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
|
// Copyright (C) 2018 Intel Corporation
//
// SPDX-License-Identifier: Apache-2.0
//
#include "mean_image.h"
#include "ie_parallel.hpp"
using namespace MKLDNNPlugin;
using namespace InferenceEngine;
MeanImage::MeanImage() : meanBuffer(nullptr) {
}
void MeanImage::Load(const MKLDNNDims& inputDims, InputInfo::Ptr inputInfo) {
PreProcessInfo &pp = inputInfo->getPreProcess();
size_t inChannels = pp.getNumberOfChannels();
if (inChannels == 0) {
meanBuffer = nullptr;
return;
}
if (inChannels != inputDims[1]) {
THROW_IE_EXCEPTION << "channels mismatch between mean and input";
}
ResponseDesc resp;
switch (pp.getMeanVariant()) {
case MEAN_VALUE: {
// mean image common value per channel (1x1xC)
meanValues.resize(inChannels);
for (unsigned channel = 0; channel < inChannels; channel++) {
meanValues[channel] = pp[channel]->meanValue;
}
}
break;
case MEAN_IMAGE: {
// since MKLDNN expects all channels in the same buffer - we copy it here as it comes from different channels...
auto meanWidth = pp[0]->meanData->dims()[0];
auto meanHeight = pp[0]->meanData->dims()[1];
meanBuffer = make_shared_blob<float>(Precision::FP32, CHW, { meanWidth, meanHeight, inChannels });
meanBuffer->allocate();
for (unsigned channel = 0; channel < inChannels; channel++) {
Blob::Ptr meanBlob = pp[channel]->meanData;
if (!meanBlob || meanBlob->precision() != Precision::FP32)
THROW_IE_EXCEPTION << "mean image not provided or not in Float 32";
if (meanBlob->size() != meanHeight*meanWidth) {
THROW_IE_EXCEPTION << "mean image size does not match expected network input, expecting " << meanWidth << " x " << meanHeight;
}
// todo: cast to TBlob and make sure it is floats
memcpy(meanBuffer->data() + channel*meanBlob->size(), meanBlob->buffer(), meanBlob->byteSize());
}
}
break;
case NONE: {
// there is no mean image. So disable mean image step
meanBuffer = nullptr;
}
break;
default: {
THROW_IE_EXCEPTION << "Unsupported mean variant: " << pp.getMeanVariant();
}
}
}
void MeanImage::Subtract(const MKLDNNDims &inputDims, float *input) {
IE_ASSERT(input != nullptr);
if (inputDims.ndims() != 4) {
THROW_IE_EXCEPTION << "Expecting input as 4 dimension blob with format NxCxHxW.";
}
int MB = inputDims[0];
int srcSize = inputDims.size() / MB;
if (meanBuffer && meanBuffer->size()) {
const float * meanBufferValues = meanBuffer->readOnly();
parallel_for2d(MB, srcSize, [&](int mb, int i) {
input[srcSize * mb + i] -= meanBufferValues[i];
});
} else if (!meanValues.empty()) {
int C = inputDims[1];
srcSize /= inputDims[1];
parallel_for3d(MB, C, srcSize, [&](int mb, int c, int i) {
input[srcSize * mb * C + c * srcSize + i] -= meanValues[c];
});
}
}
|
