Core

plane_wave classmethod

plane_wave(cls, simulation_parameters: SimulationParameters, distance: float = 0.0, wave_direction: Any = None, initial_phase: float = 0.0) -> Self

Generate wavefront of the plane wave

gaussian_beam classmethod

gaussian_beam(cls, simulation_parameters: SimulationParameters, waist_radius: float, distance: float = 0.0, dx: float = 0.0, dy: float = 0.0) -> Self

Generates the Gaussian beam.

spherical_wave classmethod

spherical_wave(cls, simulation_parameters: SimulationParameters, distance: float, initial_phase: float = 0.0, dx: float = 0.0, dy: float = 0.0) -> Self

Generate wavefront of the spherical wave

fwhm

fwhm(self, simulation_parameters: SimulationParameters) -> tuple[float, float]

Calculates full width at half maximum of the wavefront

__init__ constructor

__init__(self, axes: dict[str, torch.Tensor]) -> None

index

index(self, name: str) -> int

Index of specific axis in the tensor. The index is negative.

__init__ constructor

__init__(self, axes: dict[str, torch.Tensor | float]) -> None

meshgrid

meshgrid(self, x_axis: str, y_axis: str)

Returns a meshgrid for a selected pair of axes. …

axes_size

axes_size(self, axs: Iterable[str]) -> torch.Size

Returns a size of axes in specified order.

to

to(self, device: str | torch.device | int) -> 'SimulationParameters'

__init__ constructor

__init__(self, data: Any, requires_grad: bool = True)

__init__ constructor

__init__(self, data: Any, min_value: Any, max_value: Any, bound_func: Callable[[torch.Tensor], torch.Tensor] = torch.sigmoid, inv_bound_func: Callable[[torch.Tensor], torch.Tensor] = sigmoid_inv, requires_grad: bool = True)

__init__ constructor

__init__(self, elements: Iterable[Element]) -> None

forward

forward(self, input_wavefront: Tensor) -> Tensor

A forward function for a network assembled from elements.

stepwise_forward

stepwise_forward(self, input_wavefront: Tensor)

Function that consistently applies forward method of each element to an input wavefront.

reverse

reverse(self, Ein: Tensor) -> Tensor

to_specs

to_specs(self) -> Iterable[ParameterSpecs | SubelementSpecs]

__init__ constructor

__init__(self, experiment_directory: str) -> None

A lightweight alternative to TensorBoard and other logging frameworks for tracking the training process, storing experiment metadata, and handling checkpoints.

The Clerk is not a new concept but a minimal implementation included in the framework to start training models without any dependencies.

save_conditions

save_conditions(self, conditions: ConditionsType)

Save the experiment conditions. The conditions are stored in the experiment directory in the file conditions.json.

load_conditions

load_conditions(self) -> ConditionsType

Read and return the experiment conditions.

set_checkpoint_targets

set_checkpoint_targets(self, targets: dict[str, StatefulTorchClass])

Set the targets with state dict to be saved in the checkpoint. The same targets will be automatically used when loading from the checkpoint.

begin

begin(self, resume: bool = False, resume_load_last_checkpoint: bool = True, autosave_checkpoint: bool = False) -> Self

Configure the clerk for a new context.

write_log

write_log(self, tag: str, data: dict, flush: bool = False)

Write new data to a log file with a specific tag. The data should be a dictionary that can be serialized into JSON.

write_checkpoint

write_checkpoint(self, metadata: object | None = None)

Write the states of selected targets into a new checkpoint file.

load_checkpoint

load_checkpoint(self, index: str | int, targets: dict[str, StatefulTorchClass] | None = None, weights_only: bool = True) -> object | None

Load the checkpoint with a specific index and apply state dicts to checkpoint targets. If the targets are not provided, the checkpoint targets are obtained from the clerk settings.

If the index is integer, that the checkpoint <index>.pt is used. Otherwise, index is used as a filename.

load_logs

load_logs(self, tag: str) -> Generator[dict, None, None]

Load logs from the file specified by the tag.

Yields

• Generator[dict, None, None] — Log data with the specific tag.

load_logs_to_pandas

load_logs_to_pandas(self, tag: str) -> 'pandas.DataFrame'

Load logs from the file specified by the tag and return them as a pandas DataFrame.

clean_checkpoints

clean_checkpoints(self)

Remove checkpoints that are not listed in checkpoints.txt. If checkpoints.txt does not exist, then remove all <n>.pt files, where <n> is an integer.

clean_backup_checkpoints

clean_backup_checkpoints(self)

Remove checkpoints that are matches backup checkpoints name pattern.

__init__ constructor

__init__(self, simulation_parameters: SimulationParameters, func = 'intensity')

forward

forward(self, input_field: Wavefront) -> torch.Tensor

Method that returns the image obtained from the incident field by a detector using self.func. in the simplest case the image on a detector is an intensities image) …