Pith. sign in

REVIEW 9 cited by

Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformers

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2505.14687 v1 pith:HWI7UAYC submitted 2025-05-20 cs.CV

Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformers

classification cs.CV
keywords attentiongrattransformerstokensdiffusionfirstgenerationimage
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
abstract

Diffusion-based Transformers have demonstrated impressive generative capabilities, but their high computational costs hinder practical deployment, for example, generating an $8192\times 8192$ image can take over an hour on an A100 GPU. In this work, we propose GRAT (\textbf{GR}ouping first, \textbf{AT}tending smartly), a training-free attention acceleration strategy for fast image and video generation without compromising output quality. The key insight is to exploit the inherent sparsity in learned attention maps (which tend to be locally focused) in pretrained Diffusion Transformers and leverage better GPU parallelism. Specifically, GRAT first partitions contiguous tokens into non-overlapping groups, aligning both with GPU execution patterns and the local attention structures learned in pretrained generative Transformers. It then accelerates attention by having all query tokens within the same group share a common set of attendable key and value tokens. These key and value tokens are further restricted to structured regions, such as surrounding blocks or criss-cross regions, significantly reducing computational overhead (e.g., attaining a \textbf{35.8$\times$} speedup over full attention when generating $8192\times 8192$ images) while preserving essential attention patterns and long-range context. We validate GRAT on pretrained Flux and HunyuanVideo for image and video generation, respectively. In both cases, GRAT achieves substantially faster inference without any fine-tuning, while maintaining the performance of full attention. We hope GRAT will inspire future research on accelerating Diffusion Transformers for scalable visual generation.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Attention Sinks in Diffusion Transformers: A Causal Analysis

    cs.CV 2026-05 unverdicted novelty 7.0

    Suppressing attention sinks in diffusion transformers does not degrade text-image alignment or most preference metrics, revealing a dissociation between generation trajectory changes and semantic output quality.

  2. Efficient Video Diffusion Models: Advancements and Challenges

    cs.CV 2026-04 unverdicted novelty 7.0

    A survey that groups efficient video diffusion methods into four paradigms—step distillation, efficient attention, model compression, and cache/trajectory optimization—and outlines open challenges for practical use.

  3. A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens

    cs.CV 2026-04 conditional novelty 7.0

    Delta tokens compress VFM feature differences into single tokens, enabling a lightweight generative world model that predicts diverse futures with far lower compute than existing approaches.

  4. DC-DiT: Adaptive Compute and Elastic Inference for Visual Generation via Dynamic Chunking

    cs.CV 2026-03 unverdicted novelty 7.0

    DC-DiT learns dynamic chunking to allocate fewer tokens to smooth or noisy regions and more to detailed or late-stage areas, cutting inference FLOPs up to 36.8% while improving FID up to 37.8% on class-conditional Ima...

  5. Attention Sinks in Diffusion Transformers: A Causal Analysis

    cs.CV 2026-05 unverdicted novelty 6.0

    Suppressing attention sinks in diffusion transformers does not degrade CLIP-T alignment at moderate levels but induces sink-specific perceptual shifts six times larger than equal-budget random masking.

  6. Attention Sinks in Diffusion Transformers: A Causal Analysis

    cs.CV 2026-05 unverdicted novelty 6.0

    Suppressing attention sinks in Stable Diffusion 3 does not degrade text-image alignment or preference metrics at mild intervention levels, though stronger suppression reveals sink-specific perceptual shifts larger tha...

  7. Frequency-Aware Flow Matching for High-Quality Image Generation

    cs.CV 2026-04 unverdicted novelty 6.0

    FreqFlow introduces frequency-aware conditioning and a two-branch architecture to flow matching, reaching FID 1.38 on ImageNet-256 and outperforming DiT and SiT.

  8. ACID: Adaptive Caching for vIDeo generation

    cs.CV 2026-07 conditional novelty 5.0

    ACID dynamically switches caching thresholds based on drift-signal rate of change, recovering near-conservative quality at substantially higher video-diffusion inference speeds.

  9. ACID: Adaptive Caching for vIDeo generation

    cs.CV 2026-07 conditional novelty 5.0

    Adaptive threshold switching on the drift signal's rate of change expands the quality-vs-speed Pareto frontier of TeaCache, EasyCache, and DiCache across HunyuanVideo, Wan 2.1, and CogVideoX.