diff --git a/examples/dnn_instance_segmentation_train_ex.cpp b/examples/dnn_instance_segmentation_train_ex.cpp index 4fbe33df6..236ca22f5 100644 --- a/examples/dnn_instance_segmentation_train_ex.cpp +++ b/examples/dnn_instance_segmentation_train_ex.cpp @@ -155,6 +155,8 @@ std::vector rgb_label_image_to_mmod_rects( // Encountered a new instance instance_indexes[rgb_label] = mmod_rects.size(); mmod_rects.emplace_back(dlib::rectangle(c, r, c, r)); + + // TODO: read the instance's class from the other png! } else { @@ -460,7 +462,7 @@ std::vector> load_all_mmod_rects(const std::vector< int main(int argc, char** argv) try { - if (argc < 2 || argc > 3) + if (argc < 2 || argc > 4) { cout << "To run this program you need a copy of the PASCAL VOC2012 dataset." << endl; cout << endl; @@ -480,7 +482,7 @@ int main(int argc, char** argv) try } // mini-batches smaller than the default can be used with GPUs having less memory - const unsigned int det_minibatch_size = argc >= 3 ? std::stoi(argv[2]) : 87; + const unsigned int det_minibatch_size = argc >= 3 ? std::stoi(argv[2]) : 75; const unsigned int seg_minibatch_size = argc >= 4 ? std::stoi(argv[3]) : 25; cout << "det mini-batch size: " << det_minibatch_size << endl; cout << "seg mini-batch size: " << seg_minibatch_size << endl;