Tim Meinhardt comments

Results 144 comments of


                                            Tim Meinhardt

Use of args.multi_frame_attention

The `track_attention` is a legacy parameter and was not used during any of the trainings.

Error while train when get dataset

What do you mean by "data set is not a pair"? Did the `generate_coco_from_mot.py` scripts finish successfully? This looks like a data path issue.

Error while train when get dataset

Attention maps and mask prediction are two separate things. However, only for our model for mask prediction the attention map plotting is implemented. If you want to generate attention maps...

Thoughts poor performance?

I think you are facing potentially two issues: 1. Generalization to a different data domain (soccer videos). However, I think our model should handle this better than in the video...

detections are not associated into the track

Debugging on a custom dataset is beyond the support of this repository. But you should check out the classification scores and thresholds in `track.yaml`.

I try to train with MOT17,when I use "multi_frame_attention = True",I get " RuntimeError: stack expects each tensor to be equal size, but got [2, 2, 21, 27, 43] at entry 0 and [2, 2, 21, 27, 42] at entry 1"

Please follow the install instructions. For example, your PyTorch version is wrong. This should explain the first error. The second one might also be resolved if not your system is...

I try to train with MOT17,when I use "multi_frame_attention = True",I get " RuntimeError: stack expects each tensor to be equal size, but got [2, 2, 21, 27, 43] at entry 0 and [2, 2, 21, 27, 42] at entry 1"

Fixing the code for the PyTorch version you are limited to. But this might involve quite a lot of code changes. The support for this goes beyond this repository.

I try to train with MOT17,when I use "multi_frame_attention = True",I get " RuntimeError: stack expects each tensor to be equal size, but got [2, 2, 21, 27, 43] at entry 0 and [2, 2, 21, 27, 42] at entry 1"

You can not just set the param to false. You must remove the entire `multi_frame` option from the command. To revert all the changes [this](https://github.com/timmeinhardt/trackformer/blob/main/cfgs/train_multi_frame.yaml) is applying.

I try to train with MOT17,when I use "multi_frame_attention = True",I get " RuntimeError: stack expects each tensor to be equal size, but got [2, 2, 21, 27, 43] at entry 0 and [2, 2, 21, 27, 42] at entry 1"

I can not check all those parameters. Try running ``` python src/train.py with \ mot17_crowdhuman \ deformable \ multi_frame \ tracking \ output_dir=models/mot17_crowdhuman_deformable_multi_frame \ ``` without the `multi_frame` line.

I try to train with MOT17,when I use "multi_frame_attention = True",I get " RuntimeError: stack expects each tensor to be equal size, but got [2, 2, 21, 27, 43] at entry 0 and [2, 2, 21, 27, 42] at entry 1"

The command seems to be missing the `tracking` and `\`. Please take a step back and see what you want to do. We can not debug your issue via github....