You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I see lots of people stating unequivocally that the deep learning boom started with Krizhevsky et al 2012. But I see little credit given to Ciresan et al, who were winning image classification competitions in 2011 with deep convnets implemented in CUDA, trained on NVIDIA GPUs
I'm referring to the German Traffic Sign Recognition Benchmark at IJCNN 2011, and the ICDAR 2011 Chinese Handwriting Recognition Competition - refs: http://www.nlpr.ia.ac.cn/events/HRcompetition/ICDAR2011%20CHR%20Competition%20Final.pdf … - http://people.idsia.ch/~juergen/nn2012traffic.pdf …
I remember asking Dan Ciresan about his thoughts on Torch7, back in 2014 -- I was a Torch user at the time. He said, in essence, that he thought it was an unusable pile of bugs, and that he still preferred writing his own CUDA. I disagreed with his position at the time (still do)
I think that there is a fundamental difference between being able to do 2d classification of text and aligned and cropped street signs and doing the more wild classification of 3d objects in image net. It's a lot harder to imagine the model memorizing the data in the latter
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
