Efficient-DLM: From Autoregressive to Diffusion Language Models, and Beyond in Speed

ArXi:2512.14067v2 Announce Type: replace-cross Diffusion language models (dLMs) have emerged as a promising paradigm that enables parallel, non-autoregressive generation, but their learning efficiency lags behind that of autoregressive (AR) language models when trained from scratch. To this end, we study AR-to-dLM conversion to transform pretrained AR models into efficient dLMs that excel in speed while preserving AR models' task accuracy.