Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusion

• URL: http://arxiv.org/abs/2505.21467v1
• Date: Tue, 27 May 2025 17:39:39 GMT
• Title: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusion
• Authors: Zhanqiu Hu, Jian Meng, Yash Akhauri, Mohamed S. Abdelfattah, Jae-sun Seo, Zhiru Zhang, Udit Gupta,
• Abstract summary: Diffusion language models offer parallel token generation and inherent bidirectionality.<n>State-of-the-art diffusion models (e.g., Dream 7B, LLaDA 8B) suffer from slow inference.<n>We introduce Guided Diffusion, a training-free method that uses a lightweight pretrained autoregressive model to supervise token unmasking.<n>For the first time, diffusion language models achieve a comparable and even faster latency as the widely adopted autoregressive models.
• Score: 16.99620863197586
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Diffusion language models offer parallel token generation and inherent bidirectionality, promising more efficient and powerful sequence modeling compared to autoregressive approaches. However, state-of-the-art diffusion models (e.g., Dream 7B, LLaDA 8B) suffer from slow inference. While they match the quality of similarly sized Autoregressive (AR) Models (e.g., Qwen2.5 7B, Llama3 8B), their iterative denoising requires multiple full-sequence forward passes, resulting in high computational costs and latency, particularly for long input prompts and long-context scenarios. Furthermore, parallel token generation introduces token incoherence problems, and current sampling heuristics suffer from significant quality drops with decreasing denoising steps. We address these limitations with two training-free techniques. First, we propose FreeCache, a Key-Value (KV) approximation caching technique that reuses stable KV projections across denoising steps, effectively reducing the computational cost of DLM inference. Second, we introduce Guided Diffusion, a training-free method that uses a lightweight pretrained autoregressive model to supervise token unmasking, dramatically reducing the total number of denoising iterations without sacrificing quality. We conduct extensive evaluations on open-source reasoning benchmarks, and our combined methods deliver up to a 34x end-to-end speedup without compromising accuracy. For the first time, diffusion language models achieve a comparable and even faster latency as the widely adopted autoregressive models. Our work successfully paved the way for scaling up the diffusion language model to a broader scope of applications across different domains.

Related papers

• Beyond Fixed: Variable-Length Denoising for Diffusion Large Language Models [74.15250326312179]
  Diffusion Large Language Models offer efficient parallel generation and capable global modeling.<n>The dominant application ofDLLMs is hindered by the need for a statically predefined generation length.<n>We introduce DAEDAL, a novel training-free denoising strategy that enables Dynamic Adaptive Length Expansion.
  arXiv  Detail & Related papers  (2025-08-01T17:56:07Z)
• 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)
• Communication-Efficient Diffusion Denoising Parallelization via Reuse-then-Predict Mechanism [18.655659400456848]
  Diffusion models have emerged as a powerful class of generative models across various modalities, including image, video, and audio synthesis.<n>We propose textbfParaStep, a novel parallelization method based on a reuse-then-predict mechanism that parallelizes diffusion inference by exploiting similarity between adjacent denoising steps.<n>ParaStep achieves end-to-end speedups of up to textbf3.88$times$ on SVD, textbf2.43$times$ on CogVideoX-2b, and textbf6.56$times
  arXiv  Detail & Related papers  (2025-05-20T06:58:40Z)
• AB-Cache: Training-Free Acceleration of Diffusion Models via Adams-Bashforth Cached Feature Reuse [19.13826316844611]
  Diffusion models have demonstrated remarkable success in generative tasks, yet their iterative denoising process results in slow inference.<n>We provide a theoretical understanding by analyzing the denoising process through the second-order Adams-Bashforth method.<n>We propose a novel caching-based acceleration approach for diffusion models, instead of directly reusing cached results.
  arXiv  Detail & Related papers  (2025-04-13T08:29:58Z)
• One-Step Diffusion Model for Image Motion-Deblurring [85.76149042561507]
  We propose a one-step diffusion model for deblurring (OSDD), a novel framework that reduces the denoising process to a single step.<n>To tackle fidelity loss in diffusion models, we introduce an enhanced variational autoencoder (eVAE), which improves structural restoration.<n>Our method achieves strong performance on both full and no-reference metrics.
  arXiv  Detail & Related papers  (2025-03-09T09:39:57Z)
• Generalized Interpolating Discrete Diffusion [65.74168524007484]
  Masked diffusion is a popular choice due to its simplicity and effectiveness.<n>We derive the theoretical backbone of a family of general interpolating discrete diffusion processes.<n>Exploiting GIDD's flexibility, we explore a hybrid approach combining masking and uniform noise.
  arXiv  Detail & Related papers  (2025-03-06T14:30:55Z)
• Optimizing for the Shortest Path in Denoising Diffusion Model [8.884907787678731]
  Shortest Path Diffusion Model (ShortDF) treats the denoising process as a shortest-path problem aimed at minimizing reconstruction error.<n>Experiments on multiple standard benchmarks demonstrate that ShortDF significantly reduces diffusion time (or steps)<n>This work, we suppose, paves the way for interactive diffusion-based applications and establishes a foundation for rapid data generation.
  arXiv  Detail & Related papers  (2025-03-05T08:47:36Z)
• CAT Pruning: Cluster-Aware Token Pruning For Text-to-Image Diffusion Models [5.406829638216823]
  Diffusion models have revolutionized generative tasks, especially in the domain of text-to-image synthesis.<n>However, their iterative denoising process demands substantial computational resources.<n>We present a novel acceleration strategy that integrates token-level pruning with caching techniques to tackle this computational challenge.
  arXiv  Detail & Related papers  (2025-02-01T13:46:02Z)
• Fast constrained sampling in pre-trained diffusion models [77.21486516041391]
  We propose an algorithm that enables fast and high-quality generation under arbitrary constraints.<n>During inference, we can interchange between gradient updates computed on the noisy image and updates computed on the final, clean image.<n>Our approach produces results that rival or surpass the state-of-the-art training-free inference approaches.
  arXiv  Detail & Related papers  (2024-10-24T14:52:38Z)
• Adv-KD: Adversarial Knowledge Distillation for Faster Diffusion Sampling [2.91204440475204]
  Diffusion Probabilistic Models (DPMs) have emerged as a powerful class of deep generative models. They rely on sequential denoising steps during sample generation. We propose a novel method that integrates denoising phases directly into the model's architecture.
  arXiv  Detail & Related papers  (2024-05-31T08:19:44Z)
• BOOT: Data-free Distillation of Denoising Diffusion Models with Bootstrapping [64.54271680071373]
  Diffusion models have demonstrated excellent potential for generating diverse images. Knowledge distillation has been recently proposed as a remedy that can reduce the number of inference steps to one or a few. We present a novel technique called BOOT, that overcomes limitations with an efficient data-free distillation algorithm.
  arXiv  Detail & Related papers  (2023-06-08T20:30:55Z)
• A Cheaper and Better Diffusion Language Model with Soft-Masked Noise [62.719656543880596]
  Masked-Diffuse LM is a novel diffusion model for language modeling, inspired by linguistic features in languages. Specifically, we design a linguistic-informed forward process which adds corruptions to the text through strategically soft-masking to better noise the textual data. We demonstrate that our Masked-Diffuse LM can achieve better generation quality than the state-of-the-art diffusion models with better efficiency.
  arXiv  Detail & Related papers  (2023-04-10T17:58:42Z)
• DiffusionBERT: Improving Generative Masked Language Models with Diffusion Models [81.84866217721361]
  DiffusionBERT is a new generative masked language model based on discrete diffusion models. We propose a new noise schedule for the forward diffusion process that controls the degree of noise added at each step. Experiments on unconditional text generation demonstrate that DiffusionBERT achieves significant improvement over existing diffusion models for text.
  arXiv  Detail & Related papers  (2022-11-28T03:25:49Z)

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.