AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

D-AR: Diffusion via Autoregressive Models

Ziteng Gao, Mike Zheng Shou

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Discrete diffusion solvers

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use linear CFG schedule instead of no CFG (ablation)D-AR-L · ImageNet 256x256 class-conditionalgFID: 7.43 -> 2.44Improved
use linear CFG schedule instead of constant 1.5 (ablation)D-AR-XL · ImageNet 256x256 class-conditionalgFID: 3.39 -> 2.09Improved
increase time-shift beta from 1 to 2 (ablation)sequential diffusion tokenizer · ImageNet val (lightweight tokenizer)rFID: 39.75 -> 28.65Improved
increase time-shift beta from 2 to 4 (ablation)sequential diffusion tokenizer · ImageNet val (lightweight tokenizer)rFID: 28.65 -> 27.10No clear effect
increase conditioning groups K from 8 to 16 (ablation)sequential diffusion tokenizer · ImageNet val (lightweight tokenizer)rFID: 28.65 -> 49.07Hurt
decrease conditioning groups K from 8 to 4 (ablation)sequential diffusion tokenizer · ImageNet val (lightweight tokenizer)rFID: 28.65 -> 23.04Improved
use K=8 vs K=4 tokens for AR training (ablation)D-AR-B · ImageNet 256x256gFID: 53.70 -> 54.26No clear effect
use K=8 vs K=16 tokens for AR training (ablation)D-AR-B · ImageNet 256x256gFID: 63.69 -> 54.26Improved
generate more AR tokens (64 to 256) before decodingD-AR-L · ImageNet 256x256gFID: 7.38 -> 2.44Improved
replace LlamaGen tokenizer with sequential diffusion tokenizertokenizer, 256 tokens · ImageNet valrFIDImproved
do not use adversarial loss (ablation)sequential diffusion tokenizer · ImageNettraining stabilityNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy