Dual Sparse Long-Short Term Transformer for Video Shadow Detection Dataset ViSha dataset is available at ViSha Homepage. Requirement python==3.8 cuda==11.3 torch==1.11.0 torchvision==0.12.0 medpy==0.4.0 numpy==1.24.4 pillow==10.0.0 scipy==1.10.1 tqdm==4.65.0 Testing python test.py Our pretrained model is available here(code:u2mg).