1 citations · 1 across the 7 of their papers we have counts for
4 papers · 1 filter
4Director: Controlling Video World Models with Rigid 3D Geometry
Wei Cao, Hao Zhang, Vikram Voleti +5
Precise control over camera and object motion is essential for professional video production. Existing methods control objects only coarsely, through image-plane cues that are ambi…
SemanTok: Predictable Semantic Tokens for Efficient Autoregressive Video Generation
Mikhail Dereviannykh, Vikram Voleti, Simon Donne +3
Recent video-based world models pair the scalability of autoregressive (AR) prediction with the visual quality of diffusion models. The choice of scene tokenizer is paramount for t…
Fill My Mirror: Geometry-Constrained Mirror Inpainting
Ofek Basson, Shimon Vainer, Yacov Hel-Or +1
Mirrors are common in real-world images, yet producing geometrically consistent reflections with generative models remains challenging. Unlike most objects, mirror appearance depen…
OCTOPUS: Optimized KV Cache for Transformers via Octahedral Parametrization Under optimal Squared error quantization
Mark Boss, Vikram Voleti, Simon Donné +1
The key-value (KV) cache dominates memory bandwidth and footprint in long-context autoregressive inference. Recent rotation-preconditioned codecs (TurboQuant, PolarQuant) show that…