megenginelite-sys 1.8.2

A safe megenginelite wrapper in Rust
Documentation 
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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
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
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232

/**
 * \file dnn/src/naive/roi_pooling/opr_impl.cpp
 * MegEngine is Licensed under the Apache License, Version 2.0 (the "License")
 *
 * Copyright (c) 2014-2021 Megvii Inc. All rights reserved.
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT ARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 */
#include "src/naive/roi_pooling/opr_impl.h"

#include "src/common/roi_pooling_helper.h"
#include "src/common/utils.h"
#include "src/naive/handle.h"

namespace {

using namespace megdnn;
using Param = param::ROIPooling;

template <typename T, typename Pooler>
void forward_impl(
        _megdnn_tensor_in src, _megdnn_tensor_in rois, _megdnn_tensor_in dst,
        _megdnn_tensor_out index, float spatial_scale) {
    auto C = src.layout.shape[1], IH = src.layout.shape[2], IW = src.layout.shape[3];
    auto OH = dst.layout.shape[2], OW = dst.layout.shape[3];

    auto total_nr_elem = dst.layout.total_nr_elems();
    auto pooled_height = OH, pooled_width = OW;
    auto height = IH, width = IW;
    auto channels = C;
    for (size_t i = 0; i < total_nr_elem; ++i) {
        int pw = i % pooled_width;
        int ph = (i / pooled_width) % pooled_height;
        int c = (i / pooled_width / pooled_height) % channels;
        int n = i / pooled_width / pooled_height / channels;
        auto rois_ptr = rois.ptr<T>() + n * 5;
        int roi_batch_ind = rois_ptr[0];
        int roi_start_w = round(rois_ptr[1] * spatial_scale);
        int roi_start_h = round(rois_ptr[2] * spatial_scale);
        int roi_end_w = round(rois_ptr[3] * spatial_scale);
        int roi_end_h = round(rois_ptr[4] * spatial_scale);
        // Force malformed ROIs to be 1x1
        int roi_width = std::max(roi_end_w - roi_start_w + 1, 1);
        int roi_height = std::max(roi_end_h - roi_start_h + 1, 1);
        float bin_size_h =
                static_cast<float>(roi_height) / static_cast<float>(pooled_height);
        float bin_size_w =
                static_cast<float>(roi_width) / static_cast<float>(pooled_width);

        int hstart = static_cast<int>(floor(static_cast<float>(ph) * bin_size_h));
        int wstart = static_cast<int>(floor(static_cast<float>(pw) * bin_size_w));
        int hend = static_cast<int>(ceil(static_cast<float>(ph + 1) * bin_size_h));
        int wend = static_cast<int>(ceil(static_cast<float>(pw + 1) * bin_size_w));
        // Add roi offsets and clip to input boundaries
        hstart = std::min<int>(std::max(hstart + roi_start_h, 0), height);
        hend = std::min<int>(std::max(hend + roi_start_h, 0), height);
        wstart = std::min<int>(std::max(wstart + roi_start_w, 0), width);
        wend = std::min<int>(std::max(wend + roi_start_w, 0), width);

        Pooler pooler;
        auto feat_map_ptr =
                src.ptr<T>() + (roi_batch_ind * channels + c) * height * width;
        for (int h = hstart; h < hend; ++h) {
            for (int w = wstart; w < wend; ++w) {
                int bottom_i = h * width + w;
                pooler.feed(feat_map_ptr[bottom_i], bottom_i);
            }
        }
        pooler.writeback_val(dst.ptr<T>()[i]);
        pooler.writeback_idx(index.ptr<dt_int32>()[i]);
    }
}

template <typename T>
void forward(
        _megdnn_tensor_in src, _megdnn_tensor_in rois, _megdnn_tensor_out dst,
        _megdnn_tensor_out index, const Param& param) {
    using namespace ::megdnn::roi_pooling;
    switch (param.mode) {
        case param::ROIPooling::Mode::MAX:
            forward_impl<T, MaxPooler<T>>(src, rois, dst, index, param.scale);
            break;
        case param::ROIPooling::Mode::AVERAGE:
            forward_impl<T, AveragePooler<T>>(src, rois, dst, index, param.scale);
            break;
        default:
            megdnn_assert_internal(false);
    }
}

template <typename T, typename BwdPooler>
void backward_impl(
        _megdnn_tensor_in diff, _megdnn_tensor_in src, _megdnn_tensor_in rois,
        _megdnn_tensor_in index, _megdnn_tensor_out grad, float scale) {
    auto batch_size = src.layout.shape[0], channels = src.layout.shape[1],
         height = src.layout.shape[2], width = src.layout.shape[3];
    auto total_nr_elem = batch_size * channels * height * width;
    auto num_rois = rois.layout.shape[0];
    auto spatial_scale = scale;

    auto pooled_height = diff.layout.shape[2], pooled_width = diff.layout.shape[3];
    for (size_t i = 0; i < total_nr_elem; ++i) {
        // (n, c, h, w) coords in bottom data
        int w = i % width;
        int h = (i / width) % height;
        int c = (i / width / height) % channels;
        int n = i / width / height / channels;

        T gradient = T(0);
        // Accumulate gradient over all ROIs that pooled this element
        for (size_t roi_n = 0; roi_n < num_rois; ++roi_n) {
            const T* offset_rois = rois.ptr<T>() + roi_n * 5;
            int roi_batch_ind = offset_rois[0];
            if (n != roi_batch_ind) {
                continue;
            }

            int roi_start_w = round(offset_rois[1] * spatial_scale);
            int roi_start_h = round(offset_rois[2] * spatial_scale);
            int roi_end_w = round(offset_rois[3] * spatial_scale);
            int roi_end_h = round(offset_rois[4] * spatial_scale);

            // Skip if ROI doesn't include (h, w)
            const bool in_roi =
                    (w >= roi_start_w && w <= roi_end_w && h >= roi_start_h &&
                     h <= roi_end_h);
            if (!in_roi) {
                continue;
            }

            int offset = (roi_n * channels + c) * pooled_height * pooled_width;
            const T* offset_src_diff = diff.ptr<T>() + offset;
            const int* offset_fp_idx = index.ptr<dt_int32>() + offset;

            // Compute feasible set of pooled units that could have pooled
            // this bottom unit

            // Force malformed ROIs to be 1x1
            int roi_width = std::max(roi_end_w - roi_start_w + 1, 1);
            int roi_height = std::max(roi_end_h - roi_start_h + 1, 1);

            float bin_size_h =
                    static_cast<float>(roi_height) / static_cast<float>(pooled_height);
            float bin_size_w =
                    static_cast<float>(roi_width) / static_cast<float>(pooled_width);

            int phstart = floor(static_cast<float>(h - roi_start_h) / bin_size_h);
            int phend = ceil(static_cast<float>(h - roi_start_h + 1) / bin_size_h);
            int pwstart = floor(static_cast<float>(w - roi_start_w) / bin_size_w);
            int pwend = ceil(static_cast<float>(w - roi_start_w + 1) / bin_size_w);

            phstart = std::min<int>(std::max(phstart, 0), pooled_height);
            phend = std::min<int>(std::max(phend, 0), pooled_height);
            pwstart = std::min<int>(std::max(pwstart, 0), pooled_width);
            pwend = std::min<int>(std::max(pwend, 0), pooled_width);

            for (int ph = phstart; ph < phend; ++ph) {
                for (int pw = pwstart; pw < pwend; ++pw) {
                    BwdPooler pooler;
                    pooler.update(
                            ph, pw, h, w, bin_size_h, bin_size_w, roi_start_h,
                            roi_start_w, pooled_height, pooled_width, height, width,
                            offset_src_diff, offset_fp_idx, gradient);
                }
            }
        }
        grad.ptr<T>()[i] = gradient;
    }
}

template <typename T>
void backward(
        _megdnn_tensor_in diff, _megdnn_tensor_in src, _megdnn_tensor_in rois,
        _megdnn_tensor_in index, _megdnn_tensor_out grad, const Param& param) {
    using namespace ::megdnn::roi_pooling;
    switch (param.mode) {
        case param::ROIPooling::Mode::MAX:
            backward_impl<T, BwdMaxPooler<T>>(
                    diff, src, rois, index, grad, param.scale);
            break;
        case param::ROIPooling::Mode::AVERAGE:
            backward_impl<T, BwdAveragePooler<T>>(
                    diff, src, rois, index, grad, param.scale);
            break;
        default:
            megdnn_assert_internal(false);
    }
}

}  // anonymous namespace

namespace megdnn {
namespace naive {

void ROIPoolingForwardImpl::exec(
        _megdnn_tensor_in src, _megdnn_tensor_in rois, _megdnn_tensor_out dst,
        _megdnn_tensor_out index, _megdnn_workspace workspace) {
    check_exec(src.layout, rois.layout, dst.layout, index.layout, workspace.size);
#define cb(DType)                                                                \
    if (src.layout.dtype == DType()) {                                           \
        MEGDNN_DISPATCH_CPU_KERN_OPR(forward<typename DTypeTrait<DType>::ctype>( \
                src, rois, dst, index, param()));                                \
        return;                                                                  \
    }
    MEGDNN_FOREACH_COMPUTING_DTYPE_FLOAT(cb)
#undef cb
    megdnn_throw("bad dtype");
}

void ROIPoolingBackwardImpl::exec(
        _megdnn_tensor_in diff, _megdnn_tensor_in src, _megdnn_tensor_in rois,
        _megdnn_tensor_in index, _megdnn_tensor_out grad, _megdnn_workspace workspace) {
    check_exec(
            diff.layout, src.layout, rois.layout, index.layout, grad.layout,
            workspace.size);
#define cb(DType)                                                                 \
    if (diff.layout.dtype == DType()) {                                           \
        MEGDNN_DISPATCH_CPU_KERN_OPR(backward<typename DTypeTrait<DType>::ctype>( \
                diff, src, rois, index, grad, param()));                          \
        return;                                                                   \
    }
    MEGDNN_FOREACH_COMPUTING_DTYPE_FLOAT(cb)
#undef cb
    megdnn_throw("bad dtype");
}

}  // namespace naive
}  // namespace megdnn

// vim: syntax=cpp.doxygen