Pengxiang Li

Thank you very much for your suggestion. We will discuss this in the meeting.😀

Sorry, I gave the wrong configuration file. The correct one is as follows. Just **delete** this parameter. ```python dataset_type = 'CocoDataset' img_norm_cfg = dict( mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375],...

You are using a `metric` method that is not supported. When running `python test.py`, please note the argument after `--eval`.

Does the file `gt.txt` exist in the path `data/MOT20/test/MOT20-04/gt/`? Have you checked？

Well, I see. You want to eval the **test** set, but you can't do it locally. The **MOT challenge** doesn't provide gt.txt for test set verification, so `--eval track` may...

You're welcome. Hope everything goes well with you.🙂

Hello, For some reasons, the weight of qdtrack on TAO has not been given. You can download it at this temporary link first. https://drive.google.com/file/d/1U4XUWcfpQQOv06oWVPOt0AIqPQqtfE2U/view?usp=sharing Sorry for the inconvenience.

Thank you for your feedback！ Please change your file referring to this PR: https://github.com/open-mmlab/mmtracking/pull/538

```python model = dict( **detector**=dict( backbone=dict(xxxx) ) ) ``` The correct form is as above.(`**` is no use, just as a reminder, please remember to delete) In general, to change...

Just like this, you should use ``` model = dict( detector=dict( backbone=dict(xxxx) ) ) ``` instead of ``` model = dict( backbone=dict(xxxx) ) ```