الملخص
تمت ترجمة أجزاء من هذه الصفحة آلياً وقد تحتوي على أخطاء.
Diffusion Transformers (DiTs) have become the standard backbone for high-quality generative modeling, yet deploying them in conditional generation tasks remains computationally prohibitive because bidirectional joint attention repeatedly processes large reference streams. While existing optimization schemes mitigate generic temporal redundancy, they typically rely on coarse-grained static reuse and overlook the distinct dynamics of references and targets. Specifically, we observe that reference representations often evolve slowly along the generation trajectory, while the target often assigns little attention mass to them; reference drift and this target-to-reference exposure jointly shape how strongly stale reference states affect the target. To exploit these patterns, we introduce \RefAdapt, a training-free framework for adaptive control of joint attention between references and targets. Instead of rigid static strategies, \RefAdapt combines consecutive target-Q change with previously observed target-to-reference attention mass to control reference computation adaptively at block granularity. Under ultra-few-step settings, \RefAdapt enables speedups of up to $2.097\times$ on 4-step MiniMax H3 and $3.54\times$ on 8-step Qwen Image Edit, while maintaining comparable visual quality.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Tang, J., Fan, J., Zhou, Q., Liu, Q., Bian, J., & Li, Z. (2026). RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع. https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers
MLA 9
Tang, Jian, et al. "RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع." https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers.
شيكاغو (المؤلف–التاريخ)
Tang, Jian, Jiawei Fan, Qiannan Zhou, Qingbin Liu, Jiang Bian, and Zang Li. 2026. "RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع." https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers.
هارفارد
Tang, J., Fan, J., Zhou, Q., Liu, Q., Bian, J. and Li, Z. (2026) 'RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع', Available at: https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers.
فانكوفر
Tang J, Fan J, Zhou Q, Liu Q, Bian J, Li Z. RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع. https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers
IEEE
J. Tang, J. Fan, Q. Zhou, Q. Liu, J. Bian, and Z. Li, "RefAdapt-DiT: الانتباه المشترك التكيفي لمحوّلات الانتشار المشروطة بالمرجع," https://omanscience.com/ar/articles/refadapt-dit-adaptive-joint-attention-for-reference-conditioned-diffusion-transformers.