BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171532Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T171600
DTEND;TZID=Asia/Tokyo:20241205T172800
UID:siggraphasia_SIGGRAPH Asia 2024_sess137_papers_572@linklings.com
SUMMARY:Fast High-Resolution Image Synthesis with Latent Adversarial Diffu
 sion Distillation
DESCRIPTION:Axel Sauer, Frederic Boesel, Tim Dockhorn, Andreas Blattmann, 
 Patrick Esser, and Robin Rombach (Black Forest Labs)\n\nDiffusion models a
 re the main driver of progress in image and video synthesis, but suffer fr
 om slow inference speed. Distillation methods, like the recently introduce
 d adversarial diffusion distillation (ADD) aim to shift the model from man
 y-shot to single-step inference, albeit at the cost of expensive and diffi
 cult optimization due to its reliance on a fixed pretrained DINOv2 discrim
 inator. We introduce Latent Adversarial Diffusion Distillation (LADD), a n
 ovel distillation approach overcoming the limitations of ADD. In contrast 
 to pixel-based ADD, LADD utilizes generative features from pretrained late
 nt diffusion models. This approach simplifies training and enhances perfor
 mance, enabling high-resolution  multi-aspect ratio image synthesis. \nWe 
 apply LADD to Stable Diffusion 3 (8B) to obtain SD3-Turbo, a fast model th
 at matches the performance of state-of-the-art text-to-image generators us
 ing only four unguided sampling steps. Moreover, we systematically investi
 gate its scaling behavior and demonstrate LADD's effectiveness in various 
 applications such as image editing and inpainting.\n\nRegistration Categor
 y: Full Access, Full Access Supporter\n\nLanguage Format: English Language
 \n\nSession Chair: Michael Rubinstein (Google)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_572&sess=sess137
END:VEVENT
END:VCALENDAR