Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DMax: Aggressive Parallel Decoding for dLLMs

Zigeng Chen, Gongfan Fang, Xinyin Ma, Ruonan Yu, Xinchao Wang

Abstract

We present DMax, a new paradigm for efficient diffusion language models (dLLMs). It mitigates error accumulation in parallel decoding, enabling aggressive decoding parallelism while preserving generation quality. Unlike conventional masked dLLMs that decode through a binary mask-to-token transition, DMax reformulates decoding as a progressive self-refinement from mask embeddings to token embeddings. At the core of our approach is On-Policy Uniform Training, a novel training strategy that efficiently unifies masked and uniform dLLMs, equipping the model to recover clean tokens from both masked inputs and its own erroneous predictions. Building on this foundation, we further propose Soft Parallel Decoding. We represent each intermediate decoding state as an interpolation between the predicted token embedding and the mask embedding, enabling iterative self-revising in embedding space. Extensive experiments across a variety of benchmarks demonstrate the effectiveness of DMax. Compared with the original LLaDA-2.0-mini, our method improves TPF on GSM8K from 2.04 to 5.47 while preserving accuracy. On MBPP, it increases TPF from 2.71 to 5.86 while maintaining comparable performance. On two H200 GPUs, our model achieves an average of 1,338 TPS at batch size 1. Code is available at: https://github.com/czg1225/DMax

DMax: Aggressive Parallel Decoding for dLLMs

Abstract

We present DMax, a new paradigm for efficient diffusion language models (dLLMs). It mitigates error accumulation in parallel decoding, enabling aggressive decoding parallelism while preserving generation quality. Unlike conventional masked dLLMs that decode through a binary mask-to-token transition, DMax reformulates decoding as a progressive self-refinement from mask embeddings to token embeddings. At the core of our approach is On-Policy Uniform Training, a novel training strategy that efficiently unifies masked and uniform dLLMs, equipping the model to recover clean tokens from both masked inputs and its own erroneous predictions. Building on this foundation, we further propose Soft Parallel Decoding. We represent each intermediate decoding state as an interpolation between the predicted token embedding and the mask embedding, enabling iterative self-revising in embedding space. Extensive experiments across a variety of benchmarks demonstrate the effectiveness of DMax. Compared with the original LLaDA-2.0-mini, our method improves TPF on GSM8K from 2.04 to 5.47 while preserving accuracy. On MBPP, it increases TPF from 2.71 to 5.86 while maintaining comparable performance. On two H200 GPUs, our model achieves an average of 1,338 TPS at batch size 1. Code is available at: https://github.com/czg1225/DMax

Paper Structure

This paper contains 11 sections, 10 equations, 4 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Methodology
  4. On-Policy Uniform Training
  5. Soft Parallel Decoding
  6. Experiments
  7. Experimental Setup
  8. Experimental Results
  9. Ablation Study
  10. Related Work
  11. Conclusion

Figures (4)

  • Figure 1: Comparison between the original LLaDA-2.0-mini and our proposed DMax. Unlike the original binary mask-to-token decoding process, DMax introduces a self-revising mask-to-hybrid-embedding-to-token process, enabling highly parallel decoding without accuracy dropping.
  • Figure 2: Overview of the proposed On-Policy Uniform Training.
  • Figure 3: Overview of the Soft Parallel Decoding procedure in DMax.
  • Figure 4: Comparison of accuracy-TPF trade-off curves between original LLaDA-2.0-mini model and our method. We present curves on GSM8K, MATH500, HumanEval and MBPP benchmarks.