,21 +162,20 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then #### `tf.contrib.distributions.MultivariateNormalDiag.allow_nan_stats` {#MultivariateNormalDiag.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -395,7 +394,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -412,7 +411,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -575,8 +574,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -740,7 +739,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiag.validate_args` {#MultivariateNormalDiag.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.md index fcab6b60bd..0353b095bf 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.md @@ -10,21 +10,20 @@ MultivariateNormalDiag with `diag_stddev = softplus(diag_stddev)`. #### `tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.allow_nan_stats` {#MultivariateNormalDiagWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -243,7 +242,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -260,7 +259,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -423,8 +422,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -588,7 +587,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiagWithSoftplusScale.validate_args` {#MultivariateNormalDiagWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.QuantizedDistribution.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.QuantizedDistribution.md index 4cf3f02198..42bdb1eb24 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.QuantizedDistribution.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.QuantizedDistribution.md @@ -64,19 +64,19 @@ the `distribution`. * <b>`distribution`</b>: The base distribution class to transform. Typically an instance of `Distribution`. * <b>`low`</b>: `Tensor` with same `dtype` as this distribution and shape - able to be added to samples. Should be a whole number. Default `None`. + able to be added to samples. Should be a whole number. Default `None`. If provided, base distribution's `prob` should be defined at `low`. * <b>`high`</b>: `Tensor` with same `dtype` as this distribution and shape - able to be added to samples. Should be a whole number. Default `None`. + able to be added to samples. Should be a whole number. Default `None`. If provided, base distribution's `prob` should be defined at `high - 1`. `high` must be strictly greater than `low`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -90,21 +90,20 @@ the `distribution`. #### `tf.contrib.distributions.QuantizedDistribution.allow_nan_stats` {#QuantizedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -327,7 +326,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -344,7 +343,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -413,7 +412,7 @@ P[Y = y] := P[X <= low], if y == low, ``` -The base distribution's `log_cdf` method must be defined on `y - 1`. If the +The base distribution's `log_cdf` method must be defined on `y - 1`. If the base distribution has a `log_survival_function` method results will be more accurate for large values of `y`, and in this case the `log_survival_function` must also be defined on `y - 1`. @@ -531,8 +530,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -579,7 +578,7 @@ P[Y = y] := P[X <= low], if y == low, ``` -The base distribution's `cdf` method must be defined on `y - 1`. If the +The base distribution's `cdf` method must be defined on `y - 1`. If the base distribution has a `survival_function` method, results will be more accurate for large values of `y`, and in this case the `survival_function` must also be defined on `y - 1`. @@ -709,7 +708,7 @@ The base distribution's `cdf` method must be defined on `y - 1`. #### `tf.contrib.distributions.QuantizedDistribution.validate_args` {#QuantizedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentT.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentT.md index 6d8bff6612..569e5caab3 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentT.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentT.md @@ -28,7 +28,7 @@ Y = loc + scale * X ``` Notice that `scale` has semantics more similar to standard deviation than -variance. However it is not actually the std. deviation; the Student's +variance. However it is not actually the std. deviation; the Student's t-distribution std. dev. is `scale sqrt(df / (df - 2))` when `df > 2`. #### Examples @@ -83,22 +83,22 @@ supports broadcasting (e.g. `df + loc + scale` is a valid operation). ##### Args: -* <b>`df`</b>: Numeric `Tensor`. The degrees of freedom of the distribution(s). - `df` must contain only positive values. -* <b>`loc`</b>: Numeric `Tensor`. The mean(s) of the distribution(s). -* <b>`scale`</b>: Numeric `Tensor`. The scaling factor(s) for the distribution(s). - Note that `scale` is not technically the standard deviation of this - distribution but has semantics more similar to standard deviation than - variance. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`df`</b>: Floating-point `Tensor`. The degrees of freedom of the + distribution(s). `df` must contain only positive values. +* <b>`loc`</b>: Floating-point `Tensor`. The mean(s) of the distribution(s). +* <b>`scale`</b>: Floating-point `Tensor`. The scaling factor(s) for the + distribution(s). Note that `scale` is not technically the standard + deviation of this distribution but has semantics more similar to + standard deviation than variance. +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -110,21 +110,20 @@ supports broadcasting (e.g. `df + loc + scale` is a valid operation). #### `tf.contrib.distributions.StudentT.allow_nan_stats` {#StudentT.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -329,7 +328,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -346,7 +345,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -442,7 +441,7 @@ Mean. Additional documentation from `StudentT`: The mean of Student's T equals `loc` if `df > 1`, otherwise it is -`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised +`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised rather than returning `NaN`. @@ -492,8 +491,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -641,7 +640,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.StudentT.validate_args` {#StudentT.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.md index 8dac7a1de9..e21e0cca75 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.md @@ -10,21 +10,20 @@ StudentT with `df = floor(abs(df))` and `scale = softplus(scale)`. #### `tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.allow_nan_stats` {#StudentTWithAbsDfSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -229,7 +228,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -246,7 +245,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -342,7 +341,7 @@ Mean. Additional documentation from `StudentT`: The mean of Student's T equals `loc` if `df > 1`, otherwise it is -`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised +`NaN`. If `self.allow_nan_stats=True`, then an exception will be raised rather than returning `NaN`. @@ -392,8 +391,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -541,7 +540,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.StudentTWithAbsDfSoftplusScale.validate_args` {#StudentTWithAbsDfSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.TransformedDistribution.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.TransformedDistribution.md index 53876d7959..4c3d0fb0b2 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.TransformedDistribution.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard1/tf.contrib.distributions.TransformedDistribution.md @@ -17,7 +17,7 @@ We now describe how a `TransformedDistribution` alters the input/outputs of a Write `cdf(Y=y)` for an absolutely continuous cumulative distribution function of random variable `Y`; write the probability density function `pdf(Y=y) := d^k / (dy_1,...,dy_k) cdf(Y=y)` for its derivative wrt to `Y` evaluated at -`y`. Assume that `Y = g(X)` where `g` is a deterministic diffeomorphism, +`y`. Assume that `Y = g(X)` where `g` is a deterministic diffeomorphism, i.e., a non-random, continuous, differentiable, and invertible function. Write the inverse of `g` as `X = g^{-1}(Y)` and `(J o g)(x)` for the Jacobian of `g` evaluated at `x`. @@ -92,7 +92,7 @@ log_normal = ds.TransformedDistribution( forward_fn=tf.exp, inverse_fn=tf.log, inverse_log_det_jacobian_fn=( - lambda y: -tf.reduce_sum(tf.log(y), reduction_indices=-1)), + lambda y: -tf.reduce_sum(tf.log(y), axis=-1)), name="LogNormalTransformedDistribution") ``` @@ -108,7 +108,7 @@ normal = ds.TransformedDistribution( A `TransformedDistribution`'s batch- and event-shape are implied by the base distribution unless explicitly overridden by `batch_shape` or `event_shape` -arguments. Specifying an overriding `batch_shape` (`event_shape`) is +arguments. Specifying an overriding `batch_shape` (`event_shape`) is permitted only if the base distribution has scalar batch-shape (event-shape). The bijector is applied to the distribution as if the distribution possessed the overridden shape(s). The following example demonstrates how to construct a @@ -149,11 +149,11 @@ Construct a Transformed Distribution. `batch_shape`; valid only if `distribution.is_scalar_batch()`. * <b>`event_shape`</b>: `integer` vector `Tensor` which overrides `distribution` `event_shape`; valid only if `distribution.is_scalar_event()`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. Default: +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: `bijector.name + distribution.name`. @@ -161,21 +161,20 @@ Construct a Transformed Distribution. #### `tf.contrib.distributions.TransformedDistribution.allow_nan_stats` {#TransformedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -387,7 +386,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -404,7 +403,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -537,8 +536,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -679,7 +678,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.TransformedDistribution.validate_args` {#TransformedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Categorical.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Categorical.md index 49a2f63ee8..6e2b72c7f3 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Categorical.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Categorical.md @@ -58,36 +58,35 @@ Initialize Categorical distributions using class log-probabilities. represents a vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Categorical.allow_nan_stats` {#Categorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -292,7 +291,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -309,7 +308,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -449,8 +448,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -598,7 +597,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Categorical.validate_args` {#Categorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Chi2.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Chi2.md index e9fb06bcaf..76a28f8d17 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Chi2.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Chi2.md @@ -34,37 +34,36 @@ Construct Chi2 distributions with parameter `df`. * <b>`df`</b>: Floating point tensor, the degrees of freedom of the - distribution(s). `df` must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution + distribution(s). `df` must contain only positive values. +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Chi2.allow_nan_stats` {#Chi2.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -276,7 +275,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -293,7 +292,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -389,7 +388,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -432,8 +431,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -581,7 +580,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Chi2.validate_args` {#Chi2.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ConditionalDistribution.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ConditionalDistribution.md index 69ec5bc92f..97d31bb273 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ConditionalDistribution.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ConditionalDistribution.md @@ -14,25 +14,28 @@ Constructs the `Distribution`. * <b>`dtype`</b>: The type of the event samples. `None` implies no type-enforcement. -* <b>`is_continuous`</b>: Python boolean. If `True` this - `Distribution` is continuous over its supported domain. +* <b>`is_continuous`</b>: Python `bool`. If `True` this `Distribution` is continuous + over its supported domain. * <b>`reparameterization_type`</b>: Instance of `ReparameterizationType`. If `distributions.FULLY_REPARAMETERIZED`, this `Distribution` can be reparameterized in terms of some standard distribution with a function whose Jacobian is constant for the support - of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, + of the standard distribution. If `distributions.NOT_REPARAMETERIZED`, then no such reparameterization is available. -* <b>`validate_args`</b>: Python boolean. Whether to validate input with asserts. - If `validate_args` is `False`, and the inputs are invalid, - correct behavior is not guaranteed. -* <b>`allow_nan_stats`</b>: Python boolean. If `False`, raise an - exception if a statistic (e.g., mean, mode) is undefined for any batch - member. If True, batch members with valid parameters leading to - undefined statistics will return `NaN` for this statistic. -* <b>`parameters`</b>: Python dictionary of parameters used to instantiate this +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution + parameters are checked for validity despite possibly degrading runtime + performance. When `False` invalid inputs may silently render incorrect + outputs. +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics + (e.g., mean, mode, variance) use the value "`NaN`" to indicate the + result is undefined. When `False`, an exception is raised if one or + more of the statistic's batch members are undefined. +* <b>`parameters`</b>: Python `dict` of parameters used to instantiate this `Distribution`. -* <b>`graph_parents`</b>: Python list of graph prerequisites of this `Distribution`. -* <b>`name`</b>: A name for this distribution. Default: subclass name. +* <b>`graph_parents`</b>: Python `list` of graph prerequisites of this + `Distribution`. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: + subclass name. ##### Raises: @@ -44,21 +47,20 @@ Constructs the `Distribution`. #### `tf.contrib.distributions.ConditionalDistribution.allow_nan_stats` {#ConditionalDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -240,7 +242,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -257,7 +259,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -340,8 +342,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -442,7 +444,7 @@ denotes expectation, and `stddev.shape = batch_shape + event_shape`. #### `tf.contrib.distributions.ConditionalDistribution.validate_args` {#ConditionalDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ReparameterizationType.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ReparameterizationType.md index 503f71b8ec..35e5d87db8 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ReparameterizationType.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.ReparameterizationType.md @@ -8,7 +8,7 @@ one of two possible properties for samples from a distribution: `NOT_REPARAMETERIZED`: Samples from the distribution are not fully reparameterized, and straight-through gradients are either partially - unsupported or are not supported at all. In this case, for purposes of + unsupported or are not supported at all. In this case, for purposes of e.g. RL or variational inference, it is generally safest to wrap the sample results in a `stop_gradients` call and instead use policy gradients / surrogate loss instead. diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Uniform.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Uniform.md index 25673c069f..a3455aa9ea 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Uniform.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.Uniform.md @@ -49,15 +49,15 @@ Initialize a batch of Uniform distributions. have `low < high`. * <b>`high`</b>: Floating point tensor, upper boundary of the output interval. Must have `low < high`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -69,21 +69,20 @@ Initialize a batch of Uniform distributions. #### `tf.contrib.distributions.Uniform.allow_nan_stats` {#Uniform.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -288,7 +287,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -305,7 +304,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -445,8 +444,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -594,7 +593,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Uniform.validate_args` {#Uniform.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.WishartCholesky.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.WishartCholesky.md index acc1fa42bd..156e009dd4 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.WishartCholesky.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.WishartCholesky.md @@ -36,7 +36,7 @@ chol_scale = tf.cholesky(...) # Shape is [3, 3]. dist = tf.contrib.distributions.WishartCholesky(df=df, scale=chol_scale) # Evaluate this on an observation in R^3, returning a scalar. -x = ... # A 3x3 positive definite matrix. +x = ... # A 3x3 positive definite matrix. dist.prob(x) # Shape is [], a scalar. # Evaluate this on a two observations, each in R^{3x3}, returning a length two @@ -69,41 +69,40 @@ Construct Wishart distributions. or equal to dimension of the scale matrix. * <b>`scale`</b>: `float` or `double` `Tensor`. The Cholesky factorization of the symmetric positive definite scale matrix of the distribution. -* <b>`cholesky_input_output_matrices`</b>: `Boolean`. Any function which whose input - or output is a matrix assumes the input is Cholesky and returns a +* <b>`cholesky_input_output_matrices`</b>: Python `bool`. Any function which whose + input or output is a matrix assumes the input is Cholesky and returns a Cholesky factored matrix. Example `log_prob` input takes a Cholesky and `sample_n` returns a Cholesky when `cholesky_input_output_matrices=True`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.WishartCholesky.allow_nan_stats` {#WishartCholesky.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -322,7 +321,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -339,7 +338,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -486,8 +485,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -642,7 +641,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.WishartCholesky.validate_args` {#WishartCholesky.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.bijector.Bijector.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.bijector.Bijector.md index 6e3d6d78b6..bc383ea122 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.bijector.Bijector.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard2/tf.contrib.distributions.bijector.Bijector.md @@ -4,7 +4,7 @@ A `Bijector` implements a [diffeomorphism](https://en.wikipedia.org/wiki/Diffeomorphism), i.e., a bijective, differentiable function. A `Bijector` is used by `TransformedDistribution` but can be generally used for transforming a -`Distribution` generated `Tensor`. A `Bijector` is characterized by three +`Distribution` generated `Tensor`. A `Bijector` is characterized by three operations: 1. Forward Evaluation @@ -22,7 +22,7 @@ operations: "The log of the determinant of the matrix of all first-order partial derivatives of the inverse function." Useful for inverting a transformation to compute one probability in terms - of another. Geometrically, the det(Jacobian) is the volume of the + of another. Geometrically, the det(Jacobian) is the volume of the transformation and is used to scale the probability. By convention, transformations of random variables are named in terms of the @@ -34,7 +34,7 @@ Example Use: - Basic properties: ```python - x = ... # A tensor. + x = ... # A tensor. # Evaluate forward transformation. fwd_x = my_bijector.forward(x) x == my_bijector.inverse(fwd_x) @@ -91,7 +91,7 @@ Example transformations: if self.event_ndims is None: raise ValueError("Jacobian requires known event_ndims.") event_dims = array_ops.shape(x)[-self.event_ndims:] - return math_ops.reduce_sum(x, reduction_indices=event_dims) + return math_ops.reduce_sum(x, axis=event_dims) ``` - "Affine" @@ -116,8 +116,8 @@ Example of why a `Bijector` needs to understand sample, batch, event partitioning: - Consider the `Exp` `Bijector` applied to a `Tensor` which has sample, batch, - and event (S, B, E) shape semantics. Suppose - the `Tensor`'s partitioned-shape is `(S=[4], B=[2], E=[3, 3])`. + and event (S, B, E) shape semantics. Suppose the `Tensor`'s + partitioned-shape is `(S=[4], B=[2], E=[3, 3])`. For `Exp`, the shape of the `Tensor` returned by `forward` and `inverse` is unchanged, i.e., `[4, 2, 3, 3]`. However the shape returned by @@ -132,7 +132,7 @@ Subclass Requirements: - If the `Bijector`'s use is limited to `TransformedDistribution` (or friends like `QuantizedDistribution`) then depending on your use, you may not need - to implement all of `_forward` and `_inverse` functions. Examples: + to implement all of `_forward` and `_inverse` functions. Examples: 1. Sampling (e.g., `sample`) only requires `_forward`. 2. Probability functions (e.g., `prob`, `cdf`, `survival`) only require `_inverse` (and related). @@ -140,7 +140,7 @@ Subclass Requirements: `_inverse` can be implemented as a cache lookup. See `Example Use` [above] which shows how these functions are used to - transform a distribution. (Note: `_forward` could theoretically be + transform a distribution. (Note: `_forward` could theoretically be implemented as a cache lookup but this would require controlling the underlying sample generation mechanism.) @@ -158,7 +158,7 @@ Subclass Requirements: - Subclasses should implement `_forward_event_shape`, `_forward_event_shape_tensor` (and `inverse` counterparts) if the - transformation is shape-changing. By default the event-shape is assumed + transformation is shape-changing. By default the event-shape is assumed unchanged from input. Tips for implementing `_inverse` and `_inverse_log_det_jacobian`: @@ -167,14 +167,14 @@ Tips for implementing `_inverse` and `_inverse_log_det_jacobian`: can be implemented as a cache lookup. - The inverse `log o det o Jacobian` can be implemented as the negative of the - forward `log o det o Jacobian`. This is useful if the `inverse` is + forward `log o det o Jacobian`. This is useful if the `inverse` is implemented as a cache or the inverse Jacobian is computationally more expensive (e.g., `CholeskyOuterProduct` `Bijector`). The following demonstrates the suggested implementation. ```python def _inverse_and_log_det_jacobian(self, y): - x = # ... implement inverse, possibly via cache. + x = ... # implement inverse, possibly via cache. return x, -self._forward_log_det_jac(x) # Note negation. ``` @@ -233,10 +233,10 @@ See `Bijector` subclass docstring for more details and specific examples. * <b>`event_ndims`</b>: number of dimensions associated with event coordinates. * <b>`graph_parents`</b>: Python list of graph prerequisites of this `Bijector`. -* <b>`is_constant_jacobian`</b>: `Boolean` indicating that the Jacobian is not a +* <b>`is_constant_jacobian`</b>: Python `bool` indicating that the Jacobian is not a function of the input. -* <b>`validate_args`</b>: `Boolean`, default `False`. Whether to validate input with - asserts. If `validate_args` is `False`, and the inputs are invalid, +* <b>`validate_args`</b>: Python `bool`, default `False`. Whether to validate input + with asserts. If `validate_args` is `False`, and the inputs are invalid, correct behavior is not guaranteed. * <b>`dtype`</b>: `tf.dtype` supported by this `Bijector`. `None` means dtype is not enforced. @@ -489,7 +489,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Binomial.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Binomial.md index 2ee42bdf54..6baa4d2700 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Binomial.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Binomial.md @@ -72,7 +72,7 @@ Initialize a batch of Binomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor with shape broadcastable to `[N1,..., Nm]` with `m >= 0` and the same dtype as `probs` or - `logits`. Defines this as a batch of `N1 x ... x Nm` different Binomial + `logits`. Defines this as a batch of `N1 x ... x Nm` different Binomial distributions. Its components should be equal to integer values. * <b>`logits`</b>: Floating point tensor representing the log-odds of a positive event with shape broadcastable to `[N1,..., Nm]` `m >= 0`, and @@ -83,36 +83,35 @@ Initialize a batch of Binomial distributions. `[N1,..., Nm]` `m >= 0`, `probs in [0, 1]`. Each entry represents the probability of success for independent Binomial distributions. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Binomial.allow_nan_stats` {#Binomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -310,7 +309,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -327,7 +326,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -443,7 +442,7 @@ Mode. Additional documentation from `Binomial`: Note that when `(1 + total_count) * probs` is an integer, there are -actually two modes. Namely, `(1 + total_count) * probs` and +actually two modes. Namely, `(1 + total_count) * probs` and `(1 + total_count) * probs - 1` are both modes. Here we return only the larger of the two modes. @@ -487,8 +486,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -656,7 +655,7 @@ Number of trials. #### `tf.contrib.distributions.Binomial.validate_args` {#Binomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.DirichletMultinomial.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.DirichletMultinomial.md index b3afe9e04a..e9d08ed6b9 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.DirichletMultinomial.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.DirichletMultinomial.md @@ -96,43 +96,42 @@ Initialize a batch of DirichletMultinomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor, whose dtype is the same as `concentration`. The shape is broadcastable to `[N1,..., Nm]` with - `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different + `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different Dirichlet multinomial distributions. Its components should be equal to integer values. * <b>`concentration`</b>: Positive floating point tensor, whose dtype is the same as `n` with shape broadcastable to `[N1,..., Nm, k]` `m >= 0`. Defines this as a batch of `N1 x ... x Nm` different `k` class Dirichlet multinomial distributions. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.DirichletMultinomial.allow_nan_stats` {#DirichletMultinomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -357,7 +356,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -374,7 +373,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -416,10 +415,10 @@ Log probability density/mass function (depending on `is_continuous`). Additional documentation from `DirichletMultinomial`: For each batch of counts, -`value = [n_0, ... ,n_{k-1}]`, `P[value]` is the probability that after sampling -`self.total_count` draws from this Dirichlet-Multinomial distribution, the -number of draws falling in class `j` is `n_j`. Since this definition is -[exchangeable]( https://en.wikipedia.org/wiki/Exchangeable_random_variables); +`value = [n_0, ..., n_{k-1}]`, `P[value]` is the probability that after +sampling `self.total_count` draws from this Dirichlet-Multinomial distribution, +the number of draws falling in class `j` is `n_j`. Since this definition is +[exchangeable](https://en.wikipedia.org/wiki/Exchangeable_random_variables); different sequences have the same counts so the probability includes a combinatorial coefficient. @@ -523,8 +522,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -562,10 +561,10 @@ Probability density/mass function (depending on `is_continuous`). Additional documentation from `DirichletMultinomial`: For each batch of counts, -`value = [n_0, ... ,n_{k-1}]`, `P[value]` is the probability that after sampling -`self.total_count` draws from this Dirichlet-Multinomial distribution, the -number of draws falling in class `j` is `n_j`. Since this definition is -[exchangeable]( https://en.wikipedia.org/wiki/Exchangeable_random_variables); +`value = [n_0, ..., n_{k-1}]`, `P[value]` is the probability that after +sampling `self.total_count` draws from this Dirichlet-Multinomial distribution, +the number of draws falling in class `j` is `n_j`. Since this definition is +[exchangeable](https://en.wikipedia.org/wiki/Exchangeable_random_variables); different sequences have the same counts so the probability includes a combinatorial coefficient. @@ -695,7 +694,7 @@ Number of trials used to construct a sample. #### `tf.contrib.distributions.DirichletMultinomial.validate_args` {#DirichletMultinomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.ExpRelaxedOneHotCategorical.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.ExpRelaxedOneHotCategorical.md index 5a91aeddab..4a1f17a6d0 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.ExpRelaxedOneHotCategorical.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.ExpRelaxedOneHotCategorical.md @@ -110,36 +110,35 @@ Initialize ExpRelaxedOneHotCategorical using class log-probabilities. the last dimension represents a vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.ExpRelaxedOneHotCategorical.allow_nan_stats` {#ExpRelaxedOneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -344,7 +343,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -361,7 +360,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -501,8 +500,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -657,7 +656,7 @@ Batchwise temperature tensor of a RelaxedCategorical. #### `tf.contrib.distributions.ExpRelaxedOneHotCategorical.validate_args` {#ExpRelaxedOneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Exponential.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Exponential.md index 594031b723..6ce65a7b2f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Exponential.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Exponential.md @@ -38,36 +38,35 @@ Construct Exponential distribution with parameter `rate`. * <b>`rate`</b>: Floating point tensor, equivalent to `1 / mean`. Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Exponential.allow_nan_stats` {#Exponential.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -272,7 +271,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -289,7 +288,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -385,7 +384,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -428,8 +427,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -577,7 +576,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Exponential.validate_args` {#Exponential.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Gamma.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Gamma.md index 003fa91793..78246d36f8 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Gamma.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Gamma.md @@ -64,15 +64,15 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). distribution(s). Must contain only positive values. * <b>`rate`</b>: Floating point tensor, the inverse scale params of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -84,21 +84,20 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). #### `tf.contrib.distributions.Gamma.allow_nan_stats` {#Gamma.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -303,7 +302,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -320,7 +319,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -416,7 +415,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -459,8 +458,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -608,7 +607,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Gamma.validate_args` {#Gamma.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.InverseGamma.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.InverseGamma.md index 238bf39b47..e908ceac08 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.InverseGamma.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.InverseGamma.md @@ -64,15 +64,15 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). distribution(s). Must contain only positive values. * <b>`rate`</b>: Floating point tensor, the inverse scale params of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -85,21 +85,20 @@ supports broadcasting (e.g. `concentration + rate` is a valid operation). #### `tf.contrib.distributions.InverseGamma.allow_nan_stats` {#InverseGamma.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -304,7 +303,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -321,7 +320,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -411,7 +410,7 @@ Additional documentation from `InverseGamma`: The mean of an inverse gamma distribution is `rate / (concentration - 1)`, when `concentration > 1`, and `NaN` -otherwise. If `self.allow_nan_stats` is `False`, an exception will be +otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN` @@ -466,8 +465,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -615,7 +614,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.InverseGamma.validate_args` {#InverseGamma.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Multinomial.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Multinomial.md index 420e28bf73..796aece469 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Multinomial.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.Multinomial.md @@ -82,7 +82,7 @@ Initialize a batch of Multinomial distributions. * <b>`total_count`</b>: Non-negative floating point tensor with shape broadcastable to `[N1,..., Nm]` with `m >= 0`. Defines this as a batch of - `N1 x ... x Nm` different Multinomial distributions. Its components + `N1 x ... x Nm` different Multinomial distributions. Its components should be equal to integer values. * <b>`logits`</b>: Floating point tensor representing the log-odds of a positive event with shape broadcastable to `[N1,..., Nm, k], m >= 0`, @@ -90,40 +90,39 @@ Initialize a batch of Multinomial distributions. `N1 x ... x Nm` different `k` class Multinomial distributions. Only one of `logits` or `probs` should be passed in. * <b>`probs`</b>: Positive floating point tensor with shape broadcastable to - `[N1,..., Nm, k]` `m >= 0` and same dtype as `total_count`. Defines + `[N1,..., Nm, k]` `m >= 0` and same dtype as `total_count`. Defines this as a batch of `N1 x ... x Nm` different `k` class Multinomial distributions. `probs`'s components in the last portion of its shape should sum to `1`. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Multinomial.allow_nan_stats` {#Multinomial.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -321,7 +320,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -338,7 +337,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -494,8 +493,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -666,7 +665,7 @@ Number of trials used to construct a sample. #### `tf.contrib.distributions.Multinomial.validate_args` {#Multinomial.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.NormalWithSoftplusScale.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.NormalWithSoftplusScale.md index 5cd96f6d19..b9d6592fdb 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.NormalWithSoftplusScale.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.NormalWithSoftplusScale.md @@ -10,21 +10,20 @@ Normal with softplus applied to `scale`. #### `tf.contrib.distributions.NormalWithSoftplusScale.allow_nan_stats` {#NormalWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -222,7 +221,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -239,7 +238,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -379,8 +378,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -528,7 +527,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.NormalWithSoftplusScale.validate_args` {#NormalWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.OneHotCategorical.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.OneHotCategorical.md index 835cbffe13..6cf39c39de 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.OneHotCategorical.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard3/tf.contrib.distributions.OneHotCategorical.md @@ -9,11 +9,11 @@ Categorical has event_dim=() while OneHotCategorical has event_dim=K, where K is the number of classes. This class provides methods to create indexed batches of OneHotCategorical -distributions. If the provided `logits` or `probs` is rank 2 or higher, for +distributions. If the provided `logits` or `probs` is rank 2 or higher, for every fixed set of leading dimensions, the last dimension represents one -single OneHotCategorical distribution. When calling distribution +single OneHotCategorical distribution. When calling distribution functions (e.g. `dist.prob(x)`), `logits` and `x` are broadcast to the -same shape (if possible). In all cases, the last dimension of `logits,x` +same shape (if possible). In all cases, the last dimension of `logits,x` represents single OneHotCategorical distributions. #### Examples @@ -66,36 +66,35 @@ Initialize OneHotCategorical distributions using class log-probabilities. vector of probabilities for each class. Only one of `logits` or `probs` should be passed in. * <b>`dtype`</b>: The type of the event samples (default: int32). -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.OneHotCategorical.allow_nan_stats` {#OneHotCategorical.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -300,7 +299,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -317,7 +316,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -457,8 +456,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -606,7 +605,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.OneHotCategorical.validate_args` {#OneHotCategorical.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Invert.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Invert.md index 43a32752d6..ceafe4514f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Invert.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Invert.md @@ -29,9 +29,9 @@ return -self.inverse_log_det_jacobian(y, **kwargs) * <b>`bijector`</b>: Bijector instance. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. - - - @@ -287,7 +287,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Softplus.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Softplus.md index 7d021a9480..49527a0aea 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Softplus.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.bijector.Softplus.md @@ -275,7 +275,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.normal_conjugates_known_scale_predictive.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.normal_conjugates_known_scale_predictive.md index 6349d89fb0..fdbae8aaf4 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.normal_conjugates_known_scale_predictive.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.contrib.distributions.normal_conjugates_known_scale_predictive.md @@ -4,7 +4,7 @@ Posterior predictive Normal distribution w. conjugate prior on the mean. This model assumes that `n` observations (with sum `s`) come from a Normal with unknown mean `loc` (described by the Normal `prior`) -and known variance `scale^2`. The "known scale predictive" +and known variance `scale**2`. The "known scale predictive" is the distribution of new observations, conditioned on the existing observations and our prior. @@ -14,20 +14,20 @@ distribution(s) (also assumed Normal), and statistical estimates `s` (the sum(s) of the observations) and `n` (the number(s) of observations). -Calculates the Normal distribution(s) `p(x | sigma^2)`: +Calculates the Normal distribution(s) `p(x | sigma**2)`: ``` -p(x | sigma^2) = int N(x | mu, sigma^2) N(mu | prior.loc, prior.scale**2) dmu - = N(x | prior.loc, 1/(sigma^2 + prior.scale**2)) +p(x | sigma**2) = int N(x | mu, sigma**2)N(mu | prior.loc, prior.scale**2) dmu + = N(x | prior.loc, 1 / (sigma**2 + prior.scale**2)) ``` Returns the predictive posterior distribution object, with parameters -`(loc', scale'^2)`, where: +`(loc', scale'**2)`, where: ``` -sigma_n^2 = 1/(1/sigma0^2 + n/sigma^2), -mu' = (mu0/sigma0^2 + s/sigma^2) * sigma_n^2. -sigma'^2 = sigma_n^2 + sigma^2, +sigma_n**2 = 1/(1/sigma0**2 + n/sigma**2), +mu' = (mu0/sigma0**2 + s/sigma**2) * sigma_n**2. +sigma'**2 = sigma_n**2 + sigma**2, ``` Distribution parameters from `prior`, as well as `scale`, `s`, and `n`. @@ -40,8 +40,8 @@ will broadcast in the case of multidimensional sets of parameters. the prior distribution having parameters `(loc0, scale0)`. * <b>`scale`</b>: tensor of type `dtype`, taking values `scale > 0`. The known stddev parameter(s). -* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. -* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. +* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. +* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. ##### Returns: diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Affine.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Affine.md index 29a6a5662c..09cbf17110 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Affine.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Affine.md @@ -83,34 +83,34 @@ specified then `scale += IdentityMatrix`. Otherwise specifying a ##### Args: -* <b>`shift`</b>: Numeric `Tensor`. If this is set to `None`, no shift is applied. +* <b>`shift`</b>: Floating-point `Tensor`. If this is set to `None`, no shift is + applied. * <b>`scale_identity_multiplier`</b>: floating point rank 0 `Tensor` representing a scaling done to the identity matrix. When `scale_identity_multiplier = scale_diag = scale_tril = None` then `scale += IdentityMatrix`. Otherwise no scaled-identity-matrix is added to `scale`. -* <b>`scale_diag`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_diag` has shape [N1, N2, ... k], which represents a k x k +* <b>`scale_diag`</b>: Floating-point `Tensor` representing the diagonal matrix. + `scale_diag` has shape [N1, N2, ... k], which represents a k x k diagonal matrix. When `None` no diagonal term is added to `scale`. -* <b>`scale_tril`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_diag` has shape [N1, N2, ... k, k], which represents a k x k +* <b>`scale_tril`</b>: Floating-point `Tensor` representing the diagonal matrix. + `scale_diag` has shape [N1, N2, ... k, k], which represents a k x k lower triangular matrix. When `None` no `scale_tril` term is added to `scale`. The upper triangular elements above the diagonal are ignored. -* <b>`scale_perturb_factor`</b>: Numeric `Tensor` representing factor matrix with - last two dimensions of shape `(k, r)`. - When `None`, no rank-r update is added to `scale`. -* <b>`scale_perturb_diag`</b>: Numeric `Tensor` representing the diagonal matrix. - `scale_perturb_diag` has shape [N1, N2, ... r], which represents an - r x r Diagonal matrix. - When `None` low rank updates will take the form `scale_perturb_factor * - scale_perturb_factor.T`. +* <b>`scale_perturb_factor`</b>: Floating-point `Tensor` representing factor matrix + with last two dimensions of shape `(k, r)`. When `None`, no rank-r + update is added to `scale`. +* <b>`scale_perturb_diag`</b>: Floating-point `Tensor` representing the diagonal + matrix. `scale_perturb_diag` has shape [N1, N2, ... r], which + represents an `r x r` diagonal matrix. When `None` low rank updates will + take the form `scale_perturb_factor * scale_perturb_factor.T`. * <b>`event_ndims`</b>: Scalar `int32` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 1. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -365,7 +365,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Chain.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Chain.md index 3970bd44ad..98a4130981 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Chain.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Chain.md @@ -40,10 +40,10 @@ Instantiates `Chain` bijector. * <b>`bijectors`</b>: Python list of bijector instances. An empty list makes this bijector equivalent to the `Identity` bijector. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. Default: E.g., - `Chain([Exp(), Softplus()]).name == "chain_of_exp_of_softplus"`. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. Default: + E.g., `Chain([Exp(), Softplus()]).name == "chain_of_exp_of_softplus"`. ##### Raises: @@ -304,7 +304,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Exp.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Exp.md index 33e7d15496..9fde10ec22 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Exp.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard5/tf.contrib.distributions.bijector.Exp.md @@ -27,9 +27,9 @@ Instantiates the `Exp` bijector. * <b>`event_ndims`</b>: Scalar `int32` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. - - - @@ -278,7 +278,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.BernoulliWithSigmoidProbs.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.BernoulliWithSigmoidProbs.md index 7c51c70b9b..b0a926f8ed 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.BernoulliWithSigmoidProbs.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.BernoulliWithSigmoidProbs.md @@ -10,21 +10,20 @@ Bernoulli with `probs = nn.sigmoid(logits)`. #### `tf.contrib.distributions.BernoulliWithSigmoidProbs.allow_nan_stats` {#BernoulliWithSigmoidProbs.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -222,7 +221,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -239,7 +238,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -383,8 +382,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -532,7 +531,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.BernoulliWithSigmoidProbs.validate_args` {#BernoulliWithSigmoidProbs.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Beta.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Beta.md index 5b2fc3113b..ed9f312153 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Beta.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Beta.md @@ -87,36 +87,35 @@ Initialize a batch of Beta distributions. * <b>`concentration0`</b>: Positive floating-point `Tensor` indicating mean number of failures; aka "beta". Otherwise has same semantics as `concentration1`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Beta.allow_nan_stats` {#Beta.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -334,7 +333,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -351,7 +350,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -460,7 +459,7 @@ Additional documentation from `Beta`: Note: The mode is undefined when `concentration1 <= 1` or `concentration0 <= 1`. If `self.allow_nan_stats` is `True`, `NaN` -is used for undefined modes. If `self.allow_nan_stats` is `False` an +is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -503,8 +502,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -658,7 +657,7 @@ Sum of concentration parameters. #### `tf.contrib.distributions.Beta.validate_args` {#Beta.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Laplace.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Laplace.md index 3a814777a8..3f31604508 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Laplace.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Laplace.md @@ -38,15 +38,15 @@ broadcasting (e.g., `loc / scale` is a valid operation). of the distribution. * <b>`scale`</b>: Positive floating point tensor which characterizes the spread of the distribution. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -58,21 +58,20 @@ broadcasting (e.g., `loc / scale` is a valid operation). #### `tf.contrib.distributions.Laplace.allow_nan_stats` {#Laplace.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -270,7 +269,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -287,7 +286,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -427,8 +426,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -576,7 +575,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Laplace.validate_args` {#Laplace.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.LaplaceWithSoftplusScale.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.LaplaceWithSoftplusScale.md index 42a7ff92ae..998d117e8f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.LaplaceWithSoftplusScale.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.LaplaceWithSoftplusScale.md @@ -10,21 +10,20 @@ Laplace with softplus applied to `scale`. #### `tf.contrib.distributions.LaplaceWithSoftplusScale.allow_nan_stats` {#LaplaceWithSoftplusScale.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -222,7 +221,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -239,7 +238,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -379,8 +378,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -528,7 +527,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.LaplaceWithSoftplusScale.validate_args` {#LaplaceWithSoftplusScale.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Logistic.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Logistic.md index fc0a45d2b3..6f0d4f2210 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Logistic.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.Logistic.md @@ -68,13 +68,13 @@ broadcasting (e.g. `loc + scale` is a valid operation). * <b>`loc`</b>: Floating point tensor, the means of the distribution(s). * <b>`scale`</b>: Floating point tensor, the scales of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. * <b>`name`</b>: The name to give Ops created by the initializer. @@ -88,21 +88,20 @@ broadcasting (e.g. `loc + scale` is a valid operation). #### `tf.contrib.distributions.Logistic.allow_nan_stats` {#Logistic.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -300,7 +299,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -317,7 +316,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -457,8 +456,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -606,7 +605,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Logistic.validate_args` {#Logistic.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.AffineLinearOperator.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.AffineLinearOperator.md index b986c37962..034f96e6cf 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.AffineLinearOperator.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.AffineLinearOperator.md @@ -6,7 +6,7 @@ If `X` is a scalar then the forward transformation is: `scale * X + shift` where `*` denotes the scalar product. Note: we don't always simply transpose `X` (but write it this way for -brevity). Actually the input `X` undergoes the following transformation +brevity). Actually the input `X` undergoes the following transformation before being premultiplied by `scale`: 1. If there are no sample dims, we call `X = tf.expand_dims(X, 0)`, i.e., @@ -21,8 +21,8 @@ before being premultiplied by `scale`: (For more details see `shape.make_batch_of_event_sample_matrices`.) The result of the above transformation is that `X` can be regarded as a batch -of matrices where each column is a draw from the distribution. After -premultiplying by `scale`, we take the inverse of this procedure. The input +of matrices where each column is a draw from the distribution. After +premultiplying by `scale`, we take the inverse of this procedure. The input `Y` also undergoes the same transformation before/after premultiplying by `inv(scale)`. @@ -60,14 +60,14 @@ Instantiates the `AffineLinearOperator` bijector. ##### Args: -* <b>`shift`</b>: Numeric `Tensor`. -* <b>`scale`</b>: Subclass of `LinearOperator`. Represents the (batch) positive +* <b>`shift`</b>: Floating-point `Tensor`. +* <b>`scale`</b>: Subclass of `LinearOperator`. Represents the (batch) positive definite matrix `M` in `R^{k x k}`. * <b>`event_ndims`</b>: Scalar `integer` `Tensor` indicating the number of dimensions associated with a particular draw from the distribution. Must be 0 or 1. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -324,7 +324,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.Identity.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.Identity.md index 970730ef0f..a6045db420 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.Identity.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard6/tf.contrib.distributions.bijector.Identity.md @@ -263,7 +263,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ConditionalTransformedDistribution.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ConditionalTransformedDistribution.md index 150a0afa9d..6607f5a275 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ConditionalTransformedDistribution.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ConditionalTransformedDistribution.md @@ -16,11 +16,11 @@ Construct a Transformed Distribution. `batch_shape`; valid only if `distribution.is_scalar_batch()`. * <b>`event_shape`</b>: `integer` vector `Tensor` which overrides `distribution` `event_shape`; valid only if `distribution.is_scalar_event()`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. Default: +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. Default: `bijector.name + distribution.name`. @@ -28,21 +28,20 @@ Construct a Transformed Distribution. #### `tf.contrib.distributions.ConditionalTransformedDistribution.allow_nan_stats` {#ConditionalTransformedDistribution.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -241,7 +240,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -258,7 +257,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -350,8 +349,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -458,7 +457,7 @@ Additional documentation from `ConditionalTransformedDistribution`: #### `tf.contrib.distributions.ConditionalTransformedDistribution.validate_args` {#ConditionalTransformedDistribution.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ExponentialWithSoftplusRate.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ExponentialWithSoftplusRate.md index a66d4e1c45..d5ccf96744 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ExponentialWithSoftplusRate.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.ExponentialWithSoftplusRate.md @@ -10,21 +10,20 @@ Exponential with softplus transform on `rate`. #### `tf.contrib.distributions.ExponentialWithSoftplusRate.allow_nan_stats` {#ExponentialWithSoftplusRate.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -229,7 +228,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -246,7 +245,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -342,7 +341,7 @@ Mode. Additional documentation from `Gamma`: The mode of a gamma distribution is `(shape - 1) / rate` when -`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, +`shape > 1`, and `NaN` otherwise. If `self.allow_nan_stats` is `False`, an exception will be raised rather than returning `NaN`. @@ -385,8 +384,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -534,7 +533,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.ExponentialWithSoftplusRate.validate_args` {#ExponentialWithSoftplusRate.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.md index 4c70d93a55..c5650c8055 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.md @@ -160,18 +160,18 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then * <b>`scale_perturb_diag`</b>: Floating-point `Tensor` representing a diagonal matrix inside the rank-`r` perturbation added to `scale`. May have shape `[B1, ..., Bb, r]`, `b >= 0`, and characterizes `b`-batches of `r x r` - diagonal matrices inside the perturbation added to `scale`. When + diagonal matrices inside the perturbation added to `scale`. When `None`, an identity matrix is used inside the perturbation. Can only be specified if `scale_perturb_factor` is also specified. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -183,21 +183,20 @@ If both `scale_diag` and `scale_identity_multiplier` are `None`, then #### `tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.allow_nan_stats` {#MultivariateNormalDiagPlusLowRank.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -416,7 +415,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -433,7 +432,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -596,8 +595,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -761,7 +760,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalDiagPlusLowRank.validate_args` {#MultivariateNormalDiagPlusLowRank.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.Normal.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.Normal.md index 43a405fd4e..5454ae907f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.Normal.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.Normal.md @@ -70,15 +70,15 @@ broadcasting (e.g. `loc + scale` is a valid operation). * <b>`loc`</b>: Floating point tensor; the means of the distribution(s). * <b>`scale`</b>: Floating point tensor; the stddevs of the distribution(s). Must contain only positive values. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to - indicate the result is undefined. When `False`, an exception is raised + indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -90,21 +90,20 @@ broadcasting (e.g. `loc + scale` is a valid operation). #### `tf.contrib.distributions.Normal.allow_nan_stats` {#Normal.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -302,7 +301,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -319,7 +318,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -459,8 +458,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -608,7 +607,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Normal.validate_args` {#Normal.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.RelaxedBernoulli.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.RelaxedBernoulli.md index 0d428433f8..f7af72d0f2 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.RelaxedBernoulli.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.RelaxedBernoulli.md @@ -116,15 +116,15 @@ Construct RelaxedBernoulli distributions. * <b>`probs`</b>: An N-D `Tensor` representing the probability of a positive event. Each entry in the `Tensor` parameterizes an independent Bernoulli distribution. Only one of `logits` or `probs` should be passed in. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -136,21 +136,20 @@ Construct RelaxedBernoulli distributions. #### `tf.contrib.distributions.RelaxedBernoulli.allow_nan_stats` {#RelaxedBernoulli.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -362,7 +361,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -379,7 +378,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -519,8 +518,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -675,7 +674,7 @@ Distribution parameter for the location. #### `tf.contrib.distributions.RelaxedBernoulli.validate_args` {#RelaxedBernoulli.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.Inline.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.Inline.md index 95e118cef3..c941b5ef65 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.Inline.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.Inline.md @@ -7,7 +7,7 @@ exp = Inline( forward_fn=tf.exp, inverse_fn=tf.log, inverse_log_det_jacobian_fn=( - lambda y: -tf.reduce_sum(tf.log(y), reduction_indices=-1)), + lambda y: -tf.reduce_sum(tf.log(y), axis=-1)), name="exp") ``` @@ -35,11 +35,11 @@ Creates a `Bijector` from callables. static event shape changes. Default: shape is assumed unchanged. * <b>`inverse_event_shape_tensor_fn`</b>: Python callable implementing non-identical event shape changes. Default: shape is assumed unchanged. -* <b>`is_constant_jacobian`</b>: `Boolean` indicating that the Jacobian is constant - for all input arguments. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String`, name given to ops managed by this object. +* <b>`is_constant_jacobian`</b>: Python `bool` indicating that the Jacobian is + constant for all input arguments. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str`, name given to ops managed by this object. - - - @@ -288,7 +288,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.PowerTransform.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.PowerTransform.md index da166d0a2d..d95946499f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.PowerTransform.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.bijector.PowerTransform.md @@ -18,9 +18,9 @@ Instantiates the `PowerTransform` bijector. `Y = g(X) = (1 + X * c)**(1 / c)` where `c` is the `power`. * <b>`event_ndims`</b>: Python scalar indicating the number of dimensions associated with a particular draw from the distribution. -* <b>`validate_args`</b>: `Boolean` indicating whether arguments should be checked - for correctness. -* <b>`name`</b>: `String` name given to ops managed by this object. +* <b>`validate_args`</b>: Python `bool` indicating whether arguments should be + checked for correctness. +* <b>`name`</b>: Python `str` name given to ops managed by this object. ##### Raises: @@ -274,7 +274,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.kl.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.kl.md index 08378205ac..59a41b2dd4 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.kl.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard7/tf.contrib.distributions.kl.md @@ -22,7 +22,7 @@ identified in the search is used (favoring a shorter MRO distance to * <b>`dist_b`</b>: The second distribution. * <b>`allow_nan`</b>: If `False` (default), a runtime error is raised if the KL returns NaN values for any batch entry of the given - distributions. If `True`, the KL may return a NaN for the given entry. + distributions. If `True`, the KL may return a NaN for the given entry. * <b>`name`</b>: (optional) Name scope to use for created operations. ##### Returns: diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.Mixture.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.Mixture.md index 4238739932..9799e8b23e 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.Mixture.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.Mixture.md @@ -28,13 +28,13 @@ time and match `len(components)`. * <b>`components`</b>: A list or tuple of `Distribution` instances. Each instance must have the same type, be defined on the same domain, and have matching `event_shape` and `batch_shape`. -* <b>`validate_args`</b>: `Boolean`, default `False`. If `True`, raise a runtime +* <b>`validate_args`</b>: Python `bool`, default `False`. If `True`, raise a runtime error if batch or event ranks are inconsistent between cat and any of - the distributions. This is only checked if the ranks cannot be + the distributions. This is only checked if the ranks cannot be determined statically at graph construction time. -* <b>`allow_nan_stats`</b>: Boolean, default `True`. If `False`, raise an +* <b>`allow_nan_stats`</b>: Boolean, default `True`. If `False`, raise an exception if a statistic (e.g. mean/mode/etc...) is undefined for any - batch member. If `True`, batch members with valid parameters leading to + batch member. If `True`, batch members with valid parameters leading to undefined statistics will return NaN for this statistic. * <b>`name`</b>: A name for this distribution (optional). @@ -57,21 +57,20 @@ time and match `len(components)`. #### `tf.contrib.distributions.Mixture.allow_nan_stats` {#Mixture.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -248,7 +247,7 @@ distribution: \\) where \\( p \\) is the prior distribution, \\( q \\) is the variational, -and \\( H[q] \\) is the entropy of \\( q \\). If there is a lower bound +and \\( H[q] \\) is the entropy of \\( q \\). If there is a lower bound \\( G[q] \\) such that \\( H[q] \geq G[q] \\) then it can be used in place of \\( H[q] \\). @@ -329,7 +328,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -346,7 +345,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -486,8 +485,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -628,7 +627,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Mixture.validate_args` {#Mixture.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.bijector.SoftmaxCentered.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.bijector.SoftmaxCentered.md index 3e5f7de3d6..1e47513e6b 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.bijector.SoftmaxCentered.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.bijector.SoftmaxCentered.md @@ -2,7 +2,7 @@ Bijector which computes `Y = g(X) = exp([X 0]) / sum(exp([X 0]))`. To implement [softmax](https://en.wikipedia.org/wiki/Softmax_function) as a bijection, the forward transformation appends a value to the input and the -inverse removes this coordinate. The appended coordinate represents a pivot, +inverse removes this coordinate. The appended coordinate represents a pivot, e.g., `softmax(x) = exp(x-c) / sum(exp(x-c))` where `c` is the implicit last coordinate. @@ -23,7 +23,7 @@ bijector.SoftmaxCentered(event_ndims=1).inverse([0.2, 0.3, 0.4, 0.1]) At first blush it may seem like the [Invariance of domain]( https://en.wikipedia.org/wiki/Invariance_of_domain) theorem implies this -implementation is not a bijection. However, the appended dimension +implementation is not a bijection. However, the appended dimension makes the (forward) image non-open and the theorem does not directly apply. - - - @@ -278,7 +278,8 @@ Note: Jacobian is either constant for both forward and inverse or neither. ##### Returns: - `Boolean`. + +* <b>`is_constant_jacobian`</b>: Python `bool`. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.matrix_diag_transform.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.matrix_diag_transform.md index 6f0edd8304..1f39a487ba 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.matrix_diag_transform.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.matrix_diag_transform.md @@ -9,7 +9,7 @@ Create a trainable covariance defined by a Cholesky factor: matrix_values = tf.contrib.layers.fully_connected(activations, 4) matrix = tf.reshape(matrix_values, (batch_size, 2, 2)) -# Make the diagonal positive. If the upper triangle was zero, this would be a +# Make the diagonal positive. If the upper triangle was zero, this would be a # valid Cholesky factor. chol = matrix_diag_transform(matrix, transform=tf.nn.softplus) @@ -31,7 +31,7 @@ mu = tf.contrib.layers.fully_connected(activations, 2) # This is a fully trainable multivariate normal! dist = tf.contrib.distributions.MVNCholesky(mu, chol) -# Standard log loss. Minimizing this will "train" mu and chol, and then dist +# Standard log loss. Minimizing this will "train" mu and chol, and then dist # will be a distribution predicting labels as multivariate Gaussians. loss = -1 * tf.reduce_mean(dist.log_prob(labels)) ``` @@ -41,9 +41,9 @@ loss = -1 * tf.reduce_mean(dist.log_prob(labels)) * <b>`matrix`</b>: Rank `R` `Tensor`, `R >= 2`, where the last two dimensions are equal. -* <b>`transform`</b>: Element-wise function mapping `Tensors` to `Tensors`. To - be applied to the diagonal of `matrix`. If `None`, `matrix` is returned - unchanged. Defaults to `None`. +* <b>`transform`</b>: Element-wise function mapping `Tensors` to `Tensors`. To + be applied to the diagonal of `matrix`. If `None`, `matrix` is returned + unchanged. Defaults to `None`. * <b>`name`</b>: A name to give created ops. Defaults to "matrix_diag_transform". diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.normal_conjugates_known_scale_posterior.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.normal_conjugates_known_scale_posterior.md index 9b2cce1cc4..554d553e77 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.normal_conjugates_known_scale_posterior.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard8/tf.contrib.distributions.normal_conjugates_known_scale_posterior.md @@ -4,7 +4,7 @@ Posterior Normal distribution with conjugate prior on the mean. This model assumes that `n` observations (with sum `s`) come from a Normal with unknown mean `loc` (described by the Normal `prior`) -and known variance `scale^2`. The "known scale posterior" is +and known variance `scale**2`. The "known scale posterior" is the distribution of the unknown `loc`. Accepts a prior Normal distribution object, having parameters @@ -14,12 +14,12 @@ and statistical estimates `s` (the sum(s) of the observations) and `n` (the number(s) of observations). Returns a posterior (also Normal) distribution object, with parameters -`(loc', scale'^2)`, where: +`(loc', scale'**2)`, where: ``` -mu ~ N(mu', sigma'^2) -sigma'^2 = 1/(1/sigma0^2 + n/sigma^2), -mu' = (mu0/sigma0^2 + s/sigma^2) * sigma'^2. +mu ~ N(mu', sigma'**2) +sigma'**2 = 1/(1/sigma0**2 + n/sigma**2), +mu' = (mu0/sigma0**2 + s/sigma**2) * sigma'**2. ``` Distribution parameters from `prior`, as well as `scale`, `s`, and `n`. @@ -32,8 +32,8 @@ will broadcast in the case of multidimensional sets of parameters. the prior distribution having parameters `(loc0, scale0)`. * <b>`scale`</b>: tensor of type `dtype`, taking values `scale > 0`. The known stddev parameter(s). -* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. -* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. +* <b>`s`</b>: Tensor of type `dtype`. The sum(s) of observations. +* <b>`n`</b>: Tensor of type `int`. The number(s) of observations. ##### Returns: diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.BetaWithSoftplusConcentration.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.BetaWithSoftplusConcentration.md index 62a630022c..e01f84653b 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.BetaWithSoftplusConcentration.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.BetaWithSoftplusConcentration.md @@ -10,21 +10,20 @@ Beta with softplus transform of `concentration1` and `concentration0`. #### `tf.contrib.distributions.BetaWithSoftplusConcentration.allow_nan_stats` {#BetaWithSoftplusConcentration.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -242,7 +241,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -259,7 +258,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -368,7 +367,7 @@ Additional documentation from `Beta`: Note: The mode is undefined when `concentration1 <= 1` or `concentration0 <= 1`. If `self.allow_nan_stats` is `True`, `NaN` -is used for undefined modes. If `self.allow_nan_stats` is `False` an +is used for undefined modes. If `self.allow_nan_stats` is `False` an exception is raised when one or more modes are undefined. @@ -411,8 +410,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -566,7 +565,7 @@ Sum of concentration parameters. #### `tf.contrib.distributions.BetaWithSoftplusConcentration.validate_args` {#BetaWithSoftplusConcentration.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.MultivariateNormalTriL.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.MultivariateNormalTriL.md index 5c79f551bf..4bd0c96189 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.MultivariateNormalTriL.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.MultivariateNormalTriL.md @@ -121,15 +121,15 @@ Additional leading dimensions (if any) will index batches. * <b>`scale_tril`</b>: Floating-point, lower-triangular `Tensor` with non-zero diagonal elements. `scale_tril` has shape `[B1, ..., Bb, k, k]` where `b >= 0` and `k` is the event size. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. ##### Raises: @@ -141,21 +141,20 @@ Additional leading dimensions (if any) will index batches. #### `tf.contrib.distributions.MultivariateNormalTriL.allow_nan_stats` {#MultivariateNormalTriL.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -374,7 +373,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -391,7 +390,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -554,8 +553,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -719,7 +718,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.MultivariateNormalTriL.validate_args` {#MultivariateNormalTriL.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.Poisson.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.Poisson.md index 8e5e27f795..1a52643a32 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.Poisson.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.Poisson.md @@ -23,36 +23,35 @@ Initialize a batch of Poisson distributions. * <b>`rate`</b>: Floating point tensor, the rate parameter of the distribution(s). `rate` must be positive. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.Poisson.allow_nan_stats` {#Poisson.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -257,7 +256,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -274,7 +273,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -426,8 +425,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -582,7 +581,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.Poisson.validate_args` {#Poisson.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.WishartFull.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.WishartFull.md index db8aefb189..eefa558fca 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.WishartFull.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard9/tf.contrib.distributions.WishartFull.md @@ -32,7 +32,7 @@ scale = ... # Shape is [3, 3]; positive definite. dist = tf.contrib.distributions.WishartFull(df=df, scale=scale) # Evaluate this on an observation in R^3, returning a scalar. -x = ... # A 3x3 positive definite matrix. +x = ... # A 3x3 positive definite matrix. dist.prob(x) # Shape is [], a scalar. # Evaluate this on a two observations, each in R^{3x3}, returning a length two @@ -65,41 +65,40 @@ Construct Wishart distributions. or equal to dimension of the scale matrix. * <b>`scale`</b>: `float` or `double` `Tensor`. The symmetric positive definite scale matrix of the distribution. -* <b>`cholesky_input_output_matrices`</b>: `Boolean`. Any function which whose input - or output is a matrix assumes the input is Cholesky and returns a +* <b>`cholesky_input_output_matrices`</b>: Python `bool`. Any function which whose + input or output is a matrix assumes the input is Cholesky and returns a Cholesky factored matrix. Example `log_prob` input takes a Cholesky and `sample_n` returns a Cholesky when `cholesky_input_output_matrices=True`. -* <b>`validate_args`</b>: Python `Boolean`, default `False`. When `True` distribution +* <b>`validate_args`</b>: Python `bool`, default `False`. When `True` distribution parameters are checked for validity despite possibly degrading runtime performance. When `False` invalid inputs may silently render incorrect outputs. -* <b>`allow_nan_stats`</b>: Python `Boolean`, default `True`. When `True`, statistics +* <b>`allow_nan_stats`</b>: Python `bool`, default `True`. When `True`, statistics (e.g., mean, mode, variance) use the value "`NaN`" to indicate the - result is undefined. When `False`, an exception is raised if one or + result is undefined. When `False`, an exception is raised if one or more of the statistic's batch members are undefined. -* <b>`name`</b>: `String` name prefixed to Ops created by this class. +* <b>`name`</b>: Python `str` name prefixed to Ops created by this class. - - - #### `tf.contrib.distributions.WishartFull.allow_nan_stats` {#WishartFull.allow_nan_stats} -Python boolean describing behavior when a stat is undefined. +Python `bool` describing behavior when a stat is undefined. -Stats return +/- infinity when it makes sense. E.g., the variance -of a Cauchy distribution is infinity. However, sometimes the -statistic is undefined, e.g., if a distribution's pdf does not achieve a -maximum within the support of the distribution, the mode is undefined. -If the mean is undefined, then by definition the variance is undefined. -E.g. the mean for Student's T for df = 1 is undefined (no clear way to say -it is either + or - infinity), so the variance = E[(X - mean)^2] is also -undefined. +Stats return +/- infinity when it makes sense. E.g., the variance of a +Cauchy distribution is infinity. However, sometimes the statistic is +undefined, e.g., if a distribution's pdf does not achieve a maximum within +the support of the distribution, the mode is undefined. If the mean is +undefined, then by definition the variance is undefined. E.g. the mean for +Student's T for df = 1 is undefined (no clear way to say it is either + or - +infinity), so the variance = E[(X - mean)**2] is also undefined. ##### Returns: -* <b>`allow_nan_stats`</b>: Python boolean. +* <b>`allow_nan_stats`</b>: Python `bool`. - - - @@ -318,7 +317,7 @@ Indicates that `batch_shape == []`. ##### Returns: -* <b>`is_scalar_batch`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_batch`</b>: `bool` scalar `Tensor`. - - - @@ -335,7 +334,7 @@ Indicates that `event_shape == []`. ##### Returns: -* <b>`is_scalar_event`</b>: `Boolean` `scalar` `Tensor`. +* <b>`is_scalar_event`</b>: `bool` scalar `Tensor`. - - - @@ -482,8 +481,8 @@ param_shapes with static (i.e. `TensorShape`) shapes. This is a class method that describes what key/value arguments are required to instantiate the given `Distribution` so that a particular shape is -returned for that instance's call to `sample()`. Assumes that -the sample's shape is known statically. +returned for that instance's call to `sample()`. Assumes that the sample's +shape is known statically. Subclasses should override class method `_param_shapes` to return constant-valued tensors when constant values are fed. @@ -638,7 +637,7 @@ survival_function(x) = P[X > x] #### `tf.contrib.distributions.WishartFull.validate_args` {#WishartFull.validate_args} -Python boolean indicated possibly expensive checks are enabled. +Python `bool` indicating possibly expensive checks are enabled. - - - |