MultiScalePrior

class jolideco.priors.MultiScalePrior(prior, n_levels=2, weights=None, cycle_spin=True, anti_alias=True)

Bases: Prior

Multiscale prior

Apply a given prior per resolution level and sum up the log likelihood contributions across all resolution levels.

Parameters:

• prior (Prior) – Prior instance

• n_levels (int, optional) – Number of multiscale levels

• weights (list of floats) – Weight to be applied per level.

• cycle_spin (bool) – Apply cycle spin.

• anti_alias (bool) – Apply Gaussian smoothing before downsampling.

Initialize internal Module state, shared by both nn.Module and ScriptModule.

Attributes Summary

weights

Weights

Methods Summary

__call__(flux)	Evaluate the prior
to_dict()	Convert to dict

Attributes Documentation

weights

Weights

Methods Documentation

__call__(flux)

Evaluate the prior

Parameters:

flux (~pytorch.Tensor) – Reconstructed flux

Returns:

log_prior – Summed log prior over all overlapping patches.

Return type:

float

to_dict()

Convert to dict