530 lines
17 KiB
C++
530 lines
17 KiB
C++
// Copyright (c) 2022 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/conv_transpose_kernel.h"
|
|
|
|
#include "glog/logging.h"
|
|
|
|
#include "paddle/phi/backends/xpu/enforce_xpu.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/phi/kernels/cpu/conv_util.h"
|
|
#include "paddle/phi/kernels/full_kernel.h"
|
|
#include "paddle/phi/kernels/xpu/conv_utils_xpu.h"
|
|
#include "paddle/phi/kernels/xpu/xpu_api_wrapper.h"
|
|
#ifdef PADDLE_WITH_XPU_XRE5
|
|
#include "xpudnn/xpudnn.h"
|
|
namespace xpudnn = baidu::xpu::xpudnn;
|
|
#endif
|
|
namespace phi {
|
|
|
|
template <typename T, typename Context>
|
|
void Conv2dTransposeKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& filter,
|
|
const std::vector<int>& strides,
|
|
const std::vector<int>& paddings,
|
|
const std::vector<int>& output_padding,
|
|
const IntArray& output_size,
|
|
const std::string& padding_algorithm,
|
|
int groups,
|
|
const std::vector<int>& dilations,
|
|
const std::string& data_format,
|
|
DenseTensor* out) {
|
|
using XPUType = typename XPUTypeTrait<T>::Type;
|
|
if (x.numel() == 0 || filter.numel() == 0) {
|
|
Full<T, Context>(dev_ctx, out->dims(), 0, out);
|
|
return;
|
|
}
|
|
dev_ctx.template Alloc<T>(out);
|
|
|
|
PADDLE_ENFORCE_EQ(
|
|
data_format == "NHWC" || data_format == "NDHWC",
|
|
false,
|
|
errors::InvalidArgument(
|
|
("XPU do support data_format is NCHW in conv_transpose op.")));
|
|
|
|
DDim in_data_dims = slice_ddim(x.dims(), 2, x.dims().size());
|
|
DDim filter_data_dims = slice_ddim(filter.dims(), 2, filter.dims().size());
|
|
|
|
#ifdef PADDLE_WITH_XPU_XRE5
|
|
std::vector<int64_t> ksize = vectorize<int64_t>(filter_data_dims);
|
|
std::vector<int64_t> paddings_ =
|
|
std::vector<int64_t>(paddings.begin(), paddings.end());
|
|
std::vector<int64_t> dilations_ =
|
|
std::vector<int64_t>(dilations.begin(), dilations.end());
|
|
std::vector<int64_t> strides_ =
|
|
std::vector<int64_t>(strides.begin(), strides.end());
|
|
UpdatePaddingAndDilation(&paddings_,
|
|
&dilations_,
|
|
padding_algorithm,
|
|
in_data_dims,
|
|
strides_,
|
|
ksize);
|
|
|
|
const int64_t batch_size = x.dims()[0];
|
|
const int64_t img_yc = x.dims()[1];
|
|
const int64_t img_xc = out->dims()[1];
|
|
const int64_t img_xh = out->dims()[2];
|
|
const int64_t img_xw = out->dims()[3];
|
|
|
|
int fc_calc_type = FCCalcType<XPUType>();
|
|
if (fc_calc_type == XPUFCCalcType::FC_INT32) {
|
|
int r = xpudnn::conv2d_transpose_fusion_v2<float, float, float, int32_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
static_cast<int64_t>(groups),
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
xpu::Activation_t::LINEAR,
|
|
true,
|
|
nullptr);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_fusion_v2");
|
|
} else if (fc_calc_type == XPUFCCalcType::FC_FLOAT) {
|
|
int r = xpudnn::conv2d_transpose_fusion_v2<float, float, float, float>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
static_cast<int64_t>(groups),
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
xpu::Activation_t::LINEAR,
|
|
true,
|
|
nullptr);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_fusion_v2");
|
|
} else if (fc_calc_type == XPUFCCalcType::FC_INT32_WITH_LL) {
|
|
if (output_size.size()) {
|
|
VLOG(4) << "int_with_ll quantization is not supported when output_size "
|
|
"is specified, "
|
|
<< "use int31 instead";
|
|
int r = xpudnn::conv2d_transpose_fusion_v2<float, float, float, int32_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
static_cast<int64_t>(groups),
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
xpu::Activation_t::LINEAR,
|
|
true,
|
|
nullptr);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_fusion_v2");
|
|
} else {
|
|
// xpu::conv2d_transpose_v2 do not support int_with_ll now
|
|
// use xpu::conv2d_transpose
|
|
int64_t img_yh = static_cast<int64_t>(x.dims()[2]);
|
|
int64_t img_yw = static_cast<int64_t>(x.dims()[3]);
|
|
int r = xpudnn::
|
|
conv2d_transpose_fusion_v2<float, float, float, int_with_ll_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_yh,
|
|
img_yw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
static_cast<int64_t>(groups),
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
xpu::Activation_t::LINEAR,
|
|
true,
|
|
nullptr);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose");
|
|
}
|
|
} else {
|
|
int r =
|
|
xpudnn::conv2d_transpose_fusion_v2<XPUType, XPUType, XPUType, int16_t>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType*>(x.data<T>()),
|
|
reinterpret_cast<const XPUType*>(filter.data<T>()),
|
|
reinterpret_cast<XPUType*>(out->data<T>()),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
static_cast<int64_t>(groups),
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
xpu::Activation_t::LINEAR,
|
|
true,
|
|
nullptr);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_fusion_v2");
|
|
}
|
|
#else
|
|
std::vector<int64_t> ksize = vectorize<int64_t>(filter_data_dims);
|
|
std::vector<int64_t> strides_ =
|
|
std::vector<int64_t>(strides.begin(), strides.end());
|
|
std::vector<int64_t> paddings_ =
|
|
std::vector<int64_t>(paddings.begin(), paddings.end());
|
|
std::vector<int64_t> dilations_ =
|
|
std::vector<int64_t>(dilations.begin(), dilations.end());
|
|
UpdatePaddingAndDilation(&paddings_,
|
|
&dilations_,
|
|
padding_algorithm,
|
|
in_data_dims,
|
|
strides_,
|
|
ksize);
|
|
|
|
const int64_t batch_size = x.dims()[0];
|
|
const int64_t img_yc = x.dims()[1];
|
|
const int64_t img_xc = out->dims()[1];
|
|
const int64_t img_xh = out->dims()[2];
|
|
const int64_t img_xw = out->dims()[3];
|
|
|
|
int fc_calc_type = FCCalcType<XPUType>();
|
|
if (fc_calc_type == XPUFCCalcType::FC_INT32) {
|
|
int r = xpu::conv2d_transpose_v2<float, float, float, int32_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
true);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_v2");
|
|
} else if (fc_calc_type == XPUFCCalcType::FC_FLOAT) {
|
|
int r = xpu::conv2d_transpose_v2<float, float, float, float>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
true);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_v2");
|
|
} else if (fc_calc_type == XPUFCCalcType::FC_INT32_WITH_LL) {
|
|
if (output_size.size()) {
|
|
VLOG(4) << "int_with_ll quantization is not supported when output_size "
|
|
"is specified, "
|
|
<< "use int31 instead";
|
|
int r = xpu::conv2d_transpose_v2<float, float, float, int32_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
true);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_v2");
|
|
} else {
|
|
// xpu::conv2d_transpose_v2 do not support int_with_ll now
|
|
// use xpu::conv2d_transpose
|
|
int64_t img_yh = x.dims()[2];
|
|
int64_t img_yw = x.dims()[3];
|
|
int r = xpu::conv2d_transpose<float, float, float, int_with_ll_t>(
|
|
dev_ctx.x_context(),
|
|
x.data<float>(),
|
|
filter.data<float>(),
|
|
out->data<float>(),
|
|
batch_size,
|
|
img_yc,
|
|
img_yh,
|
|
img_yw,
|
|
img_xc,
|
|
ksize,
|
|
strides,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
true);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose");
|
|
}
|
|
} else {
|
|
int r = xpu::conv2d_transpose_v2<XPUType, XPUType, XPUType, int16_t>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType*>(x.data<T>()),
|
|
reinterpret_cast<const XPUType*>(filter.data<T>()),
|
|
reinterpret_cast<XPUType*>(out->data<T>()),
|
|
batch_size,
|
|
img_yc,
|
|
img_xh,
|
|
img_xw,
|
|
img_xc,
|
|
ksize,
|
|
strides,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
true);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "conv2d_transpose_v2");
|
|
}
|
|
#endif
|
|
}
|
|
|
|
template <typename T, typename Context>
|
|
void Conv3dTransposeKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& filter,
|
|
const std::vector<int>& strides,
|
|
const std::vector<int>& paddings,
|
|
const std::vector<int>& output_padding,
|
|
const std::vector<int>& output_size,
|
|
const std::string& padding_algorithm,
|
|
int groups,
|
|
const std::vector<int>& dilations,
|
|
const std::string& data_format,
|
|
DenseTensor* out) {
|
|
using XPUType = typename XPUTypeTrait<T>::Type;
|
|
|
|
dev_ctx.template Alloc<T>(out);
|
|
|
|
// data_format follow the legacy interface:
|
|
// https://github.com/PaddlePaddle/Paddle/blob/639abfd4/python/paddle/nn/functional/conv.py#L1726
|
|
PADDLE_ENFORCE_EQ(
|
|
data_format == "NCHW" || data_format == "NHWC",
|
|
true,
|
|
errors::InvalidArgument(
|
|
("XPU only support data_format is NCHW(in Python, it is specified as "
|
|
"NCDHW) or NHWC(in Python, it is specified as NDHWC) in "
|
|
"conv3d_transpose op.")));
|
|
|
|
DDim in_data_dims;
|
|
if (data_format == "NHWC") {
|
|
in_data_dims = slice_ddim(x.dims(), 1, x.dims().size() - 1);
|
|
} else {
|
|
in_data_dims = slice_ddim(x.dims(), 2, x.dims().size());
|
|
}
|
|
DDim filter_data_dims = slice_ddim(filter.dims(), 2, filter.dims().size());
|
|
|
|
std::vector<int64_t> ksize = vectorize<int64_t>(filter_data_dims);
|
|
std::vector<int64_t> paddings_ =
|
|
std::vector<int64_t>(paddings.begin(), paddings.end());
|
|
std::vector<int64_t> dilations_ =
|
|
std::vector<int64_t>(dilations.begin(), dilations.end());
|
|
std::vector<int64_t> strides_ =
|
|
std::vector<int64_t>(strides.begin(), strides.end());
|
|
UpdatePaddingAndDilation(&paddings_,
|
|
&dilations_,
|
|
padding_algorithm,
|
|
in_data_dims,
|
|
strides_,
|
|
ksize);
|
|
|
|
for (int64_t dilation : dilations_) {
|
|
PADDLE_ENFORCE_LE(
|
|
dilation,
|
|
1,
|
|
errors::Unimplemented(
|
|
"XPU do not support dilation > 1 in conv3d_transpose."));
|
|
}
|
|
|
|
int64_t batch_size = x.dims()[0];
|
|
int64_t img_yc = x.dims()[1];
|
|
int64_t img_yd = x.dims()[2];
|
|
int64_t img_yh = x.dims()[3];
|
|
int64_t img_yw = x.dims()[4];
|
|
int64_t img_xc = out->dims()[1];
|
|
|
|
bool is_ndhwc = false;
|
|
if (data_format == "NHWC") {
|
|
img_yc = x.dims()[4];
|
|
img_yd = x.dims()[1];
|
|
img_yh = x.dims()[2];
|
|
img_yw = x.dims()[3];
|
|
img_xc = out->dims()[4];
|
|
is_ndhwc = true;
|
|
|
|
PADDLE_ENFORCE_LE(
|
|
groups,
|
|
1,
|
|
errors::Unimplemented("XPU do not support group > 1 when data_format "
|
|
"is NHWC(in Python, it is specified as NDHWC) "
|
|
"in conv3d_transpose."));
|
|
}
|
|
|
|
const XPUType* filter_data =
|
|
reinterpret_cast<const XPUType*>(filter.data<T>());
|
|
|
|
int fc_calc_type = GetConvCalcType<XPUType>();
|
|
PD_VISIT_XPU_CONV_TYPES(XPUType, fc_calc_type, "conv3d_transpose", [&] {
|
|
using XPUTypeFP16 = typename XPUTypeTrait<phi::float16>::Type;
|
|
using RealTGEMM = std::conditional_t<
|
|
(
|
|
// 如果 XPUType 是 XPUTypeFP16 且 TGEMM 不是 FP16 或 int16
|
|
(std::is_same_v<XPUType, XPUTypeFP16> &&
|
|
!std::is_same_v<TGEMM, XPUTypeFP16> &&
|
|
!std::is_same_v<TGEMM, int16_t>) ||
|
|
|
|
// 如果 XPUType 是 float 且 TGEMM 不是 int32、int16 或 tfloat32
|
|
(std::is_same_v<XPUType, float> &&
|
|
!std::is_same_v<TGEMM, int32_t> &&
|
|
!std::is_same_v<TGEMM, int16_t> &&
|
|
!std::is_same_v<TGEMM, tfloat32>)),
|
|
std::conditional_t<std::is_same_v<XPUType, XPUTypeFP16>,
|
|
XPUTypeFP16,
|
|
tfloat32>,
|
|
TGEMM>;
|
|
|
|
int ret = xpudnn::conv3d_transpose<XPUType, XPUType, XPUType, RealTGEMM>(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPUType*>(x.data<T>()),
|
|
filter_data, // filter的shape固定为[yc, xc, fd, fh, fw],
|
|
reinterpret_cast<XPUType*>(out->data<T>()),
|
|
batch_size,
|
|
img_yc,
|
|
img_yd,
|
|
img_yh,
|
|
img_yw,
|
|
img_xc,
|
|
ksize,
|
|
strides_,
|
|
paddings_,
|
|
dilations_,
|
|
groups,
|
|
nullptr,
|
|
nullptr,
|
|
nullptr,
|
|
is_ndhwc);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(ret, "conv3d_transpose");
|
|
});
|
|
}
|
|
|
|
template <typename T, typename Context>
|
|
void DepthwiseConv2dTransposeKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& filter,
|
|
const std::vector<int>& strides,
|
|
const std::vector<int>& paddings,
|
|
const std::vector<int>& output_padding,
|
|
const IntArray& output_size,
|
|
const std::string& padding_algorithm,
|
|
int groups,
|
|
const std::vector<int>& dilations,
|
|
const std::string& data_format,
|
|
DenseTensor* out) {
|
|
Conv2dTransposeKernel<T, Context>(dev_ctx,
|
|
x,
|
|
filter,
|
|
strides,
|
|
paddings,
|
|
output_padding,
|
|
output_size,
|
|
padding_algorithm,
|
|
groups,
|
|
dilations,
|
|
data_format,
|
|
out);
|
|
}
|
|
|
|
} // namespace phi
|
|
PD_REGISTER_KERNEL(depthwise_conv2d_transpose,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::DepthwiseConv2dTransposeKernel,
|
|
float,
|
|
phi::float16) {}
|
|
|
|
PD_REGISTER_KERNEL(conv2d_transpose,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::Conv2dTransposeKernel,
|
|
float,
|
|
phi::float16) {}
|
|
|
|
PD_REGISTER_KERNEL(conv3d_transpose,
|
|
XPU,
|
|
ALL_LAYOUT,
|
|
phi::Conv3dTransposeKernel,
|
|
float,
|
|
phi::float16) {}
|