liangheming Goto Github PK

atssv1

ATSS (retina) 39.6mAP on COCO,640px(max side),42.95fps(RTX 2080TI)<<Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection>>

auto_assignv2

An unofficial pytorch implementation of <<AutoAssign: Differentiable Label Assignment for Dense Object Detection>>

bev-perception

Bird's Eye View Perception

cascade_rcnn

pytorch implement of CascadeRCNN,736px(max side),41.2mAP(COCO),21.94fps(RTX 2080TI)

condinst

detr

efficientdet

faster_rcnnv1

pytorch implement of fasterRCNN,736px(max side),39.4mAP(COCO),30.21fps(RTX 2080TI)

fastlio2_ros2

ROS2 / FAST_LIO / PGO / Online Re-Localization / Consistent Map with BA or HBA

fastlio2_sam_lc

fastlio2 with loop closure and online localization

fcosv1

FCOS ,640px(max side),mAP 38.2,fps45.45(RTX 2080TI).faster! better! <<Fully Convolutional One-Stage Object Detection>>

free_anchorv1

pytorch implement of FreeAnchor(not strict),640 px(max side),39.5mAP on COCO,43.18fps(RTX 2080TI)

gast-net-3dposeestimation

A Graph Attention Spatio-temporal Convolutional Networks for 3D Human Pose Estimation in Video (GAST-Net)

gfocal

GFocal 40.6mAP(COCO) 640px(max side) 44.05 fps(rtx 2080ti)

glim

GLIM: versatile and extensible range-based 3D localization and mapping framework

ig-lio_sam_lc

iG-LIO with Loop Closure(PGO) and Online Re-Localize

kinetics-i3d

Convolutional neural network model for video classification trained on the Kinetics dataset.

light_detection

light-weight detection

m_fastlivo

FAST-LIVO easy to read, esay to understand

m_vins

VINS_Mono FOR ros2

maskrcnn

pytorch mask rcnn 736px，box mAP 39.0, seg mAP 33.7 (COCO val) 23.64fps(RTX 2080Ti)

monodepth_v2

MonoDepthv2

my_vins_mono

objectdetectionimbalance

Lists the papers related to imbalance problems in object detection

paav1

PAA ,resnet50+fpn,640px(max side),41.4mAP,43.68fps(RTX 2080TI)<<Probabilistic Anchor Assignment with IoU Prediction for Object Detection>>

pointpillar3d

lidar based 3d dectection

pytorch-i3d

ranger-deep-learning-optimizer

Ranger - a synergistic optimizer using RAdam (Rectified Adam), Gradient Centralization and LookAhead in one codebase

retinanetv1

pytorch implement of retinanet,37.4 mAp(coco) at 640px(max side) ,44.24fps(RTX2080TI)

scaled_yolo