Imagine if folks used custom chips instead of standard gaming chips for deep learning. ~100X boost possible...https://twitter.com/mappingbabel/status/676853665500495872 …
@nachiketkapre @adapteva Then you'd have to use eg. SGD instead of batch methods like BFGS. Afaik SGD usually converges much much slower..
-
-
@oe1cxw@nachiketkapre@adapteva Depends: wall time may dec., trading off higher iterations count (low. rate) for lower per-iteration-cost.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@oe1cxw@nachiketkapre@adapteva Trade-off: Table 2 in http://leon.bottou.org/publications/pdf/compstat-2010.pdf … / discussion around 22 min mark here: https://youtu.be/l5JqUvTdZts?t=22m …Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.