1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
|
Bijector which inverts another Bijector.
Example Use: [ExpGammaDistribution (see Background & Context)](
https://reference.wolfram.com/language/ref/ExpGammaDistribution.html)
models `Y=log(X)` where `X ~ Gamma`.
```python
exp_gamma_distribution = TransformedDistribution(
distribution=Gamma(concentration=1., rate=2.),
bijector=bijector.Invert(bijector.Exp())
```
- - -
#### `tf.contrib.distributions.bijector.Invert.__init__(bijector, validate_args=False, name=None)` {#Invert.__init__}
Creates a `Bijector` which swaps the meaning of `inverse` and `forward`.
Note: An inverted bijector's `inverse_log_det_jacobian` is often more
efficient if the base bijector implements `_forward_log_det_jacobian`. If
`_forward_log_det_jacobian` is not implemented then the following code is
used:
```python
y = self.inverse(x, **kwargs)
return -self.inverse_log_det_jacobian(y, **kwargs)
```
##### Args:
* <b>`bijector`</b>: Bijector instance.
* <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.
- - -
#### `tf.contrib.distributions.bijector.Invert.bijector` {#Invert.bijector}
- - -
#### `tf.contrib.distributions.bijector.Invert.dtype` {#Invert.dtype}
dtype of `Tensor`s transformable by this distribution.
- - -
#### `tf.contrib.distributions.bijector.Invert.event_ndims` {#Invert.event_ndims}
Returns then number of event dimensions this bijector operates on.
- - -
#### `tf.contrib.distributions.bijector.Invert.forward(x, name='forward')` {#Invert.forward}
Returns the forward `Bijector` evaluation, i.e., X = g(Y).
##### Args:
* <b>`x`</b>: `Tensor`. The input to the "forward" evaluation.
* <b>`name`</b>: The name to give this op.
##### Returns:
`Tensor`.
##### Raises:
* <b>`TypeError`</b>: if `self.dtype` is specified and `x.dtype` is not
`self.dtype`.
* <b>`NotImplementedError`</b>: if `_forward` is not implemented.
- - -
#### `tf.contrib.distributions.bijector.Invert.forward_event_shape(input_shape)` {#Invert.forward_event_shape}
Shape of a single sample from a single batch as a `TensorShape`.
Same meaning as `forward_event_shape_tensor`. May be only partially defined.
##### Args:
* <b>`input_shape`</b>: `TensorShape` indicating event-portion shape passed into
`forward` function.
##### Returns:
* <b>`forward_event_shape_tensor`</b>: `TensorShape` indicating event-portion shape
after applying `forward`. Possibly unknown.
- - -
#### `tf.contrib.distributions.bijector.Invert.forward_event_shape_tensor(input_shape, name='forward_event_shape_tensor')` {#Invert.forward_event_shape_tensor}
Shape of a single sample from a single batch as an `int32` 1D `Tensor`.
##### Args:
* <b>`input_shape`</b>: `Tensor`, `int32` vector indicating event-portion shape
passed into `forward` function.
* <b>`name`</b>: name to give to the op
##### Returns:
* <b>`forward_event_shape_tensor`</b>: `Tensor`, `int32` vector indicating
event-portion shape after applying `forward`.
- - -
#### `tf.contrib.distributions.bijector.Invert.forward_log_det_jacobian(x, name='forward_log_det_jacobian')` {#Invert.forward_log_det_jacobian}
Returns both the forward_log_det_jacobian.
##### Args:
* <b>`x`</b>: `Tensor`. The input to the "forward" Jacobian evaluation.
* <b>`name`</b>: The name to give this op.
##### Returns:
`Tensor`.
##### Raises:
* <b>`TypeError`</b>: if `self.dtype` is specified and `y.dtype` is not
`self.dtype`.
* <b>`NotImplementedError`</b>: if neither `_forward_log_det_jacobian`
nor {`_inverse`, `_inverse_log_det_jacobian`} are implemented.
- - -
#### `tf.contrib.distributions.bijector.Invert.graph_parents` {#Invert.graph_parents}
Returns this `Bijector`'s graph_parents as a Python list.
- - -
#### `tf.contrib.distributions.bijector.Invert.inverse(y, name='inverse')` {#Invert.inverse}
Returns the inverse `Bijector` evaluation, i.e., X = g^{-1}(Y).
##### Args:
* <b>`y`</b>: `Tensor`. The input to the "inverse" evaluation.
* <b>`name`</b>: The name to give this op.
##### Returns:
`Tensor`.
##### Raises:
* <b>`TypeError`</b>: if `self.dtype` is specified and `y.dtype` is not
`self.dtype`.
* <b>`NotImplementedError`</b>: if neither `_inverse` nor
`_inverse_and_inverse_log_det_jacobian` are implemented.
- - -
#### `tf.contrib.distributions.bijector.Invert.inverse_and_inverse_log_det_jacobian(y, name='inverse_and_inverse_log_det_jacobian')` {#Invert.inverse_and_inverse_log_det_jacobian}
Returns both the inverse evaluation and inverse_log_det_jacobian.
Enables possibly more efficient calculation when both inverse and
corresponding Jacobian are needed.
See `inverse()`, `inverse_log_det_jacobian()` for more details.
##### Args:
* <b>`y`</b>: `Tensor`. The input to the "inverse" Jacobian evaluation.
* <b>`name`</b>: The name to give this op.
##### Returns:
`Tensor`.
##### Raises:
* <b>`TypeError`</b>: if `self.dtype` is specified and `y.dtype` is not
`self.dtype`.
* <b>`NotImplementedError`</b>: if neither `_inverse_and_inverse_log_det_jacobian`
nor {`_inverse`, `_inverse_log_det_jacobian`} are implemented.
- - -
#### `tf.contrib.distributions.bijector.Invert.inverse_event_shape(output_shape)` {#Invert.inverse_event_shape}
Shape of a single sample from a single batch as a `TensorShape`.
Same meaning as `inverse_event_shape_tensor`. May be only partially defined.
##### Args:
* <b>`output_shape`</b>: `TensorShape` indicating event-portion shape passed into
`inverse` function.
##### Returns:
* <b>`inverse_event_shape_tensor`</b>: `TensorShape` indicating event-portion shape
after applying `inverse`. Possibly unknown.
- - -
#### `tf.contrib.distributions.bijector.Invert.inverse_event_shape_tensor(output_shape, name='inverse_event_shape_tensor')` {#Invert.inverse_event_shape_tensor}
Shape of a single sample from a single batch as an `int32` 1D `Tensor`.
##### Args:
* <b>`output_shape`</b>: `Tensor`, `int32` vector indicating event-portion shape
passed into `inverse` function.
* <b>`name`</b>: name to give to the op
##### Returns:
* <b>`inverse_event_shape_tensor`</b>: `Tensor`, `int32` vector indicating
event-portion shape after applying `inverse`.
- - -
#### `tf.contrib.distributions.bijector.Invert.inverse_log_det_jacobian(y, name='inverse_log_det_jacobian')` {#Invert.inverse_log_det_jacobian}
Returns the (log o det o Jacobian o inverse)(y).
Mathematically, returns: `log(det(dX/dY))(Y)`. (Recall that: `X=g^{-1}(Y)`.)
Note that `forward_log_det_jacobian` is the negative of this function.
##### Args:
* <b>`y`</b>: `Tensor`. The input to the "inverse" Jacobian evaluation.
* <b>`name`</b>: The name to give this op.
##### Returns:
`Tensor`.
##### Raises:
* <b>`TypeError`</b>: if `self.dtype` is specified and `y.dtype` is not
`self.dtype`.
* <b>`NotImplementedError`</b>: if neither `_inverse_log_det_jacobian` nor
`_inverse_and_inverse_log_det_jacobian` are implemented.
- - -
#### `tf.contrib.distributions.bijector.Invert.is_constant_jacobian` {#Invert.is_constant_jacobian}
Returns true iff the Jacobian is not a function of x.
Note: Jacobian is either constant for both forward and inverse or neither.
##### Returns:
* <b>`is_constant_jacobian`</b>: Python `bool`.
- - -
#### `tf.contrib.distributions.bijector.Invert.name` {#Invert.name}
Returns the string name of this `Bijector`.
- - -
#### `tf.contrib.distributions.bijector.Invert.validate_args` {#Invert.validate_args}
Returns True if Tensor arguments will be validated.
|
