causica.data_generation.samplers.noise_dist_sampler¶

Module Contents¶

Classes¶

class causica.data_generation.samplers.noise_dist_sampler.NoiseModuleSampler[source]¶

Bases: causica.data_generation.samplers.sampler.Sampler[causica.distributions.noise.NoiseModule]

An interface of a univariate noise sampler

abstract sample() → causica.distributions.noise.NoiseModule[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.JointNoiseModuleSampler(noise_dist_samplers: Mapping[str, NoiseModuleSampler])[source]¶

Bases: NoiseModuleSampler

Sampler for JointNoiseModule, given shapes and types of different variables

sample() → causica.distributions.JointNoiseModule[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.UnivariateNormalNoiseModuleSampler(std_dist: torch.distributions.Distribution, dim: int = 1)[source]¶

Bases: NoiseModuleSampler

Sample a UnivariateNormalNoiseModule, with standard deviation given by a distribution.

sample()[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.UnivariateLaplaceNoiseModuleSampler(std_dist: torch.distributions.Distribution, dim: int = 1)[source]¶

Bases: NoiseModuleSampler

Sample a UnivariateLaplaceNoiseModule, with standard deviation given by a distribution.

sample()[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.UnivariateCauchyNoiseModuleSampler(std_dist: torch.distributions.Distribution, dim: int = 1)[source]¶

Bases: NoiseModuleSampler

Sample a UnivariateLaplaceNoiseModule, with standard deviation given by a distribution.

sample()[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.BernoulliNoiseModuleSampler(base_logits_dist: torch.distributions.Distribution, dim: int = 1)[source]¶

Bases: NoiseModuleSampler

Sample a BernoulliNoiseModule, with base_logits given by a distribution.

sample() → causica.distributions.noise.NoiseModule[source]¶

Sample a sample type with given shape

class causica.data_generation.samplers.noise_dist_sampler.CategoricalNoiseModuleSampler(base_logits_dist: torch.distributions.Distribution | None, num_classes: int = 2)[source]¶

Bases: NoiseModuleSampler

Sample a CategoricalNoiseModule, with num_classes classes. This does not actually sample but returns the noise.

sample() → causica.distributions.noise.NoiseModule[source]¶

Sample a sample type with given shape