works on
attention-free 1efficient inference 1entropy-based selection 1token reduction 1vision-language models 1

From the 1 of 8 linked papers with an AI index.

collaborators
Showing cs.CLShow all

1 paper · 1 filter

cs.CL2026

OPTD: On-Policy Transition Distillation with Consistency-Guided Adaptive Compression for Few-Step Diffusion Language Models

Xiaocheng Lu, Hualei Zhang, Shuhan Guo +8

Diffusion language models (dLLMs) can predict many tokens in parallel, but accurate generation still requires many iterative denoising steps. Few-step distillation accelerates deco…