# Parallel Rollout Approximation for Pixel-Space Autoregressive Image Generation

Source: [Parallel Rollout Approximation for Pixel-Space Autoregressive Image Generation](<https://www.sota2.com/research/paper/6a41fbee206542251d7cca21>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-07-02T06:00:01.866Z

Authors: Jiayi Xu, Di He, Guolin Ke

Published: 2026-06-26T11:27:39.000Z

## Abstract

Pixel-space continuous-token autoregressive (AR) generation directly models images as sequences of raw pixel patches, avoiding discrete tokenization or a separately pretrained tokenizer. However, it faces coupled challenges: high-dimensional patch generation causes large single-step errors, and teacher-forced training creates a train--inference gap that makes these errors accumulate across AR steps. Existing fixes such as $x$-prediction and input noise injection only partially mitigate these issues. Exact rollout training better matches inference-time conditions, but is impractical due to prohibitively slow sequential sampling. We propose \\emph{Parallel Rollout Approximation} (PRA), a scalable framework that addresses both challenges jointly. PRA generates low-dimensional intermediate states instead of high-dimensional pixel patches, then maps them back to pixel-space tokens with a pixel decoder, preserving a pixel-in, pixel-out AR interface. It also constructs inference-like pixel inputs through the same intermediate-state-to-pixel path used at inference, independently across positions, approximating the pixel-feedback interface encountered during inference-time rollout while retaining parallel teacher-forced training. On class-conditional ImageNet-1K generation at $256\\times256$ resolution, PRA-S with 135M parameters achieves an FID of 2.58, surpassing the previous billion-scale pixel-space AR result of 3.60. Scaling to PRA-L with 511M parameters further improves FID to 1.94, establishing a new state of the art among pixel-space AR models. Beyond generation, PRA achieves higher ImageNet classification probing accuracy than other AR and diffusion baselines, suggesting its potential for unified pixel-space image generation and understanding.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Class-conditional Image Generation on ImageNet-1K 256x256 (train)](<https://www.sota2.com/research/sota/class-conditional-image-generation-on-imagenet-1k-256x256-train.md>): Class-conditional Image Generation; ImageNet-1K 256x256 1.0 (train); IS: 287.3.

- [Image Classification on ImageNet-1K (val) (Top-1 Score)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-val-top-1-score.md>): Image Classification; ImageNet-1k (val); Top-1 Acc: 68.8.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2606.27978>)

[PDF](<https://arxiv.org/pdf/2606.27978.pdf>)

[Implementation](<https://github.com/MangataX/PRA>)