Meta Byte Track
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

README.md 1.6KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465
  1. # CTracker
  2. #### Step1
  3. git clone https://github.com/pjl1995/CTracker.git and preapare dataset
  4. #### Step2
  5. add generate_half_csv.py to https://github.com/pjl1995/CTracker
  6. run generate_half_csv.py and put train_half_annots.csv in MOT17
  7. run
  8. ```
  9. python3 train.py --root_path MOT17 --csv_train train_half_annots.csv --model_dir ctracker/ --depth 50 --epochs 50
  10. ```
  11. You can also download the CTracker model trained by us: [google](https://drive.google.com/file/d/1TwBDomJx8pxD-e96mGIiTduLenUvmf1t/view?usp=sharing), [baidu(code:6p3w)](https://pan.baidu.com/s/1MaCvnHynX2Wzg81hWkqzeg)
  12. #### Step3
  13. replace https://github.com/pjl1995/CTracker/blob/master/test.py
  14. run
  15. ```
  16. python3 test.py --dataset_path MOT17 --model_dir ctracker --model_path ctracker/mot17_half_ctracker.pt
  17. ```
  18. #### Step4
  19. add eval_motchallenge.py to https://github.com/pjl1995/CTracker
  20. prepare gt_half_val.txt as CenterTrack [DATA.md](https://github.com/xingyizhou/CenterTrack/blob/master/readme/DATA.md)
  21. #### Step5
  22. run
  23. ```
  24. python3 eval_motchallenge.py --groundtruths MOT17/train --tests ctracker/results --gt_type half_val --eval_official --score_threshold -1
  25. ```
  26. # CTracker_BYTE
  27. #### Step3
  28. add mot_online to https://github.com/pjl1995/CTracker
  29. add byte_tracker.py to https://github.com/pjl1995/CTracker
  30. add test_byte.py to https://github.com/pjl1995/CTracker
  31. run
  32. ```
  33. python3 test_byte.py --dataset_path MOT17 --model_dir ctracker --model_path ctracker/mot17_half_ctracker.pt
  34. ```
  35. #### Step5
  36. run
  37. ```
  38. python3 eval_motchallenge.py --groundtruths MOT17/train --tests ctracker/results --gt_type half_val --eval_official --score_threshold -1
  39. ```