captum/attr/_core/integrated_gradients.py [317:347]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        if step_sizes_and_alphas is None:
            # retrieve step size and scaling factor for specified
            # approximation method
            step_sizes_func, alphas_func = approximation_parameters(method)
            step_sizes, alphas = step_sizes_func(n_steps), alphas_func(n_steps)
        else:
            step_sizes, alphas = step_sizes_and_alphas

        # scale features and compute gradients. (batch size is abbreviated as bsz)
        # scaled_features' dim -> (bsz * #steps x inputs[0].shape[1:], ...)
        scaled_features_tpl = tuple(
            torch.cat(
                [baseline + alpha * (input - baseline) for alpha in alphas], dim=0
            ).requires_grad_()
            for input, baseline in zip(inputs, baselines)
        )

        additional_forward_args = _format_additional_forward_args(
            additional_forward_args
        )
        # apply number of steps to additional forward args
        # currently, number of steps is applied only to additional forward arguments
        # that are nd-tensors. It is assumed that the first dimension is
        # the number of batches.
        # dim -> (bsz * #steps x additional_forward_args[0].shape[1:], ...)
        input_additional_args = (
            _expand_additional_forward_args(additional_forward_args, n_steps)
            if additional_forward_args is not None
            else None
        )
        expanded_target = _expand_target(target, n_steps)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



captum/attr/_core/layer/internal_influence.py [254:282]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        if step_sizes_and_alphas is None:
            # retrieve step size and scaling factor for specified approximation method
            step_sizes_func, alphas_func = approximation_parameters(method)
            step_sizes, alphas = step_sizes_func(n_steps), alphas_func(n_steps)
        else:
            step_sizes, alphas = step_sizes_and_alphas

        # Compute scaled inputs from baseline to final input.
        scaled_features_tpl = tuple(
            torch.cat(
                [baseline + alpha * (input - baseline) for alpha in alphas], dim=0
            ).requires_grad_()
            for input, baseline in zip(inputs, baselines)
        )

        additional_forward_args = _format_additional_forward_args(
            additional_forward_args
        )
        # apply number of steps to additional forward args
        # currently, number of steps is applied only to additional forward arguments
        # that are nd-tensors. It is assumed that the first dimension is
        # the number of batches.
        # dim -> (bsz * #steps x additional_forward_args[0].shape[1:], ...)
        input_additional_args = (
            _expand_additional_forward_args(additional_forward_args, n_steps)
            if additional_forward_args is not None
            else None
        )
        expanded_target = _expand_target(target, n_steps)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -