expressions.py¶
pycmtensor.expressions
¶
The code snippet is a part of the PyCMTensor expressions module. It defines a base class for parsing and manipulating Aesara tensor expressions. The class provides methods for parsing a tensor expression to remove parentheses and tensor operators, and returns a clean list of keywords found in the expression. It also defines a base class for expression objects, which includes overloaded operators for tensor operations such as addition, subtraction, multiplication, division, and comparison.
ExpressionParser(expression=None)
¶
Bases: object
Base class for parsing and manipulating Aesara tensor expressions.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
expression |
TensorVariable
|
The tensor expression to parse. Defaults to None. |
None
|
parse(expression)
staticmethod
¶
Parses a tensor expression to remove parentheses and tensor operators.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
expression |
TensorVariable
|
The symbolic tensor object to parse. |
required |
Returns:
| Name | Type | Description |
|---|---|---|
list |
The clean list of keywords found in the expression. |
TensorExpressions()
¶
Base class for expression objects
Param(name, value=0.0, lb=None, ub=None, status=0)
¶
Bases: TensorExpressions
Constructor for model param object
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
name |
str
|
name of parameter |
required |
value |
float
|
the default value of the parameter |
0.0
|
lb |
float
|
value lower bound |
None
|
ub |
float
|
value upper bound |
None
|
status |
int
|
if 1, do not estimate this parameter |
0
|
Attributes:
| Name | Type | Description |
|---|---|---|
init_value |
float
|
the inital value set at object creation |
shape |
list
|
the shape of the Param |
Raises:
| Type | Description |
|---|---|
ValueError
|
If |
Note
init_value is an immutable property
name
property
¶
Returns the name of the object
init_value
property
¶
Returns the initial value of the parameter
status
property
¶
Returns the status of the parameter
shape
property
¶
Returns the shape of the initial value of the parameter
__call__()
¶
Returns the shared value of the parameter
__repr__()
¶
Returns a string representation of the parameter
get_value()
¶
Returns the numpy representation of the parameter value
set_value(value)
¶
Sets the value of the shared variable
reset_value()
¶
Resets the value of the shared variable to the initial value
Beta(name, value=0.0, lb=None, ub=None, status=0)
¶
Bases: Param
Constructor for Beta parameter
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
name |
str
|
The name of the Beta parameter. |
required |
value |
float
|
The default value of the Beta parameter. |
0.0
|
lb |
float
|
The lower bound of the Beta parameter. |
None
|
ub |
float
|
The upper bound of the Beta parameter. |
None
|
status |
int
|
The status of the Beta parameter. |
0
|
Example
Specifying a Beta parameter:
b_time = Beta("b_time", value=0., lb=None, ub=0., status=0)
To obtain the raw array value from the Beta parameter:
b_time = Beta("b_time", value=-1., lb=None, ub=0., status=0)
b_time.get_value()
output:
array(-1.)
RandomDraws(name, draw_type, n_draws)
¶
Bases: TensorExpressions
Constructor for the RandomDraws class.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
name |
str
|
The name of the RandomDraw object. |
required |
draw_type |
str
|
The distribution of the draw. Can be "normal", "lognormal", "gumbel", "exponential", "gamma", or "poisson". |
required |
n_draws |
int
|
The number of draws, which determines the size of the shared tensor. |
required |
Raises:
| Type | Description |
|---|---|
NotImplementedError
|
If an unsupported draw_type is provided. |
Returns:
| Type | Description |
|---|---|
|
None. The method initializes the object and creates a shared variable. |
name
property
¶
returns the name of the random draw tensor variable
Bias(name, size, value=None)
¶
Bases: Param
Class object for neural net bias vector
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
name |
str
|
The name of the parameter. |
required |
size |
Union[tuple, list]
|
The size of the array in 1 dimension. |
required |
value |
ndarray
|
The initial values of the parameter. If |
None
|
Weight(name, size, value=None, init_type=None)
¶
Bases: Param
Class object for neural net weight matrix
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
name |
str
|
name of the parameter |
required |
size |
Union[tuple, list]
|
size of the array |
required |
value |
ndarray
|
initial values of the parameter. Defaults to |
None
|
init_type |
str
|
initialization type, see notes |
None
|
Note
Initialization types are one of the following:
-
"zeros": a 2-D array of zeros -
"he": initialization method for neural networks that takes into account the non-linearity of activation functions, e.g. ReLU or Softplus [^1] -
"glorot": initialization method that maintains the variance for symmetric activation functions, e.g. sigm, tanh [^2]
[^1] He, K., Zhang, X., Ren, S. and Sun, J., 2015. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In Proceedings of the IEEE international conference on computer vision (pp. 1026-1034). [^2] Glorot, X. and Bengio, Y., 2010, March. Understanding the difficulty of training deep feedforward neural networks. In Proceedings of the thirteenth international conference on artificial intelligence and statistics (pp. 249-256). JMLR Workshop and Conference Proceedings.
Example
Specifying a weight array:
code