Tim Meinhardt
Ah no, they are not aligned/arranged by id.
Ah no, they are not aligned/arranged by id.
Can u post the full error log?
What are your torch, torchvision, CUDA and CUDNN versions? I assume there is a version mismatch.
Please follow the [INSTALL](https://github.com/timmeinhardt/trackformer/blob/main/docs/INSTALL.md) instructions step by step and install the correct versions. This repository does not support `torch=1.8.0`.
This is most likely a version mismatch. For example, CUDNN 7.6.7 does not support CUDA 10.0. At least according to [this](https://docs.nvidia.com/deeplearning/cudnn/archives/cudnn_765/cudnn-support-matrix/index.html) support matrix.
The code is able to train and test detection results. I suggest running it with the `deformable` option as well. Is the loss increasing when you train or are you...
You are preloading the `resume: models/r50_deformable_detr-checkpoint.pth` model. But without deformable option. Please make yourself familiar with the configurations and ideally start with one of the example commands and change it...
In your first message you only wrote `with crowdhuman`. What is your full command then?
Please start with one of the presented commands and see if these work. And then build your command up from that.