@fchollet the memnn variation in keras was your idea? Did u publish it somewhere? I mean there is no per-sequence embedding and no softmax
@edersantana yes, gotta be creative with this stuff. Seems to work as well as the original on babi. You can even remove one of the branches.
-
-
@fchollet i noticed I'm implementing the original version tho. But yours seems to work pretty well without the extra passes through the dataThanks. 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.