A Unified Tracking Repository
- Distributed Training and Inference for both SOT and MOT tasks
- Integrated Popular SOT and MOT methods and evaluation benchmarks
See INSTALL.md
We have implemented several SOT methods (DiMP, SiamRPN, Stark) in this repo and support popular benchmark datasets (OTB, LaSOT, GOT10k, TrackingNet) for evaluation. For more information, see SOT_README.md
We have implemented TransTrack, FairMOT, and ByteTrack for MOT evaluation. For more information, see MOT_README.md
All models can be downloaded from google drive
A large part of code in current repo is borrowd from Detectron, PyTracking, TransTrack, ByteTrack Libraries. Many thanks for their wonderful works.