You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

test_deconvolution.cpp 4.3 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117
  1. // Tencent is pleased to support the open source community by making ncnn available.
  2. //
  3. // Copyright (C) 2019 THL A29 Limited, a Tencent company. All rights reserved.
  4. //
  5. // Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
  6. // in compliance with the License. You may obtain a copy of the License at
  7. //
  8. // https://opensource.org/licenses/BSD-3-Clause
  9. //
  10. // Unless required by applicable law or agreed to in writing, software distributed
  11. // under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
  12. // CONDITIONS OF ANY KIND, either express or implied. See the License for the
  13. // specific language governing permissions and limitations under the License.
  14. #include "layer/deconvolution.h"
  15. #include "testutil.h"
  16. static int test_deconvolution(int w, int h, int c, int outch, int kernel, int dilation, int stride, int pad, int bias, int output_pad_right, int output_pad_bottom, int output_w, int output_h)
  17. {
  18. ncnn::Mat a = RandomMat(w, h, c);
  19. if (output_w > 0 && output_h > 0 && pad != -233 && pad != -234)
  20. {
  21. pad = -233;
  22. }
  23. ncnn::ParamDict pd;
  24. pd.set(0, outch); // num_output
  25. pd.set(1, kernel); // kernel_w
  26. pd.set(2, dilation); // dilation_w
  27. pd.set(3, stride); // stride_w
  28. pd.set(4, pad); // pad_w
  29. pd.set(5, bias); // bias_term
  30. pd.set(6, outch * c * kernel * kernel);
  31. int activation_type = RAND() % 5; // 0 1 2 3 4
  32. ncnn::Mat activation_params(2);
  33. activation_params[0] = RandomFloat(-1, 0); // alpha
  34. activation_params[1] = RandomFloat(0, 1); // beta
  35. pd.set(9, activation_type);
  36. pd.set(10, activation_params);
  37. pd.set(18, output_pad_right);
  38. pd.set(19, output_pad_bottom);
  39. pd.set(20, output_w);
  40. pd.set(21, output_h);
  41. std::vector<ncnn::Mat> weights(2);
  42. weights[0] = RandomMat(outch * c * kernel * kernel);
  43. weights[1] = RandomMat(outch);
  44. int ret = test_layer<ncnn::Deconvolution>("Deconvolution", pd, weights, a);
  45. if (ret != 0)
  46. {
  47. fprintf(stderr, "test_deconvolution failed w=%d h=%d c=%d outch=%d kernel=%d dilation=%d stride=%d pad=%d bias=%d act=%d actparams=[%f,%f] output_pad_right=%d output_pad_bottom=%d output_w=%d output_h=%d\n", w, h, c, outch, kernel, dilation, stride, pad, bias, activation_type, activation_params[0], activation_params[1], output_pad_right, output_pad_bottom, output_w, output_h);
  48. }
  49. return ret;
  50. }
  51. static int test_deconvolution_0()
  52. {
  53. static const int kdsp[16][4] = {
  54. {1, 1, 1, 0},
  55. {1, 1, 2, 0},
  56. {2, 1, 1, 1},
  57. {2, 1, 2, -233},
  58. {3, 1, 1, 1},
  59. {3, 1, 2, 1},
  60. {3, 2, 1, 1},
  61. {4, 1, 1, -233},
  62. {4, 1, 2, -234},
  63. {4, 2, 1, -234},
  64. {5, 1, 1, 2},
  65. {5, 1, 2, 2},
  66. {5, 2, 2, 2},
  67. {7, 1, 1, 3},
  68. {7, 1, 2, 3},
  69. {7, 2, 1, -233},
  70. };
  71. for (int i = 0; i < 16; i++)
  72. {
  73. const int k = kdsp[i][0];
  74. const int d = kdsp[i][1];
  75. const int s = kdsp[i][2];
  76. const int p = kdsp[i][3];
  77. int ret = 0
  78. || test_deconvolution(9, 7, 1, 1, k, d, s, p, 1, 0, 0, 0, 0)
  79. || test_deconvolution(9, 7, 4, 13, k, d, s, p, 0, 1, 1, 7, 5)
  80. || test_deconvolution(9, 7, 13, 4, k, d, s, p, 1, 1, 0, 0, 0)
  81. || test_deconvolution(9, 7, 4, 8, k, d, s, p, 0, 0, 1, 0, 0)
  82. || test_deconvolution(9, 7, 8, 4, k, d, s, p, 1, 0, 0, 7, 5)
  83. || test_deconvolution(9, 7, 8, 13, k, d, s, p, 0, 2, 2, 0, 0)
  84. || test_deconvolution(9, 7, 13, 8, k, d, s, p, 1, 2, 0, 0, 0)
  85. || test_deconvolution(9, 7, 16, 16, k, d, s, p, 0, 0, 2, 7, 5);
  86. if (ret != 0)
  87. return -1;
  88. }
  89. return 0
  90. || test_deconvolution(7, 5, 24, 32, 4, 2, 2, 2, 1, 0, 0, 0, 0)
  91. || test_deconvolution(7, 5, 32, 24, 4, 2, 2, 2, 1, 0, 0, 0, 0)
  92. || test_deconvolution(7, 5, 28, 32, 4, 2, 2, 2, 1, 0, 0, 0, 0)
  93. || test_deconvolution(7, 5, 32, 28, 4, 2, 2, 2, 1, 0, 0, 0, 0)
  94. || test_deconvolution(7, 5, 26, 32, 4, 2, 2, 2, 1, 0, 0, 0, 0)
  95. || test_deconvolution(7, 5, 32, 26, 4, 2, 2, 2, 1, 0, 0, 0, 0);
  96. }
  97. int main()
  98. {
  99. SRAND(7767517);
  100. return test_deconvolution_0();
  101. }