wehub-resource-sync
94 lines
3.2 KiB
Plaintext
94 lines
3.2 KiB
Plaintext
// 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/fusion/gpu/skip_layernorm_kernel.h"
|
|
#include "paddle/common/errors.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/phi/core/tensor_utils.h"
|
|
#include "paddle/phi/kernels/funcs/skip_layernorm_functor.h"
|
|
|
|
namespace phi {
|
|
namespace fusion {
|
|
|
|
template <typename T, typename Context>
|
|
void SkipLayerNormKernel(const Context &dev_ctx,
|
|
const DenseTensor &x,
|
|
const DenseTensor &y,
|
|
const DenseTensor &scale,
|
|
const DenseTensor &bias,
|
|
const float epsilon,
|
|
const int begin_norm_axis,
|
|
DenseTensor *out) {
|
|
auto *X_d = x.data<T>();
|
|
auto *Y_d = y.data<T>();
|
|
auto *scale_d = scale.data<T>();
|
|
auto *bias_d = bias.data<T>();
|
|
|
|
out->Resize(x.dims());
|
|
auto *output_d = dev_ctx.template Alloc<T>(out, out->numel() * sizeof(T));
|
|
|
|
size_t num = 1;
|
|
for (size_t i = 0; i < x.dims().size(); i++) {
|
|
num *= x.dims()[i];
|
|
}
|
|
int64_t hidden = x.dims()[2];
|
|
// TODO(large-tensor): downstream functors may still use int
|
|
|
|
funcs::SkipLayerNormFunctor<T> skip_layer_norm_func;
|
|
|
|
if (std::is_same<T, phi::float16>::value) {
|
|
const half *X_new = reinterpret_cast<const half *>(X_d);
|
|
const half *Y_new = reinterpret_cast<const half *>(Y_d);
|
|
const half *scale_new = reinterpret_cast<const half *>(scale_d);
|
|
const half *bias_new = reinterpret_cast<const half *>(bias_d);
|
|
half *output_new = reinterpret_cast<half *>(output_d);
|
|
funcs::SkipLayerNormFunctor<half> skip_layer_norm_func;
|
|
skip_layer_norm_func(num,
|
|
hidden,
|
|
X_new,
|
|
Y_new,
|
|
scale_new,
|
|
bias_new,
|
|
output_new,
|
|
epsilon,
|
|
dev_ctx.stream());
|
|
} else {
|
|
funcs::SkipLayerNormFunctor<T> skip_layer_norm_func;
|
|
skip_layer_norm_func(num,
|
|
hidden,
|
|
X_d,
|
|
Y_d,
|
|
scale_d,
|
|
bias_d,
|
|
output_d,
|
|
epsilon,
|
|
dev_ctx.stream());
|
|
}
|
|
}
|
|
|
|
} // namespace fusion
|
|
} // namespace phi
|
|
|
|
#if defined(PADDLE_WITH_CUDA)
|
|
PD_REGISTER_KERNEL(skip_layernorm,
|
|
GPU,
|
|
ALL_LAYOUT,
|
|
phi::fusion::SkipLayerNormKernel,
|
|
float,
|
|
phi::float16) {}
|
|
#else
|
|
PD_REGISTER_KERNEL(
|
|
skip_layernorm, GPU, ALL_LAYOUT, phi::fusion::SkipLayerNormKernel, float) {}
|
|
#endif
|