A training framework perturbs self-conditioning signals in diffusion language models to match few-step inference noise, enabling up to 400x faster sampling while surpassing standard continuous diffusion performance on sequence-to-sequence tasks.
InFindings of EMNLP, pages 2401–2410
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
FastDiSS: Few-step Match Many-step Diffusion Language Model on Sequence-to-Sequence Generation--Full Version
A training framework perturbs self-conditioning signals in diffusion language models to match few-step inference noise, enabling up to 400x faster sampling while surpassing standard continuous diffusion performance on sequence-to-sequence tasks.