How to train and evaluate on UA dataset about sst HOT 14 OPEN

Also, if it is convenient for you to provide the UA pretrained weights?

Best,

from sst.

I find UA-DETRAC/Insight-MVT_Annotation_Train and UA-DETRAC/gt in your config.py, but don't know what exactly it is.

Hi, @ys1045097987 . the gt is generated by converter tool which convert the mat format ground truth data to xml format. Insight-MVT_Annotation_Train is the image folder provided by UA-DETRAC.

I suggest you to try mot17 firstly for you don't need to convert the format of the provided data.

from sst.

I already tried MOT17, it works fine. Thx for the info.

Best,

from sst.

Hi Shijie,

When I do the training on UA images, I need to parse an argument called config['ua_ignore_root'], how to generate the index file for that folder?

Best,

from sst.

Hi, @ys1045097987
The ua_ignore_root is provided by UA-DETRAC toolkit. Files in ua_ignore_root provide the information of detecting region.

from sst.

Hi Shijie@shijieS , I find it need sst300_0712_62580.pth to resume in UA. Could you provide a link of this weight for convinent ? Thanks a lot!

from sst.

Hi Shijie@shijieS , I find it need sst300_0712_62580.pth to resume in UA. Could you provide a link of this weight for convinent ? Thanks a lot!

you can write 'None' instead it

from sst.

HI
thank you for provide the code .
when i do the 'test' on UA,i can not find 'config['ua_detection_root'] = '/media/ssm/seagate/dataset/UA-DETRAC/EB' ' EB.how to get it?thank you a lot.

from sst.

hi，@shijieS,I want to run test_ua.py, can you give me the sst300_0712_62580.pth to resume in UA? [email protected],thank you very much.

from sst.

When I ran train_ua.py, after following your instructions, I ran into CUDA memory error #22 .
Can anybody help??

from sst.

Hi,When I tried to train the UA dataset, the screen display"None Processing" ,can you tell me how to deal with it?
best,

from sst.

Hello,
do you have pretrained weights for UA? Would appreciate if shared.

from sst.

anyone with an advanced checkpoint for ua_training weights? would appreciate if shared.

from sst.

Hello，I have a question about your code.I run it on MOT17 .When I submit the result to the MOT Challenge，the MOTA is 44.5 not 52.4224.So I do not know what is wrong with it.Could you tell me？Thanks

from sst.