To further test the robustness of the model against background interference, we propose an ImageNet background interference test set, ImageNet-Bg, based on the ImageNet validation set with 48,285 ...

Abstract: Deep learning has witnessed significant advancements in various tasks and has displayed exceptional performance. However, traditional deep learning techniques often necessitate the ...

All models use multi-GPU setting with a total batch size of 4096 on ImageNet-1k and 1024 on ImageNet-22k. Training from scratch on ImageNet-1k. python -m torch.distributed.launch --nproc_per_node=8 ...

Abstract: Compared to models pretrained on ImageNet, the segment anything model (SAM) has been trained on a massive segmentation corpus, excelling in both generalization ability and boundary ...