#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/utils/OpAdapter.h"
namespace acl_op {
using npu_preparation = at_npu::native::OpPreparation;
using npu_utils = at_npu::native::NpuUtils;
namespace {
at::Tensor &smooth_l1_loss_out_npu_nocheck(at::Tensor &result, const at::Tensor &self, const at::Tensor &target,
int64_t reduction, double beta)
{
if (self.numel() == 0) {
result = at_npu::native::custom_ops::_npu_dtype_cast(result, at::kFloat).fill_(NAN);
return result;
}
string reduction_str(op_plugin::utils::get_reduction_str(reduction));
at_npu::native::OpCommand cmd;
cmd.Name("SmoothL1LossV2")
.Input(self)
.Input(target)
.Output(result)
.Attr("reduction", reduction_str)
.Attr("sigma", static_cast<float>(beta))
.Run();
return result;
}
}
at::Tensor &smooth_l1_loss_out(const at::Tensor &self, const at::Tensor &target, int64_t reduction, double beta,
at::Tensor &out)
{
auto output_size = op_infer::smooth_l1_loss_npu_output_size(self, reduction);
npu_preparation::CheckOut({self, target}, out, npu_preparation::get_tensor_npu_format(self), self.scalar_type(),
output_size);
if (!npu_utils::check_match(&out)) {
at::Tensor contiguous_result = npu_utils::format_contiguous(out);
smooth_l1_loss_out_npu_nocheck(contiguous_result, self, target, reduction, beta);
npu_utils::format_fresh_view(out, contiguous_result);
} else {
smooth_l1_loss_out_npu_nocheck(out, self, target, reduction, beta);
}
return out;
}
at::Tensor smooth_l1_loss(const at::Tensor &self, const at::Tensor &target, int64_t reduction, double beta)
{
auto output_size = op_infer::smooth_l1_loss_npu_output_size(self, reduction);
at::Tensor result = npu_preparation::apply_tensor(self, output_size);
smooth_l1_loss_out_npu_nocheck(result, self, target, reduction, beta);
return result;
}
}