6 papers
Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation
Changcheng Li, Jiancan Wu, Hengheng Zhang +5
Reliable deployment of large language models (LLMs) requires accurate uncertainty estimation. Existing methods are predominantly answer-first, producing confidence only after gener…
BeCARE: Budgeted Cache Refresh for Diffusion Transformer Acceleration
Yuhang Zhang, Junxiang Qiu, Huixia Ben +4
Training-free feature caching accelerates diffusion transformer (DiT) inference by reusing or forecasting intermediate features. However, fixed schedules make compute predictable b…
Punctuation-aware Hybrid Trainable Sparse Attention for Large Language Models
Junxiang Qiu, Shuo Wang, Zhengsu Chen +4
Attention serves as the fundamental mechanism for long-context modeling in large language models (LLMs), yet dense attention becomes structurally prohibitive for long sequences due…
Accelerating Controllable Generation via Hybrid-grained Cache
Lin Liu, Huixia Ben, Shuo Wang +4
Controllable generative models have been widely used to improve the realism of synthetic visual content. However, such models must handle control conditions and content generation…
Accelerating Diffusion Transformer via Gradient-Optimized Cache
Junxiang Qiu, Lin Liu, Shuo Wang +3
Feature caching has emerged as an effective strategy to accelerate diffusion transformer (DiT) sampling through temporal feature reuse. It is a challenging problem since (1) Progre…
Accelerating Diffusion Transformer via Error-Optimized Cache
Junxiang Qiu, Shuo Wang, Jinda Lu +4
Diffusion Transformer (DiT) is a crucial method for content generation. However, it needs a lot of time to sample. Many studies have attempted to use caching to reduce the time con…