almost every ML system in production at scale is truly a distributed system.
it's not like you can just have tensorflow on a single server running predictions and call it a day, there is so much involved in the entire pipeline, and it's very distributed.
comments
Oh sorry, I forgot, logic isn't your strong suit.
it's not like you can just have tensorflow on a single server running predictions and call it a day, there is so much involved in the entire pipeline, and it's very distributed.