many object detecrtion

[itkt]

Hi, Thanks for work. Very fast work. But after starting the program a lot of objects detection (demo data): allocate output 0 , 0 , 0.126316 , 0.126316 0.00823669 , 0 , 0.149658 , 0.167737 0.0315789 , 0 , 0.231579 , 0.126316 0.1135 , 0 , 0.254921 , 0.167737 0.136842 , 0 , 0.336842 , 0.126316 0.218763 , 0 , 0.360184 , 0.167737 0.242105 , 0 , 0.442105 , 0.126316 0.324026 , 0 , 0.465447 , 0.167737 0.347368 , 0 , 0.547368 , 0.126316 0.429289 , 0 , 0.570711 , 0.167737 0.452632 , 0 , 0.652632 , 0.126316 0.534552 , 0 , 0.675974 , 0.167737 0.557895 , 0 , 0.757895 , 0.126316 0.639816 , 0 , 0.781237 , 0.167737 0.663158 , 0 , 0.863158 , 0.126316 0.745079 , 0 , 0.8865 , 0.167737 0.768421 , 0 , 0.968421 , 0.126316 0.850342 , 0 , 0.991763 , 0.167737 0.873684 , 0 , 1 , 0.126316 0 , 0.00823669 , 0.167737 , 0.149658 0.0427892 , 0.00823669 , 0.325632 , 0.149658 0.200684 , 0.00823669 , 0.483527 , 0.149658 0.358579 , 0.00823669 , 0.641421 , 0.149658 0.516473 , 0.00823669 , 0.799316 , 0.149658 0 , 0 , 0.126316 , 0.126316 0.00823669 , 0 , 0.149658 , 0.167737 0.0315789 , 0 , 0.231579 , 0.126316 0.1135 , 0 , 0.254921 , 0.167737 0.136842 , 0 , 0.336842 , 0.126316 0.218763 , 0 , 0.360184 , 0.167737 0.242105 , 0 , 0.442105 , 0.126316 0.324026 , 0 , 0.465447 , 0.167737 0.347368 , 0 , 0.547368 , 0.126316 0.429289 , 0 , 0.570711 , 0.167737 0.452632 , 0 , 0.652632 , 0.126316 0.534552 , 0 , 0.675974 , 0.167737 0.557895 , 0 , 0.757895 , 0.126316 0.639816 , 0 , 0.781237 , 0.167737 0.663158 , 0 , 0.863158 , 0.126316 0.745079 , 0 , 0.8865 , 0.167737 0.768421 , 0 , 0.968421 , 0.126316 0.850342 , 0 , 0.991763 , 0.167737 0.873684 , 0 , 1 , 0.126316 0 , 0.00823669 , 0.167737 , 0.149658 0.0427892 , 0.00823669 , 0.325632 , 0.149658 0.200684 , 0.00823669 , 0.483527 , 0.149658 0.358579 , 0.00823669 , 0.641421 , 0.149658 0.516473 , 0.00823669 , 0.799316 , 0.149658 0 , 0 , 0.126316 , 0.126316 0.00823669 , 0 , 0.149658 , 0.167737 0.0315789 , 0 , 0.231579 , 0.126316 0.1135 , 0 , 0.254921 , 0.167737 0.136842 , 0 , 0.336842 , 0.126316 0.218763 , 0 , 0.360184 , 0.167737 0.242105 , 0 , 0.442105 , 0.126316 0.324026 , 0 , 0.465447 , 0.167737 0.347368 , 0 , 0.547368 , 0.126316 0.429289 , 0 , 0.570711 , 0.167737 0.452632 , 0 , 0.652632 , 0.126316 0.534552 , 0 , 0.675974 , 0.167737 0.557895 , 0 , 0.757895 , 0.126316 0.639816 , 0 , 0.781237 , 0.167737 0.663158 , 0 , 0.863158 , 0.126316 0.745079 , 0 , 0.8865 , 0.167737 0.768421 , 0 , 0.968421 , 0.126316 0.850342 , 0 , 0.991763 , 0.167737 0.873684 , 0 , 1 , 0.126316 0 , 0.00823669 , 0.167737 , 0.149658 0.0427892 , 0.00823669 , 0.325632 , 0.149658 0.200684 , 0.00823669 , 0.483527 , 0.149658 0.358579 , 0.00823669 , 0.641421 , 0.149658 0.516473 , 0.00823669 , 0.799316 , 0.149658 0 , 0 , 0.126316 , 0.126316 0.00823669 , 0 , 0.149658 , 0.167737 0.0315789 , 0 , 0.231579 , 0.126316 0.1135 , 0 , 0.254921 , 0.167737 0.136842 , 0 , 0.336842 , 0.126316 0.218763 , 0 , 0.360184 , 0.167737 0.242105 , 0 , 0.442105 , 0.126316 0.324026 , 0 , 0.465447 , 0.167737 0.347368 , 0 , 0.547368 , 0.126316 0.429289 , 0 , 0.570711 , 0.167737 0.452632 , 0 , 0.652632 , 0.126316 0.534552 , 0 , 0.675974 , 0.167737 0.557895 , 0 , 0.757895 , 0.126316 0.639816 , 0 , 0.781237 , 0.167737 0.663158 , 0 , 0.863158 , 0.126316 0.745079 , 0 , 0.8865 , 0.167737 0.768421 , 0 , 0.968421 , 0.126316 0.850342 , 0 , 0.991763 , 0.167737 0.873684 , 0 , 1 , 0.126316 0 , 0.00823669 , 0.167737 , 0.149658 0.0427892 , 0.00823669 , 0.325632 , 0.149658 0.200684 , 0.00823669 , 0.483527 , 0.149658 0.358579 , 0.00823669 , 0.641421 , 0.149658 0.516473 , 0.00823669 , 0.799316 , 0.149658

[PiyalGeorge]

@itkt , What all changes did you do in pluginImplement.cpp and pluginImplement.h?

[itkt]

I do not change this file

[PiyalGeorge]

@itkt , thanks for the reply. did you try with a custom caffemodel or the same model in this repo?

[itkt]

I try with model in this repo. And now I try with a custom model (trained Mobilenet ssd), and have a problem CaffeParser: ERROR: Attempting to access NULL weights Weights for layer conv1/dw doesn't exist and this message for all layer

[PiyalGeorge]

I trained mobilenet-ssd model in caffe following this . This guy @Ghustwb, has mentioned in the ReadMe that he has trained his model(see his model in model folder in the repo) with 5 classes. So i also trained my mobilenet-ssd model with 5 classes(including background). then i used that 5 class model and tried that worked successfully. Try that. But currently i'm trying to do the same for 7 classes. the Ghustwb has specified some method here . But i think its not working.

[itkt]

Yes, We also trained model following tutorial. But we need only 2 class. After training we loaded weights into this project. Then start ./mobilenet have error. Can you help me start the application with model trained to 2 classes