The Year in AI Papers
The year/Independent research

Paper 2510.06590

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer

Ziyuan Huang, DanDan Zheng, Cheng Zou, Rui Liu, Xiaolong Wang, Kaixiang Ji, Weilong Chai, Jianxin Sun, Libin Wang, Yongjie Lv, Taozhi Huang, Jiajia Liu, et al.

Published
Oct 2025
Research lab
Independent
Citations
26
GitHub
143 stars

01 In brief

Summary

Ming-UniVision introduces MingTok, a continuous visual tokenizer with a three-stage architecture (low-level encoder, semantic decoder, pixel decoder) that unifies image understanding and generation in a single autoregressive framework.

Unlike discrete tokenizers, MingTok avoids quantization errors, balancing compact latents for generation with high-dimensional semantic features for understanding.

Ming-UniVision casts both tasks as next-token prediction in a shared continuous space, enabling multi-round in-context editing and reducing visual tokens by up to 66% compared to prior unified models.

Evaluations show state-of-the-art GenEval performance (0.85 overall) and competitive understanding results, though OCR and fine-grained editing lag due to latent compression.

Ablations confirm that a unified representation improves both tasks, and a Visualized Chain-of-Thought method enhances editing consistency.

Limitations include lack of interleaved pretraining and high per-token detail density, with future work planned on higher-resolution tokenization and longer editing sequences.

02 From the paper

Abstract

Visual tokenization remains a core challenge in unifying visual understanding and generation within the autoregressive paradigm. Existing methods typically employ tokenizers in discrete latent spaces to align with the tokens from large language models, where the quantization errors can limit semantic expressiveness and degrade the capability of vision-language understanding. To address this, we introduce MingTok, a new family of visual tokenizers with a continuous latent space, for unified autoregressive generation and understanding. While understanding tasks favor discriminative high-dimensional features, generation tasks prefer compact low-level codes. Thus, to reconcile these competing demands, MingTok adopts a three-stage sequential architecture involving low-level encoding, semantic expansion, and visual reconstruction. Built on top of it, Ming-UniVision eliminates the need for task-specific visual representations, and unifies diverse vision-language tasks under a single autoregrsssive prediction paradigm. By formulating both understanding and generation as next-token prediction in a shared continuous space, it seamlessly supports multi-round, in-context tasks such as iterative understanding, generation and editing. Empirically, we find that using a unified continuous visual representation reconciles the competing requirements on the tokenizers by the understanding and generation tasks, thereby leading to state-of-the-art level performance across both domains. We hope our findings will facilitate unified visual tokenization in the continuous domain. Inference code and model weights are released to benefit community.