paper-with-me

Papers

DiTFastAttnV2: Head-wise Attention Compression for Multi-Modality Diffusion Transformers

2025-03-28 · Hanling Zhang, Rundong Su, Zhihang Yuan, Pengtao Chen, Mingzhu Shen Yibo Fan, Shengen Yan, Guohao Dai, Yu Wang

Text-to-image generation models, especially Multimodal Diffusion Transformers (MMDiT), have shown remarkable progress in generating high-quality images. However, these models often face significant computational bottlenecks, particularly in attention mechanisms, which hinder their scalability and efficiency. In this paper, we introduce DiTFastAttnV2, a post-training compression method designed to accelerate attention in MMDiT. Through an in-depth analysis of MMDiT's attention patterns, we identify key differences from prior DiT-based methods and propose head-wise arrow attention and caching mechanisms to dynamically adjust attention heads, effectively bridging this gap. We also design an Efficient Fused Kernel for further acceleration. By leveraging local metric methods and optimization techniques, our approach significantly reduces the search time for optimal compression schemes to just minutes while maintaining generation quality. Furthermore, with the customized kernel, DiTFastAttnV2 achieves a 68% reduction in attention FLOPs and 1.5x end-to-end speedup on 2K image generation without compromising visual fidelity.

📄 PDF Abstract BibTeX arXiv:2503.22796

Code (0)

등록된 구현이 없습니다.

Tasks

2kImage GenerationText to Image GenerationText-to-Image Generation

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음
Diffusion Diffusion models generate samples by gradually removing noise from a signal, and their training objective can be expressed as a reweighted variational lower-bound…

Similar Papers 제목 키워드 기반

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance

2024-12-11 · Yingxin Li, Ye Li, Yuan Meng, Xinzhu Ma 외

As large language models (LLMs) continue to advance, the demand for higher quality and faster processing of long contexts across various applications is growing. KV cache is widely adopted as it stores previously generat…

Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models

2026-05-10 · Yicheng Ji, Zhizhou Zhong, Jun Zhang, Qin Yang 외 arxiv

Autoregressive (AR) video diffusion models adopt a streaming generation framework, enabling long-horizon video generation with real-time responsiveness, as exemplified by the Self Forcing training paradigm. However, exis…

Video Generation

LAVa: Layer-wise KV Cache Eviction with Dynamic Budget Allocation

2025-09-11 · Yiqun Shen, Song Yuan, Zhengze Zhang, Xiaoliang Wang 외 arxiv

KV Cache is commonly used to accelerate LLM inference with long contexts, yet its high memory demand drives the need for cache compression. Existing compression methods, however, are largely heuristic and lack dynamic bu…

Code Completion

HybridKV: Hybrid KV Cache Compression for Efficient Multimodal Large Language Model Inference

2026-04-07 · Bowen Zeng, Feiyang Ren, Jun Zhang, Xiaoling Gu 외 arxiv

Multimodal Large Language Models (MLLMs) have advanced unified reasoning over text, images, and videos, but their inference is hindered by the rapid growth of key-value (KV) caches. Each visual input expands into thousan…

itKD: Interchange Transfer-based Knowledge Distillation for 3D Object Detection

2022-05-31 · CVPR 2023 1 · Hyeon Cho, Junyong Choi, Geonwoo Baek, Wonjun Hwang

Point-cloud based 3D object detectors recently have achieved remarkable progress. However, most studies are limited to the development of network architectures for improving only their accuracy without consideration of t…

3D Object DetectionCloud DetectionComputational EfficiencyKnowledge Distillation+2