AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

DPAR: Dynamic Patchification for Efficient Autoregressive Visual Generation

Divyansh Srivastava, Akshay Mehra, Pranav Maneriker, Debopam Sanyal, Vishnu Raj, Vijay Kamarshi, Fan Du, Joshua Kimball

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 2 citations (OpenAlex)

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
add entropy gating alone to static patches (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.58 -> 3.91Hurt
add max patch length on top of entropy gating (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.91 -> 3.45Improved
add row-boundary resets (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.45 -> 3.32Improved
full entropy patchification vs static patches (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.58 -> 3.32Improved
use entropy threshold 7.8 instead of 8.0 (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.43 -> 3.32Mixed
raise max patch length from 1 to 4 (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.73 -> 3.32Improved
raise max patch length from 4 to 16 (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.32 -> 3.61Hurt
raise inference entropy threshold to 8.1 (dynamic training) (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.32 -> 3.39No clear effect
raise inference patch length to 2.16 (static training) (ablation)Static patch model · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.58 -> 25.59Hurt
train with dynamic patches rather than raster tokens (linear probe)DPAR-L vs LlamaGen-L · ImageNet linear probeTop-1 accuracy (%): 32.62 -> 37.82Improved
add repeated start/end encodings to dynamic RoPE (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.42 -> 3.31Improved
use 1 encoder layer and 4 decoder layers instead of 4 and 1 (ablation)DPAR-L · ImageNet 256 (DPAR-L, 50 epochs)FID-50K: 3.85 -> 3.32Improved

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