You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, it is a wonderful work!. I'm freshman to the yolo detector, and I encountered some problems while simply using this yolo c++ wrapper.
I complied the source code and no error occurred. Then put my custom data near the darknet, which included *.cfg, *weights, *.name, *.data. However, when I run following command:
./darknet detector test cfg_zlt/zlt.data cfg_zlt/yolo-zlt.cfg backup/yolo-zlt_900.weights data/nfpa/pos-123.jpg
No detection result outputted. It seemed the darknet didn't detect anything. But in fact, I got a correct detection results when used the darcknet compiled from original code.
In addition, I tried to train my custom dataset using your darknet, and got strange outputs like this:
memset begins
memset ends
for bbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for finished
Region Avg IOU: 0.171883, Class: 1.000000, Obj: 0.001114, No Obj: 0.003235, Avg Recall: 0.250000, count: 8
It crashed after several epochs and the error message was given:
CUDA Error: invalid configuration argument
darknet: /home/zlt/packages/yolo_cpp/src/cuda.cpp:36: void check_error(cudaError_t): Assertion `0' failed.
Aborted (core dumped)
Any suggestions?
The text was updated successfully, but these errors were encountered:
Hi, it is a wonderful work!. I'm freshman to the yolo detector, and I encountered some problems while simply using this yolo c++ wrapper.
I complied the source code and no error occurred. Then put my custom data near the darknet, which included *.cfg, *weights, *.name, *.data. However, when I run following command:
./darknet detector test cfg_zlt/zlt.data cfg_zlt/yolo-zlt.cfg backup/yolo-zlt_900.weights data/nfpa/pos-123.jpg
No detection result outputted. It seemed the darknet didn't detect anything. But in fact, I got a correct detection results when used the darcknet compiled from original code.
In addition, I tried to train my custom dataset using your darknet, and got strange outputs like this:
memset begins
memset ends
for bbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for jjjj
for bbbbb
for finished
Region Avg IOU: 0.171883, Class: 1.000000, Obj: 0.001114, No Obj: 0.003235, Avg Recall: 0.250000, count: 8
It crashed after several epochs and the error message was given:
CUDA Error: invalid configuration argument
darknet: /home/zlt/packages/yolo_cpp/src/cuda.cpp:36: void check_error(cudaError_t): Assertion `0' failed.
Aborted (core dumped)
Any suggestions?
The text was updated successfully, but these errors were encountered: