File size: 2,930 Bytes
430de99
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
// Copyright (c) Facebook, Inc. and its affiliates.
#pragma once
#include <torch/types.h>

namespace detectron2 {

at::Tensor ROIAlign_forward_cpu(
    const at::Tensor& input,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int sampling_ratio,
    bool aligned);

at::Tensor ROIAlign_backward_cpu(
    const at::Tensor& grad,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int batch_size,
    const int channels,
    const int height,
    const int width,
    const int sampling_ratio,
    bool aligned);

#if defined(WITH_CUDA) || defined(WITH_HIP)
at::Tensor ROIAlign_forward_cuda(
    const at::Tensor& input,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int sampling_ratio,
    bool aligned);

at::Tensor ROIAlign_backward_cuda(
    const at::Tensor& grad,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int batch_size,
    const int channels,
    const int height,
    const int width,
    const int sampling_ratio,
    bool aligned);
#endif

// Interface for Python
inline at::Tensor ROIAlign_forward(
    const at::Tensor& input,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int sampling_ratio,
    bool aligned) {
  if (input.is_cuda()) {
#if defined(WITH_CUDA) || defined(WITH_HIP)
    return ROIAlign_forward_cuda(
        input,
        rois,
        spatial_scale,
        pooled_height,
        pooled_width,
        sampling_ratio,
        aligned);
#else
    AT_ERROR("Not compiled with GPU support");
#endif
  }
  return ROIAlign_forward_cpu(
      input,
      rois,
      spatial_scale,
      pooled_height,
      pooled_width,
      sampling_ratio,
      aligned);
}

inline at::Tensor ROIAlign_backward(
    const at::Tensor& grad,
    const at::Tensor& rois,
    const float spatial_scale,
    const int pooled_height,
    const int pooled_width,
    const int batch_size,
    const int channels,
    const int height,
    const int width,
    const int sampling_ratio,
    bool aligned) {
  if (grad.is_cuda()) {
#if defined(WITH_CUDA) || defined(WITH_HIP)
    return ROIAlign_backward_cuda(
        grad,
        rois,
        spatial_scale,
        pooled_height,
        pooled_width,
        batch_size,
        channels,
        height,
        width,
        sampling_ratio,
        aligned);
#else
    AT_ERROR("Not compiled with GPU support");
#endif
  }
  return ROIAlign_backward_cpu(
      grad,
      rois,
      spatial_scale,
      pooled_height,
      pooled_width,
      batch_size,
      channels,
      height,
      width,
      sampling_ratio,
      aligned);
}

} // namespace detectron2