zeal-up

same issue here

@RainerKuemmerle Hi～I know it's a long time ago issue, but I still finding a way to visualize custom-types in g2o_viewer. I can load .g2o file which contain customized types to...

@meder411 I think the difference between pytorch code and the official tf code is due to the dimension arrangement difference between pytorch and tensorflow. In pytorch, N, C, H, W,...

@dybas133 Hi, I have also noticed this "bug?". Have you get more knowledge about this?