|
|
|
@ -159,7 +159,6 @@ bool LogisticRegression::train(cv::InputArray data_ip, cv::InputArray labels_ip) |
|
|
|
|
if(num_classes == 2) |
|
|
|
|
{ |
|
|
|
|
labels_l.convertTo(labels, CV_32F); |
|
|
|
|
//currently supported training methods LogisticRegression::BATCH and LogisticRegression::MINI_BATCH
|
|
|
|
|
if(this->params.train_method == LogisticRegression::BATCH) |
|
|
|
|
new_theta = compute_batch_gradient(data_t, labels, init_theta); |
|
|
|
|
else |
|
|
|
@ -176,7 +175,6 @@ bool LogisticRegression::train(cv::InputArray data_ip, cv::InputArray labels_ip) |
|
|
|
|
{ |
|
|
|
|
new_local_labels = (labels_l == it->second)/255; |
|
|
|
|
new_local_labels.convertTo(labels, CV_32F); |
|
|
|
|
// currently supported training methods LogisticRegression::BATCH and LogisticRegression::MINI_BATCH
|
|
|
|
|
if(this->params.train_method == LogisticRegression::BATCH) |
|
|
|
|
new_theta = compute_batch_gradient(data_t, labels, init_theta); |
|
|
|
|
else |
|
|
|
|