HyperAIHyperAI

Command Palette

Search for a command to run...

Console

Panoptic Segmentation On Coco Minival

Metrics

AP
PQ

Results

Performance results of various models on this benchmark

Paper Title
HyperSeg (Swin-B)-61.2HyperSeg: Towards Universal Visual Segmentation with Large Language Model
OneFormer (InternImage-H,single-scale)52.060.0OneFormer: One Transformer to Rule Universal Image Segmentation
OpenSeeD (SwinL, single-scale)53.259.5A Simple Framework for Open-Vocabulary Segmentation and Detection
MasK DINO (SwinL,single-scale)50.959.4Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation
DiNAT-L (single-scale, Mask2Former)49.258.5Dilated Neighborhood Attention Transformer
ViT-Adapter-L (single-scale, BEiTv2 pretrain, Mask2Former)48.958.4Vision Transformer Adapter for Dense Predictions
Visual Attention Network (VAN-B6 + Mask2Former)-58.2Visual Attention Network
kMaX-DeepLab (single-scale, pseudo-labels)-58.1kMaX-DeepLab: k-means Mask Transformer
HIPIE (ViT-H, single-scale)-58.1Hierarchical Open-vocabulary Universal Image Segmentation
kMaX-DeepLab (single-scale, drop query with 256 queries)-58.0kMaX-DeepLab: k-means Mask Transformer
OneFormer (DiNAT-L, single-scale)49.258.0OneFormer: One Transformer to Rule Universal Image Segmentation
FocalNet-L (Mask2Former (200 queries))48.457.9Focal Modulation Networks
kMaX-DeepLab (single-scale)-57.9kMaX-DeepLab: k-means Mask Transformer
OneFormer (Swin-L, single-scale)49.057.9OneFormer: One Transformer to Rule Universal Image Segmentation
Mask2Former (single-scale)48.657.8Masked-attention Mask Transformer for Universal Image Segmentation
Panoptic SegFormer (single-scale)-55.8Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers
CMT-DeepLab (single-scale)-55.3CMT-DeepLab: Clustering Mask Transformers for Panoptic Segmentation
MaskFormer (single-scale)-52.7Per-Pixel Classification is Not All You Need for Semantic Segmentation
MaX-DeepLab-L (single-scale)-51.1MaX-DeepLab: End-to-End Panoptic Segmentation with Mask Transformers
Panoptic SegFormer (ResNet-101)-50.6Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers
0 of 28 row(s) selected.