SAMTok: Representing Any Mask with Two Words

TL;DR

SAMTok introduces a novel discrete mask tokenizer that enables large multimodal language models to learn pixel-wise understanding and generation capabilities efficiently, without architectural changes or complex training procedures.

Contribution

The paper proposes SAMTok, a simple yet effective mask tokenizer that converts masks into tokens, allowing existing MLLMs to acquire pixel-wise skills through standard training methods.

Findings

Achieves state-of-the-art results on multiple pixel-wise tasks.

Enables MLLMs to learn pixel-level understanding without architectural modifications.

Improves performance on GRES and GCG benchmarks with reinforcement learning.

Abstract

Pixel-wise capabilities are essential for building interactive intelligent systems. However, pixel-wise multi-modal LLMs (MLLMs) remain difficult to scale due to complex region-level encoders, specialized segmentation decoders, and incompatible training objectives. To address these challenges, we present SAMTok, a discrete mask tokenizer that converts any region mask into two special tokens and reconstructs the mask using these tokens with high fidelity. By treating masks as new language tokens, SAMTok enables base MLLMs (such as the QwenVL series) to learn pixel-wise capabilities through standard next-token prediction and simple reinforcement learning, without architectural modifications and specialized loss design. SAMTok builds on SAM2 and is trained on 209M diverse masks using a mask encoder and residual vector quantizer to produce discrete, compact, and information-rich tokens. With 5M SAMTok-formatted mask understanding and generation data samples, QwenVL-SAMTok attains state-of-the-art or comparable results on region captioning, region VQA, grounded conversation, referring segmentation, scene graph parsing, and multi-round interactive segmentation. We further introduce a textual answer-matching reward that enables efficient reinforcement learning for mask generation, delivering substantial improvements on GRES and GCG benchmarks. Our results demonstrate a scalable and straightforward paradigm for equipping MLLMs with strong pixel-wise capabilities. Our code and models are available.