deformable_conv_grad_kernel.cu 14.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
// 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.

15 16
#include "paddle/phi/kernels/deformable_conv_grad_kernel.h"

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
#include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
#include "paddle/phi/backends/gpu/gpu_context.h"
#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/impl/deformable_conv_grad_kernel_impl.h"

namespace phi {

static constexpr int kNumCUDAThreads = 512;
static constexpr int kNumMaximumNumBlocks = 4096;

static inline int NumBlocks(const int N) {
  return std::min((N + kNumCUDAThreads - 1) / kNumCUDAThreads,
                  kNumMaximumNumBlocks);
}

32
template <typename T, typename MT>
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
__global__ void ModulatedDeformableCol2imGpuKernel(
    const int nthreads,
    const T* data_col,
    const T* data_offset,
    const T* data_mask,
    const int channels,
    const int height,
    const int width,
    const int kernel_h,
    const int kernel_w,
    const int pad_h,
    const int pad_w,
    const int stride_h,
    const int stride_w,
    const int dilation_h,
    const int dilation_w,
    const int channel_per_deformable_group,
    const int batch_size,
    const int deformable_group,
    const int height_col,
    const int width_col,
54
    MT* grad_im) {
55 56
  int index = blockIdx.x * blockDim.x + threadIdx.x;
  int offset = blockDim.x * gridDim.x;
Z
Zhang Ting 已提交
57
  // using MT = typename phi::dtype::MPTypeTrait<T>::Type;
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
  for (size_t thread = index; thread < nthreads; thread += offset) {
    const int j = (thread / width_col / height_col / batch_size) % kernel_w;
    const int i =
        (thread / width_col / height_col / batch_size / kernel_w) % kernel_h;
    const int c =
        thread / width_col / height_col / batch_size / kernel_w / kernel_h;

    const int deformable_group_index = c / channel_per_deformable_group;

    int w_out = thread % width_col;
    int h_out = (thread / width_col) % height_col;
    int b = (thread / width_col / height_col) % batch_size;
    int w_in = w_out * stride_w - pad_w;
    int h_in = h_out * stride_h - pad_h;

73 74 75
    const T* data_offset_ptr =
        data_offset + (b * deformable_group + deformable_group_index) * 2 *
                          kernel_h * kernel_w * height_col * width_col;
76 77 78 79 80 81
    const int data_offset_h_ptr =
        ((2 * (i * kernel_w + j)) * height_col + h_out) * width_col + w_out;
    const int data_offset_w_ptr =
        ((2 * (i * kernel_w + j) + 1) * height_col + h_out) * width_col + w_out;
    const int data_mask_hw_ptr =
        ((i * kernel_w + j) * height_col + h_out) * width_col + w_out;
82 83 84 85
    const MT offset_h = static_cast<MT>(data_offset_ptr[data_offset_h_ptr]);
    const MT offset_w = static_cast<MT>(data_offset_ptr[data_offset_w_ptr]);
    const MT cur_inv_h_data = h_in + i * dilation_h + offset_h;
    const MT cur_inv_w_data = w_in + j * dilation_w + offset_w;
86

87
    MT cur_top_grad = static_cast<MT>(data_col[thread]);
88
    if (data_mask) {
89 90 91
      const T* data_mask_ptr =
          data_mask + (b * deformable_group + deformable_group_index) *
                          kernel_h * kernel_w * height_col * width_col;
92
      const MT mask = static_cast<MT>(data_mask_ptr[data_mask_hw_ptr]);
93 94 95 96 97 98 99 100 101 102 103
      cur_top_grad *= mask;
    }
    const int cur_h = static_cast<int>(cur_inv_h_data);
    const int cur_w = static_cast<int>(cur_inv_w_data);
    for (int dy = -2; dy <= 2; dy++) {
      for (int dx = -2; dx <= 2; dx++) {
        if (cur_h + dy >= 0 && cur_h + dy < height && cur_w + dx >= 0 &&
            cur_w + dx < width && abs(cur_inv_h_data - (cur_h + dy)) < 1 &&
            abs(cur_inv_w_data - (cur_w + dx)) < 1) {
          int cur_bottom_grad_pos =
              ((b * channels + c) * height + cur_h + dy) * width + cur_w + dx;
104 105 106 107 108 109
          MT weight = DmcnGetGradientWeight(cur_inv_h_data,
                                            cur_inv_w_data,
                                            cur_h + dy,
                                            cur_w + dx,
                                            height,
                                            width);
110 111 112 113 114 115 116 117
          paddle::platform::CudaAtomicAdd(grad_im + cur_bottom_grad_pos,
                                          weight * cur_top_grad);
        }
      }
    }
  }
}

118
template <typename T, typename MT, typename Context>
119 120 121 122 123 124 125 126 127 128 129
void ModulatedDeformableCol2im(const Context& dev_ctx,
                               const T* data_col,
                               const T* data_offset,
                               const T* data_mask,
                               const std::vector<int64_t>& im_shape,
                               const std::vector<int64_t>& col_shape,
                               const std::vector<int64_t>& kernel_shape,
                               const std::vector<int>& pad,
                               const std::vector<int>& stride,
                               const std::vector<int>& dilation,
                               const int deformable_group,
130
                               MT* grad_im) {
131 132 133 134 135
  int channel_per_deformable_group = im_shape[0] / deformable_group;
  int num_kernels = col_shape[0] * col_shape[1] * col_shape[2] * col_shape[3];
  int blocks = NumBlocks(num_kernels);
  int threads = kNumCUDAThreads;

136
  ModulatedDeformableCol2imGpuKernel<T, MT>
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157
      <<<blocks, threads, 0, dev_ctx.stream()>>>(num_kernels,
                                                 data_col,
                                                 data_offset,
                                                 data_mask,
                                                 im_shape[0],
                                                 im_shape[1],
                                                 im_shape[2],
                                                 kernel_shape[2],
                                                 kernel_shape[3],
                                                 pad[0],
                                                 pad[1],
                                                 stride[0],
                                                 stride[1],
                                                 dilation[0],
                                                 dilation[1],
                                                 channel_per_deformable_group,
                                                 col_shape[1],
                                                 deformable_group,
                                                 col_shape[2],
                                                 col_shape[3],
                                                 grad_im);
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187
}

template <typename T>
__global__ void ModulatedDeformableCol2imCoordGpuKernel(
    const int nthreads,
    const T* data_col,
    const T* data_im,
    const T* data_offset,
    const T* data_mask,
    const int channels,
    const int height,
    const int width,
    const int kernel_h,
    const int kernel_w,
    const int pad_h,
    const int pad_w,
    const int stride_h,
    const int stride_w,
    const int dilation_h,
    const int dilation_w,
    const int channel_per_deformable_group,
    const int batch_size,
    const int offset_channels,
    const int deformable_group,
    const int height_col,
    const int width_col,
    T* grad_offset,
    T* grad_mask) {
  int index = blockIdx.x * blockDim.x + threadIdx.x;
  int offset = blockDim.x * gridDim.x;
188
  using MT = typename phi::dtype::MPTypeTrait<T>::Type;
189
  for (size_t i = index; i < nthreads; i += offset) {
190
    MT val = 0, mval = 0;
191 192 193 194 195 196 197 198
    const int w = i % width_col;
    const int h = (i / width_col) % height_col;
    const int c = (i / width_col / height_col) % offset_channels;
    const int b = (i / width_col / height_col) / offset_channels;

    const int deformable_group_index = c / (2 * kernel_h * kernel_w);
    const int col_step = kernel_h * kernel_w;
    int cnt = 0;
199 200 201 202 203 204 205 206 207 208
    const T* data_col_ptr = data_col + deformable_group_index *
                                           channel_per_deformable_group *
                                           batch_size * width_col * height_col;
    const T* data_im_ptr =
        data_im + (b * deformable_group + deformable_group_index) *
                      channel_per_deformable_group / kernel_h / kernel_w *
                      height * width;
    const T* data_offset_ptr =
        data_offset + (b * deformable_group + deformable_group_index) * 2 *
                          kernel_h * kernel_w * height_col * width_col;
209 210
    const T* data_mask_ptr =
        data_mask
211 212
            ? data_mask + (b * deformable_group + deformable_group_index) *
                              kernel_h * kernel_w * height_col * width_col
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234
            : nullptr;

    const int offset_c = c - deformable_group_index * 2 * kernel_h * kernel_w;

    for (int col_c = offset_c / 2; col_c < channel_per_deformable_group;
         col_c += col_step) {
      const int col_pos =
          (((col_c * batch_size + b) * height_col) + h) * width_col + w;
      const int bp_dir = offset_c % 2;

      int j = (col_pos / width_col / height_col / batch_size) % kernel_w;
      int i =
          (col_pos / width_col / height_col / batch_size / kernel_w) % kernel_h;
      int w_out = col_pos % width_col;
      int h_out = (col_pos / width_col) % height_col;
      int w_in = w_out * stride_w - pad_w;
      int h_in = h_out * stride_h - pad_h;
      const int data_offset_h_ptr =
          (((2 * (i * kernel_w + j)) * height_col + h_out) * width_col + w_out);
      const int data_offset_w_ptr =
          (((2 * (i * kernel_w + j) + 1) * height_col + h_out) * width_col +
           w_out);
235 236 237 238 239

      const MT offset_h = static_cast<MT>(data_offset_ptr[data_offset_h_ptr]);
      const MT offset_w = static_cast<MT>(data_offset_ptr[data_offset_w_ptr]);
      MT inv_h = h_in + i * dilation_h + offset_h;
      MT inv_w = w_in + j * dilation_w + offset_w;
240 241 242
      if (inv_h <= -1 || inv_w <= -1 || inv_h >= height || inv_w >= width) {
        inv_h = inv_w = -2;
      } else {
243 244 245 246 247 248 249 250
        mval +=
            static_cast<MT>(data_col_ptr[col_pos]) *
            funcs::DmcnIm2colBilinear<T, MT>(data_im_ptr + cnt * height * width,
                                             width,
                                             height,
                                             width,
                                             inv_h,
                                             inv_w);
251
      }
252 253 254 255 256 257 258 259
      const MT weight =
          DmcnGetCoordinateWeight<T, MT>(inv_h,
                                         inv_w,
                                         height,
                                         width,
                                         data_im_ptr + cnt * height * width,
                                         width,
                                         bp_dir);
260 261 262
      if (data_mask_ptr) {
        const int data_mask_hw_ptr =
            (((i * kernel_w + j) * height_col + h_out) * width_col + w_out);
263 264
        const MT mask = static_cast<MT>(data_mask_ptr[data_mask_hw_ptr]);
        val += weight * static_cast<MT>(data_col_ptr[col_pos]) * mask;
265
      } else {
266
        val += weight * static_cast<MT>(data_col_ptr[col_pos]);
267 268 269
      }
      cnt += 1;
    }
270
    grad_offset[i] = static_cast<T>(val);
271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
    if (grad_mask && offset_c % 2 == 0)
      grad_mask[(((b * deformable_group + deformable_group_index) * kernel_h *
                      kernel_w +
                  offset_c / 2) *
                     height_col +
                 h) *
                    width_col +
                w] = mval;
  }
}

template <typename T, typename Context>
void ModulatedDeformableCol2imCoord(const Context& dev_ctx,
                                    const T* data_col,
                                    const T* data_im,
                                    const T* data_offset,
                                    const T* data_mask,
                                    const std::vector<int64_t>& im_shape,
                                    const std::vector<int64_t>& col_shape,
                                    const std::vector<int64_t>& kernel_shape,
                                    const std::vector<int>& paddings,
                                    const std::vector<int>& strides,
                                    const std::vector<int>& dilations,
                                    const int deformable_groups,
                                    T* grad_offset,
                                    T* grad_mask) {
  int num_kernels = 2 * kernel_shape[2] * kernel_shape[3] * col_shape[1] *
                    col_shape[2] * col_shape[3] * deformable_groups;
  int channel_per_deformable_group = col_shape[0] / deformable_groups;
  int blocks = NumBlocks(num_kernels);
  int threads = kNumCUDAThreads;

303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
  ModulatedDeformableCol2imCoordGpuKernel<T>
      <<<blocks, threads, 0, dev_ctx.stream()>>>(
          num_kernels,
          data_col,
          data_im,
          data_offset,
          data_mask,
          im_shape[0],
          im_shape[1],
          im_shape[2],
          kernel_shape[2],
          kernel_shape[3],
          paddings[0],
          paddings[1],
          strides[0],
          strides[1],
          dilations[0],
          dilations[1],
          channel_per_deformable_group,
          col_shape[1],
          2 * kernel_shape[2] * kernel_shape[3] * deformable_groups,
          deformable_groups,
          col_shape[2],
          col_shape[3],
          grad_offset,
          grad_mask);
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
}

template <typename T>
__global__ void FilterGradAddupGpuKernel(const int nthreads,
                                         const int n,
                                         const int height,
                                         const int width,
                                         const T* dweight_3d,
                                         T* filter_grad) {
  int index = blockIdx.x * blockDim.x + threadIdx.x;
  int offset = blockDim.x * gridDim.x;
  for (size_t i = index; i < nthreads; i += offset) {
    filter_grad[i] = filter_grad[i] + dweight_3d[i];
  }
}

template <typename T, typename Context>
void FilterGradAddup(const Context& dev_ctx,
                     const int nthreads,
                     const int n,
                     const int height,
                     const int width,
                     const T* dweight_3d,
                     T* filter_grad) {
353 354 355
  FilterGradAddupGpuKernel<T>
      <<<NumBlocks(nthreads), kNumCUDAThreads, 0, dev_ctx.stream()>>>(
          nthreads, n, height, width, dweight_3d, filter_grad);
356 357 358 359 360 361 362 363 364
}

}  // namespace phi

PD_REGISTER_KERNEL(deformable_conv_grad,
                   GPU,
                   ALL_LAYOUT,
                   phi::DeformableConvGradKernel,
                   float,
365 366
                   double,
                   paddle::platform::float16) {}