Jinkun Cao

Hi @Mobu59 , I thought that JDE Kalman FIlter means using the embeddings from the famous [JDE](https://github.com/Zhongdao/Towards-Realtime-MOT) model together with a canonical Kalman Filter. But in the following post, @HanGuangXin...

Hi @HanGuangXin , thank you for clarifying and providing more details. I always enjoy sharing my idea with the community! 1. influence of allowing non-fixed aspect ratio: I don't know...

I am keeping this issue open as many others may be interested in the combination of OC-SORT and BYTE. I wish the posts here could be helpful to them.

OC-SORT has supported BYTE from [PR #19](https://github.com/noahcao/OC_SORT/pull/19). Thanks @HanGuangXin for the contribution.

@abhigoku10 Please refer to the code contribution from PR #19

at this moment, i don't have the bandwidth to write a full C++ version. If you'd like to deploy on some mobile device, a workaround is to use the DEPLOYMENT...

Looks like the reason is somehow the parameter "exps/example/mot/yolox_s_mix_det.py" not captured by args.exp_file, which looks strange to me. Could you try to modify the default value of exp_file to the...

For the results reported in the paper, I use the detections from PermaTrack's results, please check the [doc](https://github.com/noahcao/OC_SORT/blob/master/docs/GET_STARTED.md) for details.

Currently, I have not support customized detectors here. But replacing it should be feasible by manipulating the returned model of `get_model()` [here](https://github.com/noahcao/OC_SORT/blob/bfa9aa6af44570b9049aae8fcad57773e010886a/yolox/core/trainer.py#L129).

I made a [PR](https://github.com/open-mmlab/mmtracking/pull/545) to mmtracking to support OC-SORT. As mmtracking is based on mmdetection, you can replace the detector with what you want there.