DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models
This work addresses the slow training and sampling speeds in text generation diffusion models, making them more practical for real-world applications.
The paper tackles the computational inefficiency of text diffusion models by introducing a soft absorbing state that bridges discrete and continuous spaces, achieving 4x faster training convergence and 800x faster sampling while maintaining similar output quality.
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}