Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing

• URL: http://arxiv.org/abs/2508.09192v1
• Date: Fri, 08 Aug 2025 04:51:37 GMT
• Title: Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing
• Authors: Xu Wang, Chenkai Xu, Yijie Jin, Jiachun Jin, Hao Zhang, Zhijie Deng,
• Abstract summary: Diffusion Large Language Models (dLLMs) have emerged as a promising alternative to autoregressive (AR) LLMs for text generation.<n>This paper breaks this barrier based on a simple and effective strategy named discrete diffusion forcing (D2F)<n>In this way, vanilla dLLMs are refurbished into an AR-diffusion hybrid paradigm for efficient inference.
• Score: 14.22753953706955
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Diffusion Large Language Models (dLLMs) have emerged as a promising alternative to autoregressive (AR) LLMs for text generation, with the potential to decode multiple tokens in a single iteration. However, none of the existing open-source dLLMs have achieved superior inference speed over AR LLMs of similar size. This paper breaks this barrier based on a simple and effective strategy named discrete diffusion forcing (D2F). D2F equips dLLMs with two key capabilities: (1) block-wise autoregressive generation to enable KV cache utilization; (2) prediction of following tokens without requiring completion of prior blocks for inter-block parallel decoding. In this way, the vanilla dLLMs are refurbished into an AR-diffusion hybrid paradigm for efficient inference. D2F can be implemented with an asymmetric distillation process based on pre-trained dLLMs. We further propose a pipelined parallel decoding algorithm, which enables a trade-off between efficiency and efficacy. Empirically, D2F dLLMs achieve more than $\mathbf{2.5\times}$ inference speed than LLaMA3 and Qwen2.5 on GSM8K. Compared to vanilla dLLMs like LLaDA and Dream, the acceleration can be more than $\mathbf{50\times}$ while maintaining comparable output quality. The code is available at https://github.com/zhijie-group/Discrete-Diffusion-Forcing.

Related papers

• Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models [96.0074341403456]
  Inference-time compute has re-emerged as a practical way to improve LLM reasoning.<n>Most test-time scaling (TTS) algorithms rely on autoregressive decoding.<n>We propose Prism, an efficient TTS framework for dLLMs.
  arXiv  Detail & Related papers  (2026-02-02T09:14:51Z)
• Residual Context Diffusion Language Models [90.07635240595926]
  Residual Context Diffusion (RCD) is a module that converts discarded token representations into contextual residuals and injects them back for the next denoising step.<n>RCD consistently improves frontier dLLMs by 5-10 points in accuracy with minimal extra computation overhead.
  arXiv  Detail & Related papers  (2026-01-30T13:16:32Z)
• VidLaDA: Bidirectional Diffusion Large Language Models for Efficient Video Understanding [52.69880888587866]
  Current Video Large Language Models (Video LLMs) typically encode frames via a encoder vision and employ an autoregressive (AR) LLM for understanding and generation.<n>We propose VidLaDA, a Diffusion Video LLM based on Language Models (DLMs) that leverages bidirectional attention to unlock comprehensive modeling and decode tokens in parallel.<n>Experiments show VidLaDA rivals state-of-the-art AR baselines and outperforms DLM baselines, with MARS-Cache delivering over 12x speedup without compromising accuracy.
  arXiv  Detail & Related papers  (2026-01-25T15:02:01Z)
• d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation [31.922313594074925]
  Diffusion large language models (dLLMs) offer capabilities beyond those of autoregressive (AR) LLMs.<n>Current methods typically focus on only one-side of the coin, targeting either efficiency or performance.<n>We propose d3LLM (Pseudo-Distilled Diffusion Large Language Model), striking a balance between accuracy and parallelism.
  arXiv  Detail & Related papers  (2026-01-12T14:25:36Z)
• From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs [58.640039233470766]
  We show that principled AR-to-block-diffusion adaptation is an effective and compute-efficient alternative to training DLMs from scratch.<n> NBDiff-7B (Base and Instruct) could inherit the long-context modeling and reasoning capabilities, and achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-12-07T10:28:21Z)
• Diffusion LLM with Native Variable Generation Lengths: Let [EOS] Lead the Way [23.877854550033224]
  Diffusion-based large language models (dLLMs) have exhibited substantial potential for parallel text generation.<n>Current dLLMs suffer from fixed generation lengths, which indicates the generation lengths of dLLMs have to be determined before decoding.<n>We propose to train a diffusion LLM with native variable generation lengths, abbreviated as dLLM-Var.
  arXiv  Detail & Related papers  (2025-10-28T16:32:43Z)
• dInfer: An Efficient Inference Framework for Diffusion Language Models [54.80918957287927]
  Diffusion-based large language models (dLLMs) have emerged as a promising alternative to autoregressive (AR) LLMs.<n>We present dInfer, an efficient and framework for dLLM inference.
  arXiv  Detail & Related papers  (2025-10-09T16:19:42Z)
• Fast-dLLM v2: Efficient Block-Diffusion LLM [64.38006546510337]
  Fast-dLLM v2 is a block diffusion language model that adapts pretrained AR models into dLLMs for parallel text generation.<n>This represents a 500x reduction in training data compared to full-attention diffusion LLMs such as Dream (580B tokens)
  arXiv  Detail & Related papers  (2025-09-30T14:40:18Z)
• Spiffy: Multiplying Diffusion LLM Acceleration via Lossless Speculative Decoding [40.96405124314983]
  Diffusion LLMs (dLLMs) have recently emerged as a powerful alternative to autoregressive LLMs (AR-LLMs)<n>Currently available open-source dLLMs often generate at much lower rates, typically decoding only a single token at every denoising timestep.<n>We present Spiffy, a speculative decoding algorithm that accelerates dLLM inference by $mathbf2.8-3.1times$ while provably preserving the model's output distribution.
  arXiv  Detail & Related papers  (2025-09-22T17:58:21Z)
• DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation [68.19756761027351]
  Diffusion large language models (dLLMs) are compelling alternatives to autoregressive (AR) models.<n>We investigate their denoising processes and reinforcement learning methods.<n>Our work provides deeper insight into the machinery of dLLM generation and offers an effective, diffusion-native RL training framework.
  arXiv  Detail & Related papers  (2025-06-25T17:35:47Z)
• Accelerating Diffusion LLMs via Adaptive Parallel Decoding [50.9948753314669]
  We introduce adaptive parallel decoding (APD), a novel method that dynamically adjusts the number of tokens sampled in parallel.<n>APD provides markedly higher throughput with minimal quality degradations on downstream benchmarks.
  arXiv  Detail & Related papers  (2025-05-31T06:10:10Z)
• Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding [51.711605076319216]
  Diffusion-based large language models (Diffusion LLMs) have shown promise for non-autoregressive text generation with parallel decoding capabilities.<n>We introduce a novel block-wise approximate KV Cache mechanism tailored for bidirectional diffusion models, enabling cache reuse with negligible performance drop.<n>We propose a confidence-aware parallel decoding strategy that selectively decodes tokens exceeding a confidence threshold, mitigating dependency violations and maintaining generation quality.
  arXiv  Detail & Related papers  (2025-05-28T17:39:15Z)
• SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration [10.970637831760136]
  Speculative decoding (SD) has emerged as a widely used paradigm to accelerate LLM inference without compromising quality.<n>We introduce SWIFT, an on-the-fly self-speculative decoding algorithm that adaptively selects intermediate layers of LLMs to skip during inference.<n>Our experiments demonstrate that SWIFT can achieve over a 1.3x-1.6x speedup while preserving the original distribution of the generated text.
  arXiv  Detail & Related papers  (2024-10-09T14:15:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.