arXiv:2605. 29233v2 Announce Type: replace-cross Abstract: Diffusion language models (dLLMs) generate text by iteratively denoising multiple token positions in parallel, offering an attractive alternative to strictly autoregressive decoding.
The paper introduces Window-Diffusion, a method that accelerates diffusion language model inference by pruning and caching tokens within a sliding window. It categorizes undecoded tokens into active, buffer, and far-field groups, computing only the first two while discarding the rest. Experiments on LLaDA and Dream demonstrate up to 99× speedup with minimal loss in generation quality.
DynaTE is a hardware–software co‑design that accelerates diffusion‑based large language models (dLLMs) by dynamically adapting execution to token states. It skips low‑utility token computations, uses a dimension‑reconfigurable processing array, refines locally dependent tokens with FLDD, and streams vocabulary to handle irregular token patterns. These techniques yield 2.05–2.78× speedup and 2.99–3.93× higher energy efficiency compared to existing dLLM accelerators, and outperform Jetson AGX Orin by 2.55× speedup and 6.07× energy efficiency.
The paper proposes Predictive-State Flows (PSF), a new architecture for continuous diffusion language models that separates the diffusion prior’s role to generate high‑level predictive states and the decoder’s role to resolve token‑level choices. PSF partitions sequences into blocks, uses a rectified‑flow prior to produce predictive states, and then a block‑causal decoder generates all blocks in parallel, reducing token‑level sequential depth. Experiments on OpenWebText show that a 197M‑parameter PSF model with eight predictive states achieves a generative perplexity of 22.8 and a MAUVE@512 of 0.93 while using far fewer FLOPs than a comparable ELF‑B model.
Diffusion Language Models (DLMs) exhibit strong parallel decoding capabilities by denoising multiple tokens in a single generation step. However, this parallelism comes with substantial computational...
The paper introduces a training‑free structured suffix modeling technique to accelerate Diffusion Language Models (DLMs). It partitions the suffix into local, middle, and tail regions, retaining varying numbers of tokens per region and incorporating previous decoding results into current token representations. Experiments on three DLMs show significant speedups—up to 72.81× in long‑sequence inference—while often improving performance, and the method is compatible with existing acceleration strategies.
arXiv:2508. 10875v3 Announce Type: replace-cross Abstract: Diffusion Language Models (DLMs) are rapidly emerging as a powerful and promising alternative to the dominant autoregressive (AR) paradigm.
arXiv:2609. 04010v1 Announce Type: new Abstract: Large Language Models (LLMs) owe much of their success to next-token prediction (NTP), but their autoregressive (AR) structure requires slow, sequential token generation.
The paper introduces Block Parallelism (BP) and Context‑Sharded Block Parallelism (CSBP) to improve training efficiency for Block Diffusion Language Models (BDLMs) with long contexts. By assigning each corrupted‑block computation to a separate rank and sharding the shared clean sequence, CSBP reduces communication overhead and memory usage while preserving training semantics. Experiments on 16 H200 GPUs and 8 H100 GPUs show throughput gains of up to 1.61× and 7.59×, respectively, and higher benchmark pass rates in practical fine‑tuning scenarios.
Flash-dLLM is a training‑free inference acceleration framework that improves the speed and memory efficiency of Diffusion Large Language Models (dLLMs). It tackles GPU memory I/O bottlenecks by introducing an I/O‑aware fused KV‑cache kernel and then employs a draft‑and‑verify decoding strategy that uses the dLLM itself as both drafter and verifier. Experiments on mathematical reasoning and code‑generation tasks show Flash‑dLLM outperforms existing acceleration methods, achieving up to 11.0× speedups over the Elastic‑Cache baseline.
The paper investigates how block‑diffusion language models can use a constant‑size cache to enable efficient parallel decoding. By employing sequence mixers that summarize completed blocks into a reusable state and a block‑causal training objective, the authors pretrain three 3B block‑diffusion denoisers (attention, Mamba, and hybrid) on 300 B tokens. The resulting state‑space cache remains O(1) in memory and latency regardless of context length, yielding significant speed‑up and memory savings compared to traditional attention‑based caches, especially at very long sequences.
arXiv:2606. 26120v1 Announce Type: cross Abstract: Diffusion Large Language Models (dLLMs) offer a promising alternative to autoregressive models, excelling in text generation tasks due to their bidirectional attention mechanisms.