#pragma once
// @generated by torchgen/gen.py from DispatchKeyFunction.h

// NB: The implementing C++ file is RegisterDispatchKey.cpp

// The only #includes we need are for custom classes that have defaults in the C++ API
#include <c10/core/MemoryFormat.h>
#include <c10/core/Scalar.h>
#include <ATen/core/Reduction.h>

// Forward declarations of any types needed in the operator signatures.
// We can't directly include these classes because it will cause circular include dependencies.
// This file is included by TensorBody.h, which defines the Tensor class.
#include <ATen/core/ATen_fwd.h>

namespace at {

namespace cuda {

TORCH_API ::std::tuple<at::Tensor,at::Tensor> _scaled_mm(const at::Tensor & self, const at::Tensor & mat2, const c10::optional<at::Tensor> & bias={}, c10::optional<at::ScalarType> out_dtype=c10::nullopt, const c10::optional<at::Tensor> & scale_a={}, const c10::optional<at::Tensor> & scale_b={}, const c10::optional<at::Tensor> & scale_result={});
TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> _scaled_mm_out(at::Tensor & out, at::Tensor & out_amax, const at::Tensor & self, const at::Tensor & mat2, const c10::optional<at::Tensor> & bias={}, c10::optional<at::ScalarType> out_dtype=c10::nullopt, const c10::optional<at::Tensor> & scale_a={}, const c10::optional<at::Tensor> & scale_b={}, const c10::optional<at::Tensor> & scale_result={});
TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> _scaled_mm_outf(const at::Tensor & self, const at::Tensor & mat2, const c10::optional<at::Tensor> & bias, c10::optional<at::ScalarType> out_dtype, const c10::optional<at::Tensor> & scale_a, const c10::optional<at::Tensor> & scale_b, const c10::optional<at::Tensor> & scale_result, at::Tensor & out, at::Tensor & out_amax);

} // namespace cuda
} // namespace at