Neural Search Talks β€” Zeta Alpha cover image

ColBERT + ColBERTv2: late interaction at a reasonable inference cost

Neural Search Talks β€” Zeta Alpha

00:00

The MaxSim Operation and the Difference in Performance

The G-so-Vit is much faster and it's comparably performant, not quite but comparably, right? And you can see also in a figure four, there's also a nice kind of visualization of the flops required for... Is it inference or...? Yeah, to re-rank depth K. The difference in orders of magnitude is around two to four orders of magnitude, right?Yeah, it's so much faster, exactly. I mean, I don't think we need to get a lot more in the details of the results.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app