CS @ MIT, HackMIT organizer, ML/AI fanboy
This level of accuracy made me pretty happy. Not quite state-of-the-art according to Yann Le-Cun’s website, but a solid B+ (A- if you round up). However, what was somewhat frustrating for me was how slow the script was, especially considering t…
…write this I only had my own experience and Google’s ByteNet to back this claim up. Just this week, Facebook published their fully convolutional translation model and reported a 9X speed up over LSTM based models.
Remember that the vision people got a lot of the same effect for images using convolutions. That is, instead of having an input node for each pixel, convolutions allowed the reuse of the same, small set of parameters across the entire image.