SegFormer-B5(MS, 87M #Params, ImageNet-1K pretrain)

Reported on 2 benchmarks across 2 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical1 result

Semantic SegmentationonADE20K val
mIoU· 2021-05-31
51.8
best: 62.8 (BEiT-3)
SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers arXiv:2105.15203

Audio1 result

10-shot image generationonADE20K val
mIoU· 2021-05-31
51.8
best: 62.8 (BEiT-3)
SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers arXiv:2105.15203