Maybe someone more familiar with deep learning can weigh in, but this bit
> Howard’s team was able to compete with the likes of Google by doing a lot of simple things. These include making sure that the images fed to its training algorithm were cropped correctly: “These are the obvious, dumb things that many researchers wouldn’t even think to do,” Howard says.
makes it sound like they got better performance by offloading some work to the pre-processing stage. Or maybe I don't really understand how this benchmark works?
> Howard’s team was able to compete with the likes of Google by doing a lot of simple things. These include making sure that the images fed to its training algorithm were cropped correctly: “These are the obvious, dumb things that many researchers wouldn’t even think to do,” Howard says.
makes it sound like they got better performance by offloading some work to the pre-processing stage. Or maybe I don't really understand how this benchmark works?