Related papers: Fast-dLLM: Training-free Acceleration of Diffusion…

Auto-regressive models (ARMs) have established a dominant paradigm in language modeling. However, their strictly sequential decoding paradigm imposes fundamental constraints on both inference efficiency and modeling flexibility. To address…

Accelerating the inference of large language models (LLMs) is a critical challenge in generative AI. Speculative decoding (SD) methods offer substantial efficiency gains by generating multiple tokens using a single target forward pass.…

Generating long sequences of tokens given a long-context input is a very compute-intensive inference scenario for large language models (LLMs). One prominent inference speed-up approach is to construct a smaller key-value (KV) cache,…

The transformer's context window is vital for tasks such as few-shot learning and conditional generation as it preserves previous tokens for active memory. However, as the context lengths increase, the computational costs grow…

Diffusion large language models (dLLMs) have emerged as a promising alternative for text generation, distinguished by their native support for parallel decoding. In practice, block inference is crucial for avoiding order misalignment in…

Inference accounts for the majority of latency and energy consumption in large language model (LLM) deployments, often exceeding 90% of total cost. While training-time efficiency has seen extensive progress, runtime optimization remains a…

Recent advancements in Large Language Models (LLMs) have spurred interest in numerous applications requiring robust long-range capabilities, essential for processing extensive input contexts and continuously generating extended outputs. As…

The growth of long-context Large Language Models (LLMs) significantly increases memory and bandwidth pressure during autoregressive decoding due to the expanding Key-Value (KV) cache. While accuracy-preserving KV-cache quantization (e.g.,…

Diffusion Language Models (DLMs) offer a promising alternative for language modeling by enabling parallel decoding through iterative refinement. However, most DLMs rely on hard binary masking and discrete token assignments, which hinder the…

The deployment of efficient long-context LLMs in applications like autonomous agents, long-chain reasoning, and creative writing is fundamentally bottlenecked by the linear growth of KV cache memory. Existing compression and eviction…

While Diffusion Language Models (DLMs) offer a flexible, arbitrary-order alternative to the autoregressive paradigm, their non-causal nature precludes standard KV caching, forcing costly hidden state recomputation at every decoding step.…

Key-value (KV) caching has become the de-facto to accelerate generation speed for large language models (LLMs) inference. However, the growing cache demand with increasing sequence length has transformed LLM inference to be a memory bound…

Diffusion language models promise parallel generation, yet still lag behind autoregressive (AR) models in quality. We stem this gap to a failure of introspective consistency: AR models agree with their own generations, while DLMs often do…

Recent advancements in large language models (LLMs) have significantly enhanced their knowledge and generative capabilities, leading to a surge of interest in leveraging LLMs for high-quality data synthesis. However, synthetic data…

Recent advancements in Large Visual Language Models (LVLMs) have gained significant attention due to their remarkable reasoning capabilities and proficiency in generalization. However, processing a large number of visual tokens and…

The impressive capabilities of Large Language Models (LLMs) come at the cost of substantial computational resources during deployment. While KV Cache can significantly reduce recomputation during inference, it also introduces additional…

Recent large-scale vision-language models (VLMs) have shown remarkable text-to-image generation capabilities, yet their visual fidelity remains constrained by the discrete image tokenization, which poses a major challenge. Although several…

Diffusion models demonstrate outstanding performance in image generation, but their multi-step inference mechanism requires immense computational cost. Previous works accelerate inference by leveraging layer or token cache techniques to…

Parallel decoding methods such as Jacobi decoding show promise for more efficient LLM inference as it breaks the sequential nature of the LLM decoding process and transforms it into parallelizable computation. However, in practice, it…

Diffusion large language models (dLLMs) present a promising alternative to dominant autoregressive models (ARMs) by the ability of parallel decoding at the expense of substantial computation and memory costs. Specifically, the cache…