Has anyone else notices SwiftKey flow really lose its way with predictions since the update?

Hey,

Since the update to Swiftkey, I've really really noticed a drop in the accuracy of the swiping. The words it creates are miles off what I'm writing and I'm correcting more words than it's getting right at the moment swiping, and I don't even mean to exaggerate.

Is it just me, or has anyone else noticed this?