Diffusion models have gained prominence in generating high-quality sequences of text. Nevertheless, current approaches predominantly represent discrete text within a continuous diffusion space, which incurs substantial computational overhead during training and results in slower sampling speeds. In this paper, we introduce a soft absorbing state that facilitates the diffusion model in learning to reconstruct discrete mutations based on the underlying Gaussian space, thereby enhancing its capacity to recover conditional signals. During the sampling phase, we employ state-of-the-art ODE solvers within the continuous space to expedite the sampling process. Comprehensive experimental evaluations reveal that our proposed method effectively accelerates the training convergence by 4x and generates samples of similar quality 800x faster, rendering it significantly closer to practical application. \footnote{The code is released at \url{https://github.com/Shark-NLP/DiffuSeq}

翻译：扩散模型在生成高质量文本序列方面已崭露头角。然而，当前方法主要将离散文本表征于连续扩散空间中，这在训练过程中会带来大量计算开销，并导致采样速度较慢。本文引入一种软吸收状态，使扩散模型能基于高斯空间学习重构离散突变，从而增强其恢复条件信号的能力。在采样阶段，我们在连续空间中采用最先进的ODE求解器以加速采样过程。全面的实验评估表明，我们提出的方法能有效将训练收敛速度提升4倍，并在生成质量相近的样本时实现800倍的加速，使其显著更接近实际应用。\footnote{代码已开源在 \url{https://github.com/Shark-NLP/DiffuSeq}}