Deep Learning: Computer Vision Models

Fast R-CNN (Ross Girshick): Fast Region-based Convolutional Networks for object detection

Faster R-CNN (Ross Girshick): Towards Real-Time Object Detection with Region Proposal Networks

FCOS: Fully Convolutional One-Stage Object Detection

Mask RCNN (Matterport implementation): for Object Detection and Segmentation

Microsoft: Best Practices for Computer Vision

MMDetection: an open source object detection toolbox based on PyTorch. The training speed is faster than or comparable to other codebases.

ResNeXt: Aggregated Residual Transformations for Deep Neural Networks

YOLO: Real-Time Object Detection