#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 &histc_out_nocheck(at::Tensor &result, const at::Tensor &self, int64_t bins, const at::Scalar &min,
const at::Scalar &max)
{
at_npu::native::OpCommand cmd;
cmd.Name("Histogram")
.Input(self)
.Output(result)
.Attr("bins", bins)
.Attr("min", min)
.Attr("max", max)
.Run();
return result;
}
}
at::Tensor &histc_out(const at::Tensor &self, int64_t bins, const at::Scalar &min, const at::Scalar &max,
at::Tensor &out)
{
npu_preparation::CheckOut({self}, out, self, {bins});
if (!npu_utils::check_match(&out)) {
at::Tensor contiguous_result = npu_utils::format_contiguous(out);
histc_out_nocheck(contiguous_result, self, bins, min, max);
npu_utils::format_fresh_view(out, contiguous_result);
} else {
histc_out_nocheck(out, self, bins, min, max);
}
return out;
}
at::Tensor histc(const at::Tensor &self, int64_t bins, const at::Scalar &min, const at::Scalar &max)
{
TORCH_CHECK(self.dtype() == at::kInt || self.dtype() == at::kFloat || self.dtype() == at::kHalf,
"histc input only supported Int32, Float16, Float32, but got", self.dtype(),
OPS_ERROR(ErrCode::TYPE));
bool is_fp = (self.dtype() == at::kInt) ? false : true;
at::Tensor result =
npu_preparation::apply_tensor({bins}, self.options().dtype(is_fp ? at::kFloat : at::kInt), self);
histc_out_nocheck(result, self, bins, min, max);
return result;
}
}