New code walkthrough on http://keras.io : a reproduction of the ConvMixer architecture from the "Patches are all you need" paper. Implemented by @RisingSayakhttps://keras.io/examples/vision/convmixer/ …
Show this thread
I think it's fun that vision models are converging towards an architecture that is nearly indistinguishable from Xception (back from 2016). Depthwise separable convolutions are simply the correct architecture prior for spatial perception, and it's time they become mainstream.
-
-
EfficientNet 2 had to throw some of them away thoughpic.twitter.com/h941J6IScw
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Most advancements revolve around the established ideas
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.