🤔 Ah, the eternal debate: is it the size of the #parameters or the motion of the #computation in the ocean that truly matters? 🏄‍♂️ Deep learning enthusiasts, gather 'round for this riveting tale where we learn that 1 million parameters = 1 million floating point operations! 🤯 Truly #groundbreaking stuff, folks!
https://parl.ai/projects/params_vs_compute/ #deeplearning #floatingpointoperations #debate #HackerNews #ngated