# Accelerating Diffusion Transformer via Gradient-Optimized Cache

arXiv: https://arxiv.org/abs/2503.05156 · HTML: https://mlsanity.com/paper/2503.05156/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 17 of its 44 references that are in the map.
- First posted on arXiv 7 Mar 2025.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 17 of 17)

- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Consistency Models](https://mlsanity.com/paper/2303.01469/) (2023)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis](https://mlsanity.com/paper/2310.00426/) (2023)
- [BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction](https://mlsanity.com/paper/2102.05426/) (2021)
- [Delta-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers](https://mlsanity.com/paper/2406.01125/) (2024)
- [Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching](https://mlsanity.com/paper/2406.01733/) (2024)
- [FORA: Fast-Forward Caching in Diffusion Transformer Acceleration](https://mlsanity.com/paper/2407.01425/) (2024)
- [A Signal Propagation Perspective for Pruning Neural Networks at Initialization](https://mlsanity.com/paper/1906.06307/) (2019)
- [PTQ4DiT: Post-training Quantization for Diffusion Transformers](https://mlsanity.com/paper/2405.16005/) (2024)
- [Progressive Distillation for Fast Sampling of Diffusion Models](https://mlsanity.com/paper/2202.00512/) (2022)
- [Flash Diffusion: Accelerating Any Conditional Diffusion Model for Few Steps Image Generation](https://mlsanity.com/paper/2406.02347/) (2024)
- [Token Pruning for Caching Better: 9 Times Acceleration on Stable Diffusion for Free](https://mlsanity.com/paper/2501.00375/) (2024)
- [SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers](https://mlsanity.com/paper/2411.10510/) (2024)
- [Accelerating Diffusion Transformer via Error-Optimized Cache](https://mlsanity.com/paper/2501.19243/) (2025)
- [DAMA: Data- and Model-aware Alignment of Multi-modal LLMs](https://mlsanity.com/paper/2502.01943/) (2025)

## Built on by (most cited first, 15 of 15)

- [Enhancing CLIP Robustness via Cross-Modality Alignment](https://mlsanity.com/paper/2510.24038/) (2025)
- [Astraea: A Token-wise Acceleration Framework for Video Diffusion Transformers](https://mlsanity.com/paper/2506.05096/) (2025)
- [Mixture of Distributions Matters: Dynamic Sparse Attention for Efficient Video Diffusion Transformers](https://mlsanity.com/paper/2601.11641/) (2026)
- [Denoising as Path Planning: Training-Free Acceleration of Diffusion Models with DPCache](https://mlsanity.com/paper/2602.22654/) (2026)
- [Relational Feature Caching for Accelerating Diffusion Transformers](https://mlsanity.com/paper/2602.19506/) (2026)
- [AccelAes: Accelerating Diffusion Transformers for Training-Free Aesthetic-Enhanced Image Generation](https://mlsanity.com/paper/2603.12575/) (2026)
- [Thinking with Frames: Generative Video Distortion Evaluation via Frame Reward Model](https://mlsanity.com/paper/2601.04033/) (2026)
- [Accelerating Diffusion-based Video Editing via Heterogeneous Caching: Beyond Full Computing at Sampled Denoising Timestep](https://mlsanity.com/paper/2603.24260/) (2026)
- [Plug-and-Play Fidelity Optimization for Diffusion Transformer Acceleration via Cumulative Error Minimization](https://mlsanity.com/paper/2512.23258/) (2025)
- [CUFG: Curriculum Unlearning Guided by the Forgetting Gradient](https://mlsanity.com/paper/2509.14633/) (2025)
- [H2-Cache: A Novel Hierarchical Dual-Stage Cache for High-Performance Acceleration of Generative Diffusion Models](https://mlsanity.com/paper/2510.27171/) (2025)
- [BeCARE: Budgeted Cache Refresh for Diffusion Transformer Acceleration](https://mlsanity.com/paper/2605.27075/) (2026)
- [SODA: Sensitivity-Oriented Dynamic Acceleration for Diffusion Transformer](https://mlsanity.com/paper/2603.07057/) (2026)
- [Trajectory-Consistent Calibration for Cache-Accelerated Diffusion Models](https://mlsanity.com/paper/2605.24870/) (2026)
- [DriveCache: Action-Aware Caching for Driving World Model Inference](https://mlsanity.com/paper/2608.16354/) (2026)

## Abstract (arXiv)

Feature caching has emerged as an effective strategy to accelerate diffusion transformer (DiT) sampling through temporal feature reuse. It is a challenging problem since (1) Progressive error accumulation from cached blocks significantly degrades generation quality, particularly when over 50% of blocks are cached; (2) Current error compensation approaches neglect dynamic perturbation patterns during the caching process, leading to suboptimal error correction. To solve these problems, we propose the Gradient-Optimized Cache (GOC) with two key innovations: (1) Cached Gradient Propagation: A gradient queue dynamically computes the gradient differences between cached and recomputed features. These gradients are weighted and propagated to subsequent steps, directly compensating for the approximation errors introduced by caching. (2) Inflection-Aware Optimization: Through statistical analysis of feature variation patterns, we identify critical inflection points where the denoising trajectory changes direction. By aligning gradient updates with these detected phases, we prevent conflicting gradient directions during error correction. Extensive evaluations on ImageNet demonstrate GOC's superior trade-off between efficiency and quality. With 50% cached blocks, GOC achieves IS 216.28 (26.3% higher) and FID 3.907 (43% lower) compared to baseline DiT, while maintaining identical computational costs. These improvements persist across various cache ratios, demonstrating robust adaptability to different acceleration requirements. Code is available at https://github.com/qiujx0520/GOCICCV2025.git.