vision_unlearning.utils.parameter_attribution

Attributes

logger

Classes

ParameterAttributionMethod

!!! abstract "Usage Documentation"

ParameterAttributionMethodSaliency

!!! abstract "Usage Documentation"

Module Contents

vision_unlearning.utils.parameter_attribution.logger
class vision_unlearning.utils.parameter_attribution.ParameterAttributionMethod(/, **data: Any)[source]

Bases: pydantic.BaseModel, abc.ABC

!!! abstract “Usage Documentation”

[Models](../concepts/models.md)

A base class for creating Pydantic models.

__class_vars__

The names of the class variables defined on the model.

__private_attributes__

Metadata about the private attributes of the model.

__signature__

The synthesized __init__ [Signature][inspect.Signature] of the model.

__pydantic_complete__

Whether model building is completed, or if there are still undefined fields.

__pydantic_core_schema__

The core schema of the model.

__pydantic_custom_init__

Whether the model has a custom __init__ function.

__pydantic_decorators__

Metadata containing the decorators defined on the model. This replaces Model.__validators__ and Model.__root_validators__ from Pydantic V1.

__pydantic_generic_metadata__

A dictionary containing metadata about generic Pydantic models. The origin and args items map to the [__origin__][genericalias.__origin__] and [__args__][genericalias.__args__] attributes of [generic aliases][types-genericalias], and the parameter item maps to the __parameter__ attribute of generic classes.

__pydantic_parent_namespace__

Parent namespace of the model, used for automatic rebuilding of models.

__pydantic_post_init__

The name of the post-init method for the model, if defined.

__pydantic_root_model__

Whether the model is a [RootModel][pydantic.root_model.RootModel].

__pydantic_serializer__

The pydantic-core SchemaSerializer used to dump instances of the model.

__pydantic_validator__

The pydantic-core SchemaValidator used to validate instances of the model.

__pydantic_fields__

A dictionary of field names and their corresponding [FieldInfo][pydantic.fields.FieldInfo] objects.

__pydantic_computed_fields__

A dictionary of computed field names and their corresponding [ComputedFieldInfo][pydantic.fields.ComputedFieldInfo] objects.

__pydantic_extra__

A dictionary containing extra values, if [extra][pydantic.config.ConfigDict.extra] is set to ‘allow’.

__pydantic_fields_set__

The names of fields explicitly set during instantiation.

__pydantic_private__

Values of private attributes set on the model instance.

abstract attribute(noise_scheduler: Any, text_encoder: Any, vae: Any, unet: Any, dataloader: torch.utils.data.DataLoader, device: str, weight_dtype: torch.dtype) Dict[str, torch.Tensor][source]
class vision_unlearning.utils.parameter_attribution.ParameterAttributionMethodSaliency(/, **data: Any)[source]

Bases: ParameterAttributionMethod

!!! abstract “Usage Documentation”

[Models](../concepts/models.md)

A base class for creating Pydantic models.

__class_vars__

The names of the class variables defined on the model.

__private_attributes__

Metadata about the private attributes of the model.

__signature__

The synthesized __init__ [Signature][inspect.Signature] of the model.

__pydantic_complete__

Whether model building is completed, or if there are still undefined fields.

__pydantic_core_schema__

The core schema of the model.

__pydantic_custom_init__

Whether the model has a custom __init__ function.

__pydantic_decorators__

Metadata containing the decorators defined on the model. This replaces Model.__validators__ and Model.__root_validators__ from Pydantic V1.

__pydantic_generic_metadata__

A dictionary containing metadata about generic Pydantic models. The origin and args items map to the [__origin__][genericalias.__origin__] and [__args__][genericalias.__args__] attributes of [generic aliases][types-genericalias], and the parameter item maps to the __parameter__ attribute of generic classes.

__pydantic_parent_namespace__

Parent namespace of the model, used for automatic rebuilding of models.

__pydantic_post_init__

The name of the post-init method for the model, if defined.

__pydantic_root_model__

Whether the model is a [RootModel][pydantic.root_model.RootModel].

__pydantic_serializer__

The pydantic-core SchemaSerializer used to dump instances of the model.

__pydantic_validator__

The pydantic-core SchemaValidator used to validate instances of the model.

__pydantic_fields__

A dictionary of field names and their corresponding [FieldInfo][pydantic.fields.FieldInfo] objects.

__pydantic_computed_fields__

A dictionary of computed field names and their corresponding [ComputedFieldInfo][pydantic.fields.ComputedFieldInfo] objects.

__pydantic_extra__

A dictionary containing extra values, if [extra][pydantic.config.ConfigDict.extra] is set to ‘allow’.

__pydantic_fields_set__

The names of fields explicitly set during instantiation.

__pydantic_private__

Values of private attributes set on the model instance.

attribute(noise_scheduler: Any, text_encoder: Any, vae: Any, unet: Any, dataloader: torch.utils.data.DataLoader, device: str, weight_dtype: torch.dtype) Dict[str, torch.Tensor][source]

@return saliency: keys like “down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_v.weight”, values are tensors of same shape as the parameter, containing the accumulated saliency values. Tensor are of type torch.float32.

Expected characteristics of the dataloader: * Batch size and number of workers are set according to the training arguments. * shuffled * collate behavior: Batches are created by stacking the per-example tensors (pixel_values stacked into contiguous FloatTensor) * Fields

  • pixel_values: preprocessed images, ready to be fed to the vae (i.e. resized, cropped, normalized…). Shape=[batch size, 3, resolution, resolution]

  • input_ids: tokenized captions, ready to be fed to the text encoder. Shape=[batch size, sequence length]

Expected characteristics of the model (scheduler, text encoder, vae, unet): * unet should have anabled gradients * All loaded in the same device (the one specified in the arguments) * All loaded in the same dtype (the one specified in the arguments) * Loadede from the same base model / working together coherently