capspeech/eval/src/model/revgrad_func.py

from torch.autograd import Function


class RevGrad(Function):
    @staticmethod
    def forward(ctx, input_, alpha_):
        ctx.save_for_backward(input_, alpha_)
        output = input_
        return output

    @staticmethod
    def backward(ctx, grad_output):  # pragma: no cover
        grad_input = None
        _, alpha_ = ctx.saved_tensors
        if ctx.needs_input_grad[0]:
            grad_input = -grad_output * alpha_
        return grad_input, None


revgrad = RevGrad.apply