wehub-resource-sync
118 lines
4.0 KiB
C++
118 lines
4.0 KiB
C++
// Copyright (c) 2023 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/scatter_grad_kernel.h"
|
|
|
|
#include "paddle/phi/backends/xpu/enforce_xpu.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/phi/kernels/full_kernel.h"
|
|
|
|
namespace phi {
|
|
|
|
template <typename T, typename Context>
|
|
void ScatterGradKernel(const Context &dev_ctx,
|
|
const DenseTensor &index,
|
|
const DenseTensor &updates,
|
|
const DenseTensor &out_grad,
|
|
bool overwrite,
|
|
DenseTensor *x_grad,
|
|
DenseTensor *updates_grad) {
|
|
if (out_grad.numel() == 0) {
|
|
if (x_grad) {
|
|
dev_ctx.template Alloc<T>(x_grad);
|
|
}
|
|
if (updates_grad) {
|
|
Full<T, Context>(dev_ctx, updates_grad->dims(), 0, updates_grad);
|
|
}
|
|
return;
|
|
}
|
|
if (index.numel() == 0) {
|
|
if (x_grad) {
|
|
phi::Copy<Context>(dev_ctx, out_grad, dev_ctx.GetPlace(), false, x_grad);
|
|
}
|
|
if (updates_grad) {
|
|
dev_ctx.template Alloc<T>(updates_grad);
|
|
Full<T, Context>(dev_ctx, updates_grad->dims(), 0, updates_grad);
|
|
}
|
|
return;
|
|
}
|
|
using XPUType = typename XPUTypeTrait<T>::Type;
|
|
|
|
const auto &index_type = index.dtype();
|
|
bool index_type_match =
|
|
index_type == DataType::INT32 || index_type == DataType::INT64;
|
|
PADDLE_ENFORCE_EQ(index_type_match,
|
|
true,
|
|
common::errors::InvalidArgument(
|
|
"scatter_op index holds the wrong type, it holds [%s],"
|
|
"but desires to be [%s] or [%s]",
|
|
index_type,
|
|
DataType::INT32,
|
|
DataType::INT64));
|
|
|
|
T *x_grad_data = nullptr;
|
|
T *updates_grad_data = nullptr;
|
|
if (x_grad != nullptr) {
|
|
dev_ctx.template Alloc<T>(x_grad);
|
|
x_grad_data = x_grad->data<T>();
|
|
}
|
|
if (updates_grad != nullptr) {
|
|
dev_ctx.template Alloc<T>(updates_grad);
|
|
updates_grad_data = updates_grad->data<T>();
|
|
}
|
|
|
|
std::vector<int64_t> x_grad_shape;
|
|
DDim out_dims = out_grad.dims();
|
|
for (int i = 0; i < out_dims.size(); i++) {
|
|
x_grad_shape.push_back(out_dims[i]);
|
|
}
|
|
|
|
int64_t index_size = index.numel();
|
|
|
|
int r;
|
|
if (index_type == DataType::INT32) {
|
|
auto index_data = const_cast<int *>(index.data<int>());
|
|
xpu::VectorParam<int> indices{nullptr, index_size, index_data};
|
|
r = xpu::scatter_grad<XPUType, int>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType *>(out_grad.data<T>()),
|
|
indices,
|
|
reinterpret_cast<XPUType *>(x_grad_data),
|
|
reinterpret_cast<XPUType *>(updates_grad_data),
|
|
x_grad_shape,
|
|
overwrite);
|
|
} else if (index_type == DataType::INT64) {
|
|
auto index_data = const_cast<int64_t *>(index.data<int64_t>());
|
|
xpu::VectorParam<int64_t> indices{nullptr, index_size, index_data};
|
|
r = xpu::scatter_grad<XPUType, int64_t>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType *>(out_grad.data<T>()),
|
|
indices,
|
|
reinterpret_cast<XPUType *>(x_grad_data),
|
|
reinterpret_cast<XPUType *>(updates_grad_data),
|
|
x_grad_shape,
|
|
overwrite);
|
|
}
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "scatter grad");
|
|
}
|
|
} // namespace phi
|
|
|
|
PD_REGISTER_KERNEL(scatter_grad,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::ScatterGradKernel,
|
|
float,
|
|
phi::float16,
|
|
phi::bfloat16) {}
|