KimBomm

@cha-zhang Actually.. I don't know exactly. I just use [Examples/Image/Detection/FasterRCNN/run_faster_rcnn.py](https://github.com/Microsoft/CNTK/blob/master/Examples/Image/Detection/FasterRCNN/run_faster_rcnn.py) to train.

Thanks, darknet53.cfg is sames as yolov3.cfg(diffrent # of classes). I changed `num` to 3. and It works. But it shows Division by Zero at `yolomodel.py` line 108. So i fix...

I have a same issue. The error is shown in second yolo layer. Did you solve this problem?

Hi @abeyang00 Here! [https://github.com/AlexeyAB/darknet](9https://github.com/AlexeyAB/darknet). He made loss plot for training.

Hi @lehuiwu Whatever you want. opencv doesn't required for build library. But, you need opencv for run or test.

Oh I see. I think you are testing on Windows OS. And, I used OpenCV 3.4.1. Additionally, That error will appear when the library settings are incorrect. How about installing...

This is because the corresponding dll is for release mode only. MSVC's dll calling is diffrent with gcc. So, you need build again for make dll for debug mode. Change...

I built it with Debug x64. cudnn64_7.dll exists by version of CUDA. Therefore, if possible, it is recommended that this file be in a directory such as libYOLOV3SE.dll. There is...

Hi @shkipper12. Simple, Do not use CUDA 9.1 and 9.2, It has unknown error. YOLO works on CUDA 8.0, 9.0 and 10.0.

Hi @sushant1727, I think you got this project as zip(not git clone). But, I uploaded all the dll files as Git LFS. So you have to download as `git clone`...