How to conduct HOTA MOTA IDF1 MT ML FP FN IDs FPS evaluation

[luke110]

[Xavier-Lin]

For the results with CLEAR metrics, you can run tracking scripts [such as, mot17_track_cfg.py] to get the results.

For the results with HOTA metrics, you can refer to 'tools/eval_hota.sh', modify your own file paths, and run it.

[luke110]

对于具有 CLEAR 指标的结果，您可以运行跟踪脚本 [例如 mot17_track_cfg.py] 来获取结果。

对于HOTA指标的结果，可以参考tools/eval_hota.sh，修改自己的文件路径，然后运行。

I found that every time when train is executed, the default evaluation is coco’s AR evaluation. How should I modify it so that it becomes mot’s evaluation, such as mota, hota, etc. Can you provide relevant code updates?