xmyqsh

FPN. Feature Pyramid Network

deeplab-v2. DeepLab v2 support cudnn-v5

RetinaNet. Focal loss for Dense Object Detection

Coursera. honor programming assignment

AI-Challenger-Keypoint. Lua

SNIP_on_cityscapes. my state of the art result on cityscapes

TusharWhiteboard. Translate Tushar Roy's whiteboard to code

gpu_coords_map. C++

leetcode2016. C++

dlsys-course. CSE 599W: Systems for ML

SUMU_user_pairing. communication project and paper

py-faster-rcnn. faster-rcnn upcompatible with caffe/master for supporting the latest cuDNN 5.1.5

FSN. Fully Sparse Network (5 FPS, IO bounded, 方向分类不需要取模操作，加速训练)