CrossFormer

This repository is the code for our paper CrossFormer: A Versatile Vision Transformer Based on Cross-scale Attention.

Introduction

Existing vision transformers fail to build attention among objects/features of different scales (cross-scale attention), while such ability is very important to visual tasks. CrossFormer is a versatile vision transformer which solves this problem. Its core designs contain Cross-scale Embedding Layer (CEL), Long-Short Distance Attention (L/SDA), which work together to enable cross-scale attention.

CEL blends every input embedding with multiple-scale features. L/SDA split all embeddings into several groups, and the self-attention is only computed within each group (embeddings with the same color border belong to the same group.).

Further, we also propose a dynamic position bias (DPB) module, which makes the effective yet inflexible relative position bias apply to variable image size.

Now, experiments are done on four representative visual tasks, i.e., image classification, objection detection, and instance/semantic segmentation. Results show that CrossFormer outperforms existing vision transformers in these tasks, especially in dense prediction tasks (i.e., object detection and instance/semantic segmentation). We think it is because image classification only pays attention to one object and large-scale features, while dense prediction tasks rely more on cross-scale attention.

Prerequisites

1. Libraries (Python3.6-based)

pip3 install numpy scipy Pillow pyyaml torch==1.7.0 torchvision==0.8.1 timm==0.3.2

2. Dataset: ImageNet

3. Requirements for detection/instance segmentation and semantic segmentation are listed here: detection/README.md or segmentation/README.md

Getting Started

Training

## There should be two directories under the path_to_imagenet: train and validation

## CrossFormer-T
python -u -m torch.distributed.launch --nproc_per_node 8 main.py --cfg configs/tiny_patch4_group7_224.yaml \
--batch-size 128 --data-path path_to_imagenet --output ./output

## CrossFormer-S
python -u -m torch.distributed.launch --nproc_per_node 8 main.py --cfg configs/small_patch4_group7_224.yaml \
--batch-size 128 --data-path path_to_imagenet --output ./output

## CrossFormer-B
python -u -m torch.distributed.launch --nproc_per_node 8 main.py --cfg configs/base_patch4_group7_224.yaml 
--batch-size 128 --data-path path_to_imagenet --output ./output

## CrossFormer-L
python -u -m torch.distributed.launch --nproc_per_node 8 main.py --cfg configs/large_patch4_group7_224.yaml \
--batch-size 128 --data-path path_to_imagenet --output ./output

Testing

## Take CrossFormer-T as an example
python -u -m torch.distributed.launch --nproc_per_node 1 main.py --cfg configs/tiny_patch4_group7_224.yaml \
--batch-size 128 --data-path path_to_imagenet --eval --resume path_to_crossformer-t.pth

Training scripts for objection detection: detection/README.md.

Training scripts for semantic segmentation: segmentation/README.md.

Results

Image Classification

Models trained on ImageNet-1K and evaluated on its validation set. The input image size is 224 x 224.

More results compared with other vision transformers can be seen in the paper.

Objection Detection & Instance Segmentation

Models trained on COCO 2017. Backbones are initialized with weights pre-trained on ImageNet-1K.

More results and pretrained models for objection detection: detection/README.md.

Semantic Segmentation

Models trained on ADE20K. Backbones are initialized with weights pre-trained on ImageNet-1K.

MS IOU means IOU with multi-scale testing.

More results and pretrained models for semantic segmentation: segmentation/README.md.

Citing Us

@article{crossformer2021,
  title     = {CrossFormer: A Versatile Vision Transformer Based on Cross-scale Attention},
  author    = {Wenxiao Wang and Lu Yao and Long Chen and Deng Cai and Xiaofei He and Wei Liu},
  journal   = {CoRR},
  volume    = {abs/2108.00154},
  year      = {2021},
}

Acknowledgement

Part of the code of this repository refers to Swin Transformer.