Skip to content

No consistency in generated saliencies #3

@nicenoize

Description

@nicenoize

Hello! I noticed, that the values of the generated saliencies are quite different after generating them multiple times.

While trying to find the origin of this issue i noticed:

  • in line 111 of interpret_grads_occ.py: executing model(batch[0], attention_mask=batch[1], labels=batch[2].long()) returns a different tensor after each call. I was expecting to receive the same results
  • line 157: ablator.attribute(input_embeddings, target=cls_, additional_forward_args=additional) also returns different tensors with each call

I hope you can help me clarify this

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions