DT-UFC: Universal Large Model Feature Coding via Peaky-to-Balanced Distribution Transformation

Like image coding in visual data transmission, feature coding is essential for the distributed deployment of large models by significantly reducing transmission and storage overhead. However, prior studies have mostly targeted task- or model-specific scenarios, leaving the challenge of universal feature coding across diverse large models largely unaddressed. In this paper, we present the first systematic study on universal feature coding for large models. The key challenge lies in the inherently diverse and distributionally incompatible nature of features extracted from different models. For example, features from DINOv2 exhibit highly peaky, concentrated distributions, while those from Stable Diffusion 3 (SD3) are more dispersed and uniform. This distributional heterogeneity severely hampers both compression efficiency and cross-model generalization. To address this, we propose a learned peaky-to-balanced distribution transformation, which reshapes highly skewed feature distributions into a common, balanced target space. This transformation is non-uniform, data-driven, and plug-and-play, enabling effective alignment of heterogeneous distributions without modifying downstream codecs. With this alignment, a universal codec trained on the balanced target distribution can effectively generalize to features from different models and tasks. We validate our approach on three representative large models-LLaMA3, DINOv2, and SD3-across multiple tasks and modalities. Extensive experiments show that our method achieves notable improvements in both compression efficiency and cross-model generalization over task-specific baselines. All source code will be released for future research.

@article{gao2025_2506.16495,
  title={ DT-UFC: Universal Large Model Feature Coding via Peaky-to-Balanced Distribution Transformation },
  author={ Changsheng Gao and Zijie Liu and Li Li and Dong Liu and Xiaoyan Sun and Weisi Lin },
  journal={arXiv preprint arXiv:2506.16495},
  year={ 2025 }
}