Community, Please check out and share the post: https://medium.com/apache-mxnet/gluoncv-0-3-a-new-horizon-564326364e16 for v0.3 of gluon-cv, detailing improvements to existing implementations of state of the art cv models. Particularly of note, is the accuracy improvement for the ResNet family compared to the original reference papers and to gluon-cv v0.2. In addition, several new state of the art models like Yolo v3, DeepLab v3 and MaskRCNN with accuracy greater than or equal to reference have been added.
Thanks to Zhi Zhang and Hao Jin for working to get the post published and all gluon-cv contributors.