spaces.DataAxis

spaces.DataAxis(values)

Axis for sampling from predefined data values.

Uses a fixed set of values provided by the user. This axis type is deterministic and always returns the same predefined values. The first dimension of the data will always be used as axis dimension.

Parameters

Name	Type	Description	Default
values	array - like	Predefined values to sample from.	required

Raises

Name	Type	Description
	ValueError	If values array is empty.

Examples

>>> import jax.numpy as jnp
>>> data = DataAxis([1.0, 2.5, 3.7, 4.2])
>>> values = data.generate_values()
>>> print(values)  # [1.0, 2.5, 3.7, 4.2]
>>>
>>> # Can also use JAX arrays
>>> data_jax = DataAxis(jnp.linspace(0, 1, 5))
>>> print(data_jax.size)  # 5

Attributes

Name	Description
size	Number of predefined values.

Methods

Name	Description
generate_values	Return the predefined data values.

generate_values

spaces.DataAxis.generate_values(key=None)

Return the predefined data values.

Parameters

Name	Type	Description	Default
key	jax.random.PRNGKey	Random key. Ignored for deterministic data sampling.	`None`

Returns

Name	Type	Description
	jnp.ndarray	The predefined data values.

# spaces.DataAxis { #tvboptim.spaces.DataAxis } ```python spaces.DataAxis(values) ``` Axis for sampling from predefined data values. Uses a fixed set of values provided by the user. This axis type is deterministic and always returns the same predefined values. The first dimension of the data will always be used as axis dimension. ## Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |--------|-----------------------------------|-----------------------------------|------------| | values | [array](`array`) - [like](`like`) | Predefined values to sample from. | _required_ | ## Raises {.doc-section .doc-section-raises} | Name | Type | Description | |--------|----------------------------|---------------------------| | | [ValueError](`ValueError`) | If values array is empty. | ## Examples {.doc-section .doc-section-examples} ```python >>> import jax.numpy as jnp >>> data = DataAxis([1.0, 2.5, 3.7, 4.2]) >>> values = data.generate_values() >>> print(values) # [1.0, 2.5, 3.7, 4.2] >>> >>> # Can also use JAX arrays >>> data_jax = DataAxis(jnp.linspace(0, 1, 5)) >>> print(data_jax.size) # 5 ``` ## Attributes | Name | Description | | --- | --- | | [size](#tvboptim.spaces.DataAxis.size) | Number of predefined values. | ## Methods | Name | Description | | --- | --- | | [generate_values](#tvboptim.spaces.DataAxis.generate_values) | Return the predefined data values. | ### generate_values { #tvboptim.spaces.DataAxis.generate_values } ```python spaces.DataAxis.generate_values(key=None) ``` Return the predefined data values. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |--------|---------------------------------------------------------------------|------------------------------------------------------|-----------| | key | [jax](`jax`).[random](`jax.random`).[PRNGKey](`jax.random.PRNGKey`) | Random key. Ignored for deterministic data sampling. | `None` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|---------------------------------------------------|-----------------------------| | | [jnp](`jax.numpy`).[ndarray](`jax.numpy.ndarray`) | The predefined data values. |