sfzhang15 / FaceBoxes

FaceBoxes: A CPU Real-time Face Detector with High Accuracy, IJCB, 2017
Apache License 2.0
591 stars 171 forks source link

F0731 17:32:46.966609 19825 detection_output_layer.cpp:164] Check failed: num_priors_ * num_loc_classes_ * 4 == bottom[0]->channels() (4000 vs. 25600) Number of priors must match number of location predictions. #51

Closed TomaSaya closed 5 years ago

TomaSaya commented 5 years ago

Hello, i had successfully build the ssd caffe (https://github.com/weiliu89/caffe/tree/ssd), and i already download the pre-trained model "faceboxes.caffemodel". But when i run the following two codes, RCNN_Net.reset( new Net (model_dir + "/deploy.prototxt", TEST) ); RCNN_Net->CopyTrainedLayersFrom(model_dir + "/faceboxes.caffemodel"); I got the following error messages:

F0731 17:32:46.966609 19825 detection_output_layer.cpp:164] Check failed: numpriors num_locclasses 4 == bottom[0]->channels() (4000 vs. 25600) Number of priors must match number of location predictions.

do you have any idea, thanks

sfzhang15 commented 5 years ago

@TomaSaya We modify some layers of caffe-ssd to implement our method, you should use our caffe codes.