Chang-Bin Zhang

> > > > > I got the same problem > > > > > > > > > > > > Is there any solution? > > Maybe it's...

It works well when I use gcc-9.4. @ma3252788

hi, @Ze-Yang We found that high version pytorch can not reach the supposed performance due to the bug caused by InplaceABNSync, which has been fixed in our latest commitment. You...

hi, @XiaorongLi-95 We found that high version pytorch can not reach the supposed performance due to the bug caused by InplaceABNSync, which has been fixed in our latest commitment. You...

I will update the training code on Noisy-CIFAR in one week. Thanks. @daikiokamura

Hi, @BarCodeReader You can use the bash script to download the dataset. ``` cd ./data/PASCAL/ wget https://www.dropbox.com/s/4ut7uju34madtq4/PASCALContourData.tar.gz?dl=0 tar -xvzf PASCALContourData.tar.gz cd ../../ ```

Hi, @BarCodeReader There is a ```run.sh``` to train or test the model.

Hi, @yzhou377 and @pha-nguyen , I also implemented the drivable area and road on the Nuscenes, but the performance is worse. Do you have the same trouble?

@TempleX98 Hi, I encountered the error ```Co-DETR/mmdet/datasets/builder.py", line 80, in build_dataset dataset = MultiImageMixDataset(**cp_cfg) TypeError: __init__() got an unexpected keyword argument 'filter_empty_gt' ```, when I use config ```projects/configs/co_dino/co_dino_5scale_lsj_swin_large_3x_coco.py```.

Amazing, it will spend 12 days training ```co_dino_5scale_lsj_swin_large_3x_coco.py``` with 8 GPUs. It is so strange because the whole dataset has double size samples (14786 vs 7000+). ```mmdet - INFO -...