mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2025-06-12 12:22:51 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge pull request #9310 from r2d3:svm_parallel

Browse Source
This commit is contained in:
Vadim Pisarevsky 2017-09-18 09:32:41 +00:00
commit 70c5d69640
1 changed files with 116 additions and 65 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

181
modules/ml/src/svm.cpp
View File

@ -1636,6 +1636,99 @@ public:
return true; return true;
} }
class TrainAutoBody : public ParallelLoopBody
{
public:
TrainAutoBody(const vector<SvmParams>& _parameters,
const cv::Mat& _samples,
const cv::Mat& _responses,
const cv::Mat& _labels,
const vector<int>& _sidx,
bool _is_classification,
int _k_fold,
std::vector<double>& _result) :
parameters(_parameters), samples(_samples), responses(_responses), labels(_labels),
sidx(_sidx), is_classification(_is_classification), k_fold(_k_fold), result(_result)
{}
void operator()( const cv::Range& range ) const
{
int sample_count = samples.rows;
int var_count_ = samples.cols;
size_t sample_size = var_count_*samples.elemSize();
int test_sample_count = (sample_count + k_fold/2)/k_fold;
int train_sample_count = sample_count - test_sample_count;
// Use a local instance
cv::Ptr<SVMImpl> svm = makePtr<SVMImpl>();
svm->class_labels = labels;
int rtype = responses.type();
Mat temp_train_samples(train_sample_count, var_count_, CV_32F);
Mat temp_test_samples(test_sample_count, var_count_, CV_32F);
Mat temp_train_responses(train_sample_count, 1, rtype);
Mat temp_test_responses;
for( int p = range.start; p < range.end; p++ )
{
svm->setParams(parameters[p]);
double error = 0;
for( int k = 0; k < k_fold; k++ )
{
int start = (k*sample_count + k_fold/2)/k_fold;
for( int i = 0; i < train_sample_count; i++ )
{
int j = sidx[(i+start)%sample_count];
memcpy(temp_train_samples.ptr(i), samples.ptr(j), sample_size);
if( is_classification )
temp_train_responses.at<int>(i) = responses.at<int>(j);
else if( !responses.empty() )
temp_train_responses.at<float>(i) = responses.at<float>(j);
}
// Train SVM on <train_size> samples
if( !svm->do_train( temp_train_samples, temp_train_responses ))
continue;
for( int i = 0; i < test_sample_count; i++ )
{
int j = sidx[(i+start+train_sample_count) % sample_count];
memcpy(temp_test_samples.ptr(i), samples.ptr(j), sample_size);
}
svm->predict(temp_test_samples, temp_test_responses, 0);
for( int i = 0; i < test_sample_count; i++ )
{
float val = temp_test_responses.at<float>(i);
int j = sidx[(i+start+train_sample_count) % sample_count];
if( is_classification )
error += (float)(val != responses.at<int>(j));
else
{
val -= responses.at<float>(j);
error += val*val;
}
}
}
result[p] = error;
}
}
private:
const vector<SvmParams>& parameters;
const cv::Mat& samples;
const cv::Mat& responses;
const cv::Mat& labels;
const vector<int>& sidx;
bool is_classification;
int k_fold;
std::vector<double>& result;
};
bool trainAuto( const Ptr<TrainData>& data, int k_fold, bool trainAuto( const Ptr<TrainData>& data, int k_fold,
ParamGrid C_grid, ParamGrid gamma_grid, ParamGrid p_grid, ParamGrid C_grid, ParamGrid gamma_grid, ParamGrid p_grid,
ParamGrid nu_grid, ParamGrid coef_grid, ParamGrid degree_grid, ParamGrid nu_grid, ParamGrid coef_grid, ParamGrid degree_grid,
@ -1713,15 +1806,12 @@ public:
int sample_count = samples.rows; int sample_count = samples.rows;
var_count = samples.cols; var_count = samples.cols;
size_t sample_size = var_count*samples.elemSize();
vector<int> sidx; vector<int> sidx;
setRangeVector(sidx, sample_count); setRangeVector(sidx, sample_count);
int i, j, k;
// randomly permute training samples // randomly permute training samples
for( i = 0; i < sample_count; i++ ) for( int i = 0; i < sample_count; i++ )
{ {
int i1 = rng.uniform(0, sample_count); int i1 = rng.uniform(0, sample_count);
int i2 = rng.uniform(0, sample_count); int i2 = rng.uniform(0, sample_count);
@ -1735,7 +1825,7 @@ public:
// between the k_fold parts. // between the k_fold parts.
vector<int> sidx0, sidx1; vector<int> sidx0, sidx1;
for( i = 0; i < sample_count; i++ ) for( int i = 0; i < sample_count; i++ )
{ {
if( responses.at<int>(sidx[i]) == 0 ) if( responses.at<int>(sidx[i]) == 0 )
sidx0.push_back(sidx[i]); sidx0.push_back(sidx[i]);
@ -1746,15 +1836,15 @@ public:
int n0 = (int)sidx0.size(), n1 = (int)sidx1.size(); int n0 = (int)sidx0.size(), n1 = (int)sidx1.size();
int a0 = 0, a1 = 0; int a0 = 0, a1 = 0;
sidx.clear(); sidx.clear();
for( k = 0; k < k_fold; k++ ) for( int k = 0; k < k_fold; k++ )
{ {
int b0 = ((k+1)*n0 + k_fold/2)/k_fold, b1 = ((k+1)*n1 + k_fold/2)/k_fold; int b0 = ((k+1)*n0 + k_fold/2)/k_fold, b1 = ((k+1)*n1 + k_fold/2)/k_fold;
int a = (int)sidx.size(), b = a + (b0 - a0) + (b1 - a1); int a = (int)sidx.size(), b = a + (b0 - a0) + (b1 - a1);
for( i = a0; i < b0; i++ ) for( int i = a0; i < b0; i++ )
sidx.push_back(sidx0[i]); sidx.push_back(sidx0[i]);
for( i = a1; i < b1; i++ ) for( int i = a1; i < b1; i++ )
sidx.push_back(sidx1[i]); sidx.push_back(sidx1[i]);
for( i = 0; i < (b - a); i++ ) for( int i = 0; i < (b - a); i++ )
{ {
int i1 = rng.uniform(a, b); int i1 = rng.uniform(a, b);
int i2 = rng.uniform(a, b); int i2 = rng.uniform(a, b);
@ -1764,23 +1854,12 @@ public:
} }
} }
int test_sample_count = (sample_count + k_fold/2)/k_fold;
int train_sample_count = sample_count - test_sample_count;
SvmParams best_params = params;
double min_error = FLT_MAX;
int rtype = responses.type();
Mat temp_train_samples(train_sample_count, var_count, CV_32F);
Mat temp_test_samples(test_sample_count, var_count, CV_32F);
Mat temp_train_responses(train_sample_count, 1, rtype);
Mat temp_test_responses;
// If grid.minVal == grid.maxVal, this will allow one and only one pass through the loop with params.var = grid.minVal. // If grid.minVal == grid.maxVal, this will allow one and only one pass through the loop with params.var = grid.minVal.
#define FOR_IN_GRID(var, grid) \ #define FOR_IN_GRID(var, grid) \
for( params.var = grid.minVal; params.var == grid.minVal || params.var < grid.maxVal; params.var = (grid.minVal == grid.maxVal) ? grid.maxVal + 1 : params.var * grid.logStep ) for( params.var = grid.minVal; params.var == grid.minVal || params.var < grid.maxVal; params.var = (grid.minVal == grid.maxVal) ? grid.maxVal + 1 : params.var * grid.logStep )
// Create the list of parameters to test
std::vector<SvmParams> parameters;
FOR_IN_GRID(C, C_grid) FOR_IN_GRID(C, C_grid)
FOR_IN_GRID(gamma, gamma_grid) FOR_IN_GRID(gamma, gamma_grid)
FOR_IN_GRID(p, p_grid) FOR_IN_GRID(p, p_grid)
@ -1788,51 +1867,23 @@ public:
FOR_IN_GRID(coef0, coef_grid) FOR_IN_GRID(coef0, coef_grid)
FOR_IN_GRID(degree, degree_grid) FOR_IN_GRID(degree, degree_grid)
{ {
// make sure we updated the kernel and other parameters parameters.push_back(params);
setParams(params); }
double error = 0; std::vector<double> result(parameters.size());
for( k = 0; k < k_fold; k++ ) TrainAutoBody invoker(parameters, samples, responses, class_labels, sidx,
is_classification, k_fold, result);
parallel_for_(cv::Range(0,(int)parameters.size()), invoker);
// Extract the best parameters
SvmParams best_params = params;
double min_error = FLT_MAX;
for( int i = 0; i < (int)result.size(); i++ )
{
if( result[i] < min_error )
{ {
int start = (k*sample_count + k_fold/2)/k_fold; min_error = result[i];
for( i = 0; i < train_sample_count; i++ ) best_params = parameters[i];
{
j = sidx[(i+start)%sample_count];
memcpy(temp_train_samples.ptr(i), samples.ptr(j), sample_size);
if( is_classification )
temp_train_responses.at<int>(i) = responses.at<int>(j);
else if( !responses.empty() )
temp_train_responses.at<float>(i) = responses.at<float>(j);
}
// Train SVM on <train_size> samples
if( !do_train( temp_train_samples, temp_train_responses ))
continue;
for( i = 0; i < test_sample_count; i++ )
{
j = sidx[(i+start+train_sample_count) % sample_count];
memcpy(temp_test_samples.ptr(i), samples.ptr(j), sample_size);
}
predict(temp_test_samples, temp_test_responses, 0);
for( i = 0; i < test_sample_count; i++ )
{
float val = temp_test_responses.at<float>(i);
j = sidx[(i+start+train_sample_count) % sample_count];
if( is_classification )
error += (float)(val != responses.at<int>(j));
else
{
val -= responses.at<float>(j);
error += val*val;
}
}
}
if( min_error > error )
{
min_error = error;
best_params = params;
} }
} }