sigmoid support

[WandernForte]

i'm trying to modify foolbox like following,

from typing import TypeVar, Any
from abc import ABC, abstractmethod

import torch
import eagerpy as ep
from foolbox.criteria import Criterion

class TargetedMisclassificationML(Criterion):
    """Considers those perturbed inputs adversarial whose predicted class
    matches the target classes. Multi-Label

    Args:
        target_classes: Tensor with target classes ``(batch,)``.
    """

    def __init__(self, target_classes: Any):
        super().__init__()
        self.target_classes: ep.Tensor = ep.astensor(target_classes)

    def __repr__(self) -> str:
        return f"{self.__class__.__name__}({self.target_classes!r})"

    def __call__(self, perturbed: T, outputs: T) -> T:
        outputs_, restore_type = ep.astensor_(outputs)
        del perturbed, outputs
        print(type(outputs_), type(outputs))
        # classes = outputs_.argmax(axis=-1)
        classes = torch.tensor(torch.round(torch.sigmoid(outputs_)).detach().numpy().tolist())  # ERR
        assert classes.shape == self.target_classes.shape
        is_adv = classes == self.target_classes
        return restore_type(is_adv)

will u support sigmoid function in the future? and how to implement sigmoid with eagerpy now?

[WandernForte]

i've solve it by self.

class TargetedMisclassificationML(Criterion):
    """Considers those perturbed inputs adversarial whose predicted class
    matches the target classes. Multi-Label

    Args:
        target_classes: Tensor with target classes ``(batch,)``.
    """

    def __init__(self, target_classes: Any):
        super().__init__()
        self.target_classes: ep.Tensor = ep.astensor(target_classes)
        _, self.restore_type = ep.astensor_(target_classes)
    def __repr__(self) -> str:
        return f"{self.__class__.__name__}({self.target_classes!r})"

    def __call__(self, perturbed: T, outputs: T) -> T:
        outputs_, restore_type = ep.astensor_(outputs)
        del perturbed, outputs
        classes = torch.tensor(torch.round(torch.sigmoid(self.restore_type(outputs_).to(device))).detach().numpy().tolist()).to(device)
        classes = ep.astensor(classes)
        assert classes.shape == self.target_classes.shape
        is_adv = classes == self.target_classes
        return restore_type(is_adv)