Skip to content

Latest commit

 

History

History
executable file
·
44 lines (34 loc) · 3.33 KB

MODEL_ZOO.md

File metadata and controls

executable file
·
44 lines (34 loc) · 3.33 KB

AdelaiDet Model Zoo and Baselines

Introduction

This file documents a collection of models trained with AdelaiDet in Nov, 2019.

Models

The inference time is measured on one 1080Ti based on the most recent commit on Detectron2 (ffff8ac).

More models will be released soon. Stay tuned.

COCO Object Detecton Baselines with FCOS

Name box AP download
FCOS_R_50_1x 38.7 model

COCO Instance Segmentation Baselines with BlendMask

Model Name inference time (ms/im) box AP mask AP download
Mask R-CNN 550_R_50_3x 63 39.1 35.3
BlendMask 550_R_50_3x 36 38.7 34.5 model
Mask R-CNN R_50_1x 80 38.6 35.2
BlendMask R_50_1x 73 39.9 35.8 model
Mask R-CNN R_50_3x 80 41.0 37.2
BlendMask R_50_3x 74 42.7 37.8 model
Mask R-CNN R_101_3x 100 42.9 38.6
BlendMask R_101_3x 94 44.8 39.5 model
BlendMask R_101_dcni3_5x 105 46.8 41.1 model

COCO Panoptic Segmentation Baselines with BlendMask

Model Name PQ PQTh PQSt download
Panoptic FPN R_50_3x 41.5 48.3 31.2
BlendMask R_50_3x 42.5 49.5 32.0 model
Panoptic FPN R_101_3x 43.0 49.7 32.9
BlendMask R_101_3x 44.3 51.6 33.2 model
BlendMask R_101_dcni3_5x 46.0 52.9 35.5 model

Person in Context with BlendMask

Model Name box AP mask AP download
BlendMask R_50_1x 70.6 66.7 model