#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/OpApiInterface.h"
#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/utils/op_api_common.h"
#include "torch_npu/csrc/framework/utils/UtilForOpAdapter.h"
namespace op_api {
at::Tensor& silu_backward_out(const at::Tensor& grad_output, const at::Tensor& self, at::Tensor& result)
{
DO_COMPATIBILITY(aclnnSiluBackward, acl_op::silu_backward_out(grad_output, self, result));
if (c10_npu::GetSocVersion() < c10_npu::SocVersion::Ascend950) {
at_npu::native::OpPreparation::check_tensor({grad_output, self}, result, grad_output);
}
EXEC_NPU_CMD(aclnnSiluBackward, grad_output, self, result);
return result;
}
at::Tensor silu_backward(const at::Tensor& grad_output, const at::Tensor& self)
{
DO_COMPATIBILITY(aclnnSiluBackward, acl_op::silu_backward(grad_output, self));
at::Tensor grad_input;
if (c10_npu::GetSocVersion() < c10_npu::SocVersion::Ascend950) {
grad_input = at_npu::native::OpPreparation::apply_tensor_without_format(grad_output);
} else {
at::ScalarType output_dtype = grad_output.scalar_type();
if (grad_output.scalar_type() != self.scalar_type()) {
output_dtype = at::kFloat;
}
auto output_size = op_infer::broadcast_ops_npu_output_size(grad_output, self);
grad_input = at_npu::native::OpPreparation::apply_tensor_without_format(output_size, self.options().dtype(output_dtype));
}
EXEC_NPU_CMD(aclnnSiluBackward, grad_output, self, grad_input);
return grad_input;
}
}