#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/utils/OpAdapter.h"
namespace acl_op {
using npu_preparation = at_npu::native::OpPreparation;
using tensor_list = std::tuple<at::Tensor, at::Tensor, at::Tensor, at::Tensor>;
tensor_list npu_batch_nms(const at::Tensor &self, const at::Tensor &scores, double score_threshold,
double iou_threshold, int64_t max_size_per_class, int64_t max_total_size,
bool change_coordinate_frame, bool transpose_box)
{
TORCH_CHECK(self.numel() > 0, "The input tensor [self] is an empty tensor." + OPS_ERROR(ErrCode::PARAM));
at::Tensor nmsed_boxes = npu_preparation::apply_tensor({self.size(0), max_total_size, 4}, self.options(), self);
at::Tensor nmsed_scores = npu_preparation::apply_tensor({self.size(0), max_total_size}, self.options(), self);
at::Tensor nmsed_classes = npu_preparation::apply_tensor({self.size(0), max_total_size}, self.options(), self);
at::Tensor nmsed_num = npu_preparation::apply_tensor({self.size(0)}, self.options().dtype(at::kInt), self);
at_npu::native::OpCommand cmd;
cmd.Name("BatchMultiClassNonMaxSuppression")
.Input(self)
.Input(scores)
.Output(nmsed_boxes)
.Output(nmsed_scores)
.Output(nmsed_classes)
.Output(nmsed_num)
.Attr("score_threshold", static_cast<float>(score_threshold))
.Attr("iou_threshold", static_cast<float>(iou_threshold))
.Attr("max_size_per_class", max_size_per_class)
.Attr("max_total_size", max_total_size)
.Attr("change_coordinate_frame", change_coordinate_frame)
.Attr("transpose_box", transpose_box)
.Run();
return std::tie(nmsed_boxes, nmsed_scores, nmsed_classes, nmsed_num);
}
}