Zhongdao / Towards-Realtime-MOT

Joint Detection and Embedding for fast multi-object tracking
MIT License
2.39k stars 539 forks source link

TypeError: test() got an unexpected keyword argument 'img_size' #147

Open junjieinkit opened 4 years ago

junjieinkit commented 4 years ago

https://github.com/Zhongdao/Towards-Realtime-MOT/blob/c9dd7ee438c05ef0ecdc30621a9e9789d417983d/train.py#L175 hello there, an error happens here, did someone know how to fix it?Just delete the unexpected keyword(hier image_size and nID) in the test() function? Thanks in advance

jwnirvana commented 4 years ago

I miss the same problem,it may run to delete the parameter,but the test input size should influence the inference time

jwnirvana commented 4 years ago

There should be a prolem, the img_size show in the table in reademe could be change. How it does?

junjieinkit commented 4 years ago

There should be a prolem, the img_size show in the table in reademe could be change. How it does? https://github.com/Zhongdao/Towards-Realtime-MOT/blob/c9dd7ee438c05ef0ecdc30621a9e9789d417983d/train.py#L187 you can change different config file here such as yolov3_1088x608.cfg or yolov3_576x320.cfg in cfg folder or simplely change the image_size (height or width) in one of the .cfg file like https://github.com/Zhongdao/Towards-Realtime-MOT/blob/c9dd7ee438c05ef0ecdc30621a9e9789d417983d/cfg/yolov3_1088x608.cfg#L4

kanghu-bo commented 4 years ago

how to solve it?

NaifahNurya commented 4 years ago

Got the same error at the last iteration of the first epoch after saving trained weight latest.pt; Any update on how to solve this.

_2020-06-10 11:31:24 [INFO]: 0/29 4000/4036 2.3 0.153 4.75 2.73 39.5 0.428 Traceback (most recent call last): File "train.py", line 221, in opt=opt, File "train.py", line 177, in train print_interval=40, nID=dataset.nID) TypeError: test() got an unexpected keyword argument 'imgsize'

xwwxyd commented 4 years ago

Did you write the correct test data path in your ccmcpe.json file?I have the same problem before,I solve the problem by writting the correct test data in my ccmpe.json like this: image

tcexeexe commented 3 years ago

l meet the same question too,does anyone know how to solve it?