mmtracking
mmtracking copied to clipboard
OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.
 It seems that only mmcv class have been registered in the project, how can i solve it? Wish your reply, thank you.
Hi! At first, thanks a lot for your great effort to create mmtracking! I have a question regarding MOT testing: I want to compare the performance of the MOT-models, independent...
For Siameserpn, in my dataset there are images without bounding boxes what should I initialise those values in groundtruth.txt? I tried 0,0,0,0 -1,-1,-1,-1 and blank line but nothing is working...
Thanks for your error report and we appreciate it a lot. **Checklist** 1. I have searched related issues but cannot get the expected help. 2. The bug has not been...
When I run `demo_mot_vis.py` on branch 1.x, the throughput is very low (about 2 task/s), but when I run benchmark with the same config, I got 15 FPS. 
I'm trying to run MaskTrack RCNN evaluation on the Youtube VIS 2019 using the below script ``` !python tools/test.py \ configs/vis/masktrack_rcnn/masktrack_rcnn_r50_fpn_12e_youtubevis2019.py \ --checkpoint checkpoints/masktrack_rcnn_r50_fpn_12e_youtubevis2019_20211022_194830-6ca6b91e.pth \ --out masktrack_rcnn_result_dir/results.pkl \ --eval 'track_segm'...
When i tried to run inference in docker i found this error, please provide solution for this issue. Traceback (most recent call last): File "/mmtracking/./demo/demo_mot_vis.py", line 119, in main(args) File...
Hello, I want to know how to get the search_metrics=['MOTA', 'IDF1', 'FN', 'FP', 'IDs', 'MT', 'ML']? In the detectrone2, i just need to test the video by the MOT model...