-
Is it possible to replace a ViT backbone with a regular CNN backbone like Resnet?
-
RT,
I try to run on A100(40G memory) ,but have out of memory issue ?
-
Hi,
Could you share some GPU information and Length of training time ？ I try to reproduce your paper , but found this paper have a lot of dependence on High level training platform (A100 or mo…
-
您的论文中的Table 2里边显示的Mixformer的速度是25FPS，GPU是1080Ti，请问一下您这个结果用了几个GPU？是在哪个数据集上跑的？还有就是我用的Mixformer在UAV123数据集上跑出来的结果是6.1FPS，两个2080Ti，请问这正常吗？十分感谢
-
Thank you for your excellent work!
When I load ConvMAE-v2-Base pretrained checkpoints [https://drive.google.com/file/d/1gykVKNDlRn8eiuXk5bUj1PbSnHXFzLnI/view?usp=sharing], it has cls_token paramete…
-
Could you provide a tutorial on how to train and finetune with custom dataset? And how to modify the input image size during the detection, the current code seems not to support custom image size.
-
Hi! Thanks for the opensource code.
I noticed that the mask convolution in the code only masks the residual block, but the skip connection does not have a mask, as shown in line 119 of "ConvMAE/visio…
-
ConvMAE: Masked Convolution Meets Masked Autoencoders
https://arxiv.org/abs/2205.03892
https://github.com/alpha-vl/convmae
Please consider the addition of ConvMAE on your repo.
-
Hi,
ConvMAE is an interesting work and this work seems to take it to the next level. I'm looking forward to seeing the official paper. May I know is there any plan to release the paper? Thanks!
-
- News
- Conference
- KDD 2022 notification: Accept 되신 분들 모두 축하드립니다!
- NeurIPS 2022 deadline: 모두들 수고 많으셨습니다. (supple은 26일까지)
- AI미래포럼-공학한림원 공동 주최 [Industrial AI Forum](https://www.yout…