zyayoung
zyayoung
Please check that arguments in `exps/aps.sh` are used, e.g. `--with_box_refine` is required.
We tried training on 8 V100 GPUs without checkpointing and do not encounter a CUDA OOM issue. If OOM occurs in the middle of an epoch, maybe MOTR cumulates too...
Do you mean pretrained weight of Deformable DETR or YOLOX? For YOLOX, you can use any weight or detection model as long as the boxes are correctly fed into the...
It doesn't matter, just ignore the warning messages.
Implementing batch size > 1 is possible but requires some work on handling track queries with variable lengths between batches. Using batch size 1 already uses all GPU memory on...
Yes, but we have not checked the multi-class implementation in this repo. We will check it in a few weeks.
Could you provide more details about the sampler? My guess is that you are using a batch size > 1, which is not supported in ByteTrack.
Sorry for the late reply, we will check it soon.