Loading video...
Video Failed to Load
"Cameras as Relative Positional Encoding" TLDR: comparison for conditioning transformers on cameras: token-level raymap, attention-level relative pose encodings, a (new) relative encoding Projective Positional Encoding -> camera frustums, (int|ext)insics for relative pos encoding
17,837 views • 1 year ago •via X (Twitter)
3 Comments

Alexandre Morgand1 year ago
1,2*, @brenthyi 1*, @JunchenLiu77 1*, @hangg70 1, @YiMaTweets 1,3, @akanazawa 1 ( *Equal contribution ) 1 @UCBerkeley 2 @nvidia 3 @HKUniversity Project page: Paper: Source code:

Ruilong Li1 year ago
thanks for sharing!

Alexandre Morgand1 year ago
Anytime!

