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

/**
 * \file src/opr/test/dnn/local.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 "./legacy_checker.h"

#include "megbrain/opr/dnn/local.h"
#include "megbrain/test/autocheck.h"

using namespace mgb;

namespace {

using Param = opr::Local::Param;
using Mode = Param::Mode;

Mode modes_to_check[] = {Mode::CONVOLUTION, Mode::CROSS_CORRELATION};

void local_brute(
        const std::vector<std::shared_ptr<HostTensorND>>& in_tensor,
        std::shared_ptr<HostTensorND>& out_tensor, const Param& param) {
    ASSERT_EQ(2u, in_tensor.size());
    auto in = in_tensor[0], filter = in_tensor[1];
    ASSERT_EQ(4u, in->shape().ndim);
    ASSERT_EQ(6u, filter->shape().ndim);

    int batch_size = in->shape().shape[0];
    int ic = in->shape().shape[1];
    int ih = in->shape().shape[2];
    int iw = in->shape().shape[3];

    int fh = filter->shape().shape[3];
    int fw = filter->shape().shape[4];

    int ph = param.pad_h;
    int pw = param.pad_w;

    int sh = param.stride_h;
    int sw = param.stride_w;
    ASSERT_GE(ih + 2 * ph, fh);
    ASSERT_GE(iw + 2 * pw, fw);
    int oh = (ih + 2 * ph - fh) / sh + 1;
    int ow = (iw + 2 * pw - fw) / sw + 1;
    ASSERT_EQ(static_cast<size_t>(ic), filter->shape().shape[2]);
    int oc = filter->shape().shape[5];

    out_tensor = std::make_shared<HostTensorND>(
            CompNode::load("xpu0"),
            TensorShape{
                    static_cast<size_t>(batch_size), static_cast<size_t>(oc),
                    static_cast<size_t>(oh), static_cast<size_t>(ow)});

    int pn, poc, poh, pow, pih, piw, pic, pfh, pfw;
    for (pn = 0; pn < batch_size; ++pn)
        for (poc = 0; poc < oc; ++poc)
            for (poh = 0, pih = -ph; poh < oh; ++poh, pih += sh)
                for (pow = 0, piw = -pw; pow < ow; ++pow, piw += sw) {
                    float& target = out_tensor->ptr<float>(
                            {static_cast<size_t>(pn), static_cast<size_t>(poc),
                             static_cast<size_t>(poh), static_cast<size_t>(pow)})[0];
                    target = 0;
                    for (pic = 0; pic < ic; ++pic)
                        for (pfh = 0; pfh < fh; ++pfh)
                            for (pfw = 0; pfw < fw; ++pfw) {
                                int prih, priw;
                                float img_data, filter_data;
                                if (param.mode == Param::Mode::CONVOLUTION) {
                                    prih = pih + fh - pfh - 1;
                                    priw = piw + fw - pfw - 1;
                                } else {
                                    mgb_assert(
                                            param.mode ==
                                            Param::Mode::CROSS_CORRELATION);
                                    prih = pih + pfh;
                                    priw = piw + pfw;
                                }
                                if (prih >= 0 && prih < ih && priw >= 0 && priw < iw) {
                                    img_data = in_tensor[0]->ptr<float>(
                                            {static_cast<size_t>(pn),
                                             static_cast<size_t>(pic),
                                             static_cast<size_t>(prih),
                                             static_cast<size_t>(priw)})[0];
                                } else {
                                    img_data = 0;
                                }
                                filter_data = filter->ptr<float>(
                                        {static_cast<size_t>(poh),
                                         static_cast<size_t>(pow),
                                         static_cast<size_t>(pic),
                                         static_cast<size_t>(pfh),
                                         static_cast<size_t>(pfw),
                                         static_cast<size_t>(poc)})[0];
                                target += img_data * filter_data;
                            }
                }
}

}  // anonymous namespace

TEST(TestOprDNN, LocalForward) {
    uint32_t ih = 10, ic = 16, oc = 32, ph = 0, sh = 1, fh = 2;
    for (auto mode : modes_to_check) {
        uint32_t iw = ih, fw = fh, pw = ph, sw = sh;
        uint32_t oh = (ih + 2 * ph - fh) / sh + 1, ow = (iw + 2 * pw - fw) / sw + 1;
        Param param{mode, ph, pw, sh, sw};
        size_t batch_size = 32;
        opr::test::ForwardChecker<opr::Local, 2> forward_checker(
                {{batch_size, ic, ih, iw}, {oh, ow, ic, fh, fw, oc}}, local_brute,
                param);
        forward_checker.run();
    }
}

TEST(TestOprDNN, LocalBackward) {
    uint32_t ih = 10, ic = 16, oc = 32, ph = 0, sh = 1, fh = 2;
    uint32_t iw = ih, fw = fh, pw = ph, sw = sh;
    uint32_t oh = (ih + 2 * ph - fh) / sh + 1, ow = (iw + 2 * pw - fw) / sw + 1;
    Param param{Mode::CROSS_CORRELATION, ph, pw, sh, sw};
    size_t batch_size = 32;
    opr::test::BackwardChecker<opr::Local, 2> backward_checker(
            {{batch_size, ic, ih, iw}, {oh, ow, ic, fh, fw, oc}}, param, 1e-2, 1);
    backward_checker.run();
}

TEST(TestOprDNN, GroupLocal) {
    using Checker = AutoOprChecker<2, 1>;
    opr::GroupLocal::Param param;
    auto make_graph = [&](const Checker::SymInpArray& inputs) -> Checker::SymOutArray {
        return {opr::GroupLocal::make(inputs[0], inputs[1], param)};
    };

    auto fwd = [&](Checker::NumOutArray& dest, Checker::NumInpArray inp) {
        auto&& out = dest[0];
        auto inp0 = std::make_shared<HostTensorND>(),
             inp1 = std::make_shared<HostTensorND>();
        auto sl = inp[0]->layout(), fl = inp[1]->layout().remove_axis(0);
        TensorLayout ol;
        auto group = inp[1]->layout()[0];
        sl.shape[1] /= group;
        for (size_t i = 0; i < group; ++i) {
            inp0->copy_from(inp[0]->sub(SubTensorSpec::make_from_offset_elem(
                    sl, i * sl[1] * sl[2] * sl[3])));
            inp1->copy_from(inp[1]->sub(
                    SubTensorSpec::make_from_offset_elem(fl, i * fl.total_nr_elems())));
            std::shared_ptr<HostTensorND> cur_out;
            local_brute({inp0, inp1}, cur_out, {});
            if (!i) {
                auto oshp = cur_out->shape();
                oshp[1] *= group;
                out.resize(oshp);
                ol = out.layout();
                ol[1] /= group;
            }
            out.sub(SubTensorSpec::make_from_offset_elem(ol, i * ol[1] * ol[2] * ol[3]))
                    .copy_from_fixlayout(*cur_out);
        }
    };

    Checker::RunOptions opt;
    opt.numdiff_eps = 1;
    opt.outputs_max_err = 5e-5;
    Checker checker{make_graph, fwd};
    auto run = [&](const TensorShape& ishp, size_t fs, size_t oc, size_t group) {
        size_t ic = ishp[1], ih = ishp[2], iw = ishp[3];
        TensorShape flt{group, ih - fs + 1, iw - fs + 1, ic / group,
                        fs,    fs,          oc / group};
        checker.run({ishp, flt}, opt);
    };
    run({32, 9, 2, 2}, 1, 96, 3);
    run({32, 4, 2, 3}, 2, 32, 2);
    run({32, 3, 4, 3}, 3, 16, 1);
}

// vim: syntax=cpp.doxygen foldmethod=marker foldmarker=f{{{,f}}}