aldi
aldi copied to clipboard
justinkay
Official implementation of "Align and Distill: Unifying and Improving Domain Adaptive Object Detection"
This PR implements Contrastive Mean Teacher (CMT) on top of Adaptive Teacher (AT). Some code in `cmt.py` is taken from the [original implementation](https://github.com/Shengcao-Cao/CMT).
Hi, Thank you for your research and I really like your achievement and sharing. I evaluated the model "cityscapes_baseline_strongaug_ema_foggy_val_model_best_591_ema2model.pth" with this command "python tools/train_net.py --config configs/cityscapes/Base-RCNN-FPN-Cityscapes_strongaug_ema.yaml --eval-only" I got the...
Hi, I think the source-only model AP50 should be "59.1" like other ablation study result. By the way, I would like to use the best model. Will you share the...
After successfully training the models, I am trying to use this to visualize the feature space. but after a while the process is getting killed automatically. There is no clue...
Hi Dr. Kay, I'm working with the code and am impressed with the approach. Do you have any suggestions for potential future work or improvements that could further enhance the...
I made the baseline training for the PT. and then used the priorart PT for training it on my custom dataset all training is success but when I try to...
Hi, Do you have a plan to release the final model weights for VitDet-B on cityscapes foggy? best
