Mikio L. Braun

ML and AI expert

Back to archive

Thread

2 tweets

1
Yeah, got it. It seems that we've pushed the envelope quite far in terms of performance, but I think there's also a real need for models that are still OK but can be run locally or even on CPUs, so I'm expecting some progress in that direction.
2
@rasbt And sometime academia has this tendency to forget about paths it has already taken, so yeah. But linformers etc. are probably not too long ago. Then again given the speed at which the field is moving... Thanks for you thoughts, in any case, Sebastian!