TikTok presents Depth Anything
Unleashing the Power of Large-Scale Unlabeled Data
paper page:
demo:
Depth Anything is trained on 1.5M labeled images and 62M+ unlabeled images jointly, providing the most capable Monocular Depth
@wartranslated
honestly I cannot decide which scenario is funnier, managing to lose an AEW&C aircraft to enemy air defense for the first time in world history or losing it to your own air defense
@bayraktar_1love
I mean it's certainly technically possible, but holy shit losing something to S-200 in the year 2024 is even more embarrassing than shooting it down yourself lmao
@slip_trace
x imeni y has got to be one of the most awkward constructions to translate lmao, it always feels so out of place in english no matter how you do it
@KarnesMellagio
@un_a_valeable
honestly you gotta respect how they had to fuck up literally every single possible thing in order to lose to the communists and yet they NAILED it
that's real dedication to the craft
@canderaid
I still maintain that eichenlaubmuster and erbsenmuster were divinely inspired camo patterns and a solid 90% of waffen-ss popularity is purely because of how cool they look lmao
@cattle_rattle
@Hanasaku_Yuri
saint-gaudens wasn't actually very good at sculpting horses lol, that and his fucked riding posture were the main complaints originally
fortunately for him most people no longer have any personal experience with horses or riding, so his work gets a lot less complaints now