125 lines
3.8 KiB
C++
125 lines
3.8 KiB
C++
// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
|
|
#include "paddle/phi/kernels/top_k_kernel.h"
|
|
|
|
#include "paddle/phi/backends/xpu/enforce_xpu.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/phi/kernels/full_kernel.h"
|
|
#include "paddle/phi/kernels/funcs/math_function.h"
|
|
#include "paddle/phi/kernels/xpu/xpu_mem_util.h"
|
|
|
|
namespace {
|
|
inline void GetDims(
|
|
const phi::DDim& dim, int axis, int64_t* pre, int64_t* n, int64_t* post) {
|
|
*pre = 1;
|
|
*post = 1;
|
|
*n = dim[axis];
|
|
for (int i = 0; i < axis; ++i) {
|
|
(*pre) *= dim[i];
|
|
}
|
|
for (int i = axis + 1; i < dim.size(); ++i) {
|
|
(*post) *= dim[i];
|
|
}
|
|
}
|
|
} // namespace
|
|
|
|
namespace phi {
|
|
|
|
template <typename T, typename Context>
|
|
void TopkGradKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& indices,
|
|
const DenseTensor& out_grad,
|
|
const Scalar& k_scalar,
|
|
int axis,
|
|
bool largest,
|
|
bool sorted,
|
|
DenseTensor* x_grad) {
|
|
using XPUType = typename XPUTypeTrait<T>::Type;
|
|
|
|
if (x_grad && x_grad->numel() == 0) {
|
|
dev_ctx.template Alloc<T>(x_grad);
|
|
return;
|
|
}
|
|
const auto& in_dims = x.dims();
|
|
|
|
// get the real the axis and the k
|
|
if (axis < 0) {
|
|
axis += in_dims.size();
|
|
}
|
|
|
|
// allocate the xpu memory for the x_grad
|
|
T* x_grad_data = dev_ctx.template Alloc<T>(x_grad);
|
|
const T* out_grad_data = out_grad.data<T>();
|
|
const int64_t* indices_data = indices.data<int64_t>();
|
|
|
|
if (in_dims.size() == 0) {
|
|
Copy<Context>(dev_ctx, out_grad, dev_ctx.GetPlace(), false, x_grad);
|
|
return;
|
|
}
|
|
|
|
int64_t pre, n, post;
|
|
GetDims(in_dims, axis, &pre, &n, &post);
|
|
|
|
Full<T, Context>(dev_ctx, x_grad->dims(), 0.0f, x_grad);
|
|
|
|
// launch the xpu kernel to assign the grad
|
|
int ret = xpu::scatter_element<XPUType, int64_t>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType*>(x_grad_data),
|
|
reinterpret_cast<const XPUType*>(out_grad_data),
|
|
indices_data,
|
|
reinterpret_cast<XPUType*>(x_grad_data),
|
|
vectorize(x_grad->dims()),
|
|
vectorize(out_grad.dims()),
|
|
vectorize(indices.dims()),
|
|
axis,
|
|
/*reduction=override*/ 0);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(ret, "scatter");
|
|
}
|
|
|
|
template <typename T, typename Context>
|
|
void TopkV1GradKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& indices,
|
|
const DenseTensor& out_grad,
|
|
const Scalar& k_scalar,
|
|
DenseTensor* x_grad) {
|
|
TopkGradKernel<T, Context>(
|
|
dev_ctx, x, indices, out_grad, k_scalar, -1, true, true, x_grad);
|
|
}
|
|
|
|
} // namespace phi
|
|
|
|
PD_REGISTER_KERNEL(topk_grad,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::TopkGradKernel,
|
|
float,
|
|
int,
|
|
int64_t,
|
|
phi::float16,
|
|
phi::bfloat16) {}
|
|
|
|
PD_REGISTER_KERNEL(topk_v1_grad,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::TopkV1GradKernel,
|
|
float,
|
|
int,
|
|
int64_t,
|
|
phi::float16,
|
|
phi::bfloat16) {}
|