search

CaptainEven / MCMOT

Real time one-stage multi-class & multi-object tracking based on anchor-free detection and ReID
MIT License
383 stars 82 forks source link

How to evaluate on MCMOT dataset？ #66

Closed StrongerzZ closed 3 years ago

StrongerzZ commented 3 years ago

Hi, thanks for the wonderful work!

I have been able to train correctly in VisDrone dataset and generate tracking results.

But there are some problems when I run the test_det.py, test_emb.py and track.py, I tried to fix them but failed. It seems that the above python scripts are not able to perform multi-category evaluation well. (1) Have CUDA out of memory error in test_emb.py: torch.mm (line 102). (2) Shape missmatch in test_det.py: in function mot_decode(I try to fix it but don't know if it is correct) and line 32 (3) track.py does not run properly in the VisDrone dataset. Can someone help me？Thanks a lot!

I also want to ask if MOT metrics can only be evaluated in the MOT dataset? (Just like the COCO dataset) If yes, how to evaluate on the custom dataset or my own dataset？

JAYCHOU2020 commented 2 years ago

Hi, thanks for the wonderful work!

I have been able to train correctly in VisDrone dataset and generate tracking results.

But there are some problems when I run the test_det.py, test_emb.py and track.py, I tried to fix them but failed. It seems that the above python scripts are not able to perform multi-category evaluation well. (1) Have CUDA out of memory error in test_emb.py: torch.mm (line 102). (2) Shape missmatch in test_det.py: in function mot_decode(I try to fix it but don't know if it is correct) and line 32 (3) track.py does not run properly in the VisDrone dataset. Can someone help me？Thanks a lot!

I also want to ask if MOT metrics can only be evaluated in the MOT dataset? (Just like the COCO dataset) If yes, how to evaluate on the custom dataset or my own dataset？

same qusetion

JAYCHOU2020 commented 2 years ago

Hi, thanks for the wonderful work!

I have been able to train correctly in VisDrone dataset and generate tracking results.

But there are some problems when I run the test_det.py, test_emb.py and track.py, I tried to fix them but failed. It seems that the above python scripts are not able to perform multi-category evaluation well. (1) Have CUDA out of memory error in test_emb.py: torch.mm (line 102). (2) Shape missmatch in test_det.py: in function mot_decode(I try to fix it but don't know if it is correct) and line 32 (3) track.py does not run properly in the VisDrone dataset. Can someone help me？Thanks a lot!

I also want to ask if MOT metrics can only be evaluated in the MOT dataset? (Just like the COCO dataset) If yes, how to evaluate on the custom dataset or my own dataset？

did u solve this?

Soyad-yao commented 2 years ago

@StrongerzZ Hi, i have the same question with you. I can only train mcmot in visdrone dataset and i can not evaluate the pth i got. did you solve this problem?can you help me.It will be helpful for my undergraduate research

wsw321 commented 2 years ago

@Soyad-yao Have you solved this evaluation problem? Looking forward to your reply.

wsw321 commented 2 years ago

@JAYCHOU2020 have you solved this evaluation problem? Looking forward to your reply.