AI RESEARCH
Predict-then-Diffuse: Adaptive Response Length for Compute-Budgeted Inference in Diffusion LLMs
arXiv CS.AI
•
ArXi:2605.04215v1 Announce Type: cross Diffusion-based Large Language Models (D-LLMs) represent a promising frontier in generative AI, offering fully parallel token generation that can lead to significant throughput advantages and superior GPU utilization over traditional autoregressive paradigm. However, this parallelism is constrained by the requirement of a fixed-size response length prior to generation.