cloak_one_shot

Classes:

Name	Description
CloakNoiseLayerOneShot	Applies an input independent stochastic transformation to a Tensor using ParameterWrapper,

CloakNoiseLayerOneShot

Bases: CloakNoiseLayer

Applies an input independent stochastic transformation to a Tensor using ParameterWrapper, with standard deviations parameterized by CloakStandardDeviationParameterization, and optional standard deviation-based input masking using PercentMasker.

The original Cloak Algorithm requires two steps to train. In the first step, the stochastic mapping is trained, and in the second step, masking is trained. This class allows for training in one step, by recalculating the threshold every forward pass.

For most use cases, this class is preferred over a multi-step Stained Glass Transform version, since it is simpler to use. For more advanced use cases, however, the multi-step version may be required, since it allows for more control over the masking process during training.

Parameters:

Name	Type	Description	Default
scale	tuple[float, float]	Used to set bound on the min and max standard deviation of the generated stochastic transformation.	required
percent_to_mask	float	The percentage of the outputs to mask.	required
shallow	float	A temperature-like parameter which controls the spread of the parameterization function. Controls both the magnitude of parameterized standard deviations and their rate of change with respect to rhos.	1.0
rhos_init	float	The initial values for the rhos.	-4.0
seed	int | None	Seed for the random number generator used to generate the stochastic transformation. If None, the global RNG state is used.	None

Notes

Specifying percent_to_mask==0.0 is functionally equivalent to Cloak Step 1, as no masking occurs.

Raises:

Type	Description
ValueError	If percent_to_mask is not specified.
ValueError	If percent_to_mask is not in [0, 1], inclusive.

Warning

The directly learned locs and rhos used by this class are prone to vanishing when trained with weight decay regularization. To avoid this, ensure that optimizer parameter group(s) containing locs or rhos are configured with weight_decay=0.0.

Examples:

>>> import torch
>>> noise_layer = CloakNoiseLayerOneShot(percent_to_mask=0.5, scale=(1e-4, 2.0))
>>> img = torch.rand((1, 3, 8, 8))
>>> transformed_img = noise_layer(img)
>>> transformed_img.shape == img.shape
True
>>> torch.allclose(transformed_img, img)
False

See paper: Not All Features Are Equal: Discovering Essential Features for Preserving Prediction Privacy.

Methods:

Name	Description
__call__	Transform the input data.
__getstate__	Prepare a serializable copy of self.__dict__.
__init__
__init_subclass__	Set the default dtype to torch.float32 inside all subclass __init__ methods.
__setstate__	Restore from a serialized copy of self.__dict__.
forward	Transform the input data.
get_applied_transform_components_factory	Create a function that returns the elements of the transform components ('mean' and 'std') applied during the most recent
get_transformed_output_factory	Create a function that returns the transformed output from the most recent forward pass.
initial_seed	Return the initial seed of the CPU device's random number generator.
manual_seed	Seed each of the random number generators.
reset_parameters	Reinitialize parameters and buffers.
seed	Seed each of the random number generators using a non-deterministic random number.

__call__

__call__(
    input: Tensor,
    noise_mask: Tensor | None = None,
    **kwargs: Any,
) -> torch.Tensor

Transform the input data.

Parameters:

Name	Type	Description	Default
input	Tensor	The input to transform.	required
noise_mask	Tensor | None	An optional mask that selects the elements of input to transform. Where the mask is False, the original input value is returned. Also used to select the elements of the sampled standard deviations to use to mask the input. If None, the entire input is transformed.	None
**kwargs	Any	Additional keyword arguments to the estimator modules.	required

__getstate__

__getstate__() -> dict[str, Any]

Prepare a serializable copy of self.__dict__.

__init__

__init__(
    percent_to_mask: float,
    scale: tuple[float, float],
    shallow: float = 1.0,
    rhos_init: float = -4.0,
    seed: int | None = None,
) -> None

Changed in version 0.10.0: `threshold` and `percent_threshold` parameters were removed in favor of `percent_to_mask`

__init_subclass__

__init_subclass__() -> None

Set the default dtype to torch.float32 inside all subclass __init__ methods.

__setstate__

__setstate__(state: dict[str, Any]) -> None

Restore from a serialized copy of self.__dict__.

forward

forward(
    input: Tensor,
    noise_mask: Tensor | None = None,
    **kwargs: Any,
) -> torch.Tensor

Transform the input data.

Parameters:

Name	Type	Description	Default
input	Tensor	The input to transform.	required
noise_mask	Tensor | None	An optional mask that selects the elements of input to transform. Where the mask is 0, the original input value is returned. Also used to select the elements of the sampled standard deviations to use to mask the input. If None, the entire input is transformed.	None
**kwargs	Any	Additional keyword arguments to the estimator modules.	required

Returns:

Type	Description
torch.Tensor	The transformed input data.

get_applied_transform_components_factory

get_applied_transform_components_factory() -> Callable[
    [], dict[str, torch.Tensor]
]

Create a function that returns the elements of the transform components ('mean' and 'std') applied during the most recent forward pass.

Specifically, the applied elements are those selected by the noise mask (if supplied) and standard deviation mask (if std_estimator.masker is not None). If no masks are used, all elements are returned.

The applied transform components are returned flattened.

This function is intended to be used to log histograms of the transform components.

Returns:

Type	Description
Callable[[], dict[str, torch.Tensor]]	A function that returns the the elements of the transform components applied during the most recent forward pass.

Examples:

>>> from torch import nn
>>> from stainedglass_core import model as sg_model, noise_layer as sg_noise_layer
>>> base_model = nn.Linear(20, 2)
>>> noisy_model = sg_model.NoisyModel(
...     sg_noise_layer.CloakNoiseLayer1,
...     base_model,
...     target_parameter="input",
... )
>>> get_applied_transform_components = (
...     noisy_model.noise_layer.get_applied_transform_components_factory()
... )
>>> input = torch.ones(1, 20)
>>> noise_mask = torch.tensor(5 * [False] + 15 * [True])
>>> output = noisy_model(input, noise_mask=noise_mask)
>>> applied_transform_components = get_applied_transform_components()
>>> applied_transform_components
{'mean': tensor(...), 'std': tensor(...)}
>>> {
...     component_name: component.shape
...     for component_name, component in applied_transform_components.items()
... }
{'mean': torch.Size([15]), 'std': torch.Size([15])}

get_transformed_output_factory

get_transformed_output_factory() -> Callable[
    [], torch.Tensor
]

Create a function that returns the transformed output from the most recent forward pass.

If super batching is active, only the transformed half of the super batch output is returned.

Returns:

Type	Description
Callable[[], torch.Tensor]	A function that returns the transformed output from the most recent forward pass.

Examples:

>>> from stainedglass_core import noise_layer as sg_noise_layer
>>> noise_layer = sg_noise_layer.CloakNoiseLayer1()
>>> get_transformed_output = noise_layer.get_transformed_output_factory()
>>> input = torch.ones(2, 3, 32, 32)
>>> output = noise_layer(input)
>>> transformed_output = get_transformed_output()
>>> assert output.equal(transformed_output)

initial_seed

initial_seed() -> int

Return the initial seed of the CPU device's random number generator.

manual_seed

manual_seed(seed: int | None) -> None

Seed each of the random number generators.

Setting seed to None will destroy any existing generators.

Parameters:

Name	Type	Description	Default
seed	int | None	The seed to set.	required

reset_parameters

reset_parameters() -> None

Reinitialize parameters and buffers.

This method is useful for initializing tensors created on the meta device.

seed

seed() -> None

Seed each of the random number generators using a non-deterministic random number.