Dynamically Instantiated Network

Reported on 3 benchmarks across 3 tasks · 1 paper · 3 SOTA

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

Medical1 result

Semantic SegmentationonCityscapes test
PQ· 2017-04-07
55.4
best: 68 (OneFormer (ConvNeXt-L, single-scale, Mapillary Vistas-Pretrained))
SOTA
Pixelwise Instance Segmentation with a Dynamically Instantiated Network arXiv:1704.02386

Audio1 result

10-shot image generationonCityscapes test
PQ· 2017-04-07
55.4
best: 68 (OneFormer (ConvNeXt-L, single-scale, Mapillary Vistas-Pretrained))
SOTA
Pixelwise Instance Segmentation with a Dynamically Instantiated Network arXiv:1704.02386

Computer Vision1 result

Panoptic SegmentationonCityscapes test
PQ· 2017-04-07
55.4
best: 68 (OneFormer (ConvNeXt-L, single-scale, Mapillary Vistas-Pretrained))
SOTA
Pixelwise Instance Segmentation with a Dynamically Instantiated Network arXiv:1704.02386