Diffusion language models (DLMs) have attracted wide attention for their fast, parallel decoding, but they are believed to be far less efficient to train tha…
机构:Cornell
来源:arXiv 2610.17590 | AI4Papers 论文推荐平台