Memory-Efficient Training-Free Acceleration of Diffusion Transformers with BaryCache

• URL: http://arxiv.org/abs/2608.28670v1
• Date: Mon, 24 Aug 2026 16:11:56 GMT
• Title: Memory-Efficient Training-Free Acceleration of Diffusion Transformers with BaryCache
• Abstract summary: We propose a training-free acceleration method that performs stepwise forecasting for DiT sampling using a Barycentric Extrapolator.<n>Our approach provides a favorable trade-off between memory usage and perceptual quality, while delivering up to 3.30x end-to-end sampling speedup.
• Score: 23.178001584385544
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Diffusion Transformers achieve high-fidelity image and video generation, but their iterative sampling remains expensive, for each denoising step requires large matrix operations. Existing cache-based acceleration reduces redundant computation yet increases the VRAM footprint by storing intermediate states, which can directly constrain inference batch size. In this work, we propose a training-free acceleration method that performs stepwise forecasting for DiT sampling using a Barycentric Extrapolator. By leveraging barycentric extrapolation, our predictor is numerically stable and alleviates oscillatory artifacts analogous to the Runge phenomenon during forward forecasting. Across extensive experiments on both image and video generation, our approach provides a favorable trade-off between memory usage and perceptual quality, while delivering up to 3.30x end-to-end sampling speedup compared with baseline DiT inference.

Related papers

• Test-time Sparsity for Extreme Fast Action Diffusion [15.679362579177267]
  Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature.<n>We propose test-time sparsity to tackle this challenge, which aims to accelerate action diffusion by dynamically predicting prunable residual computations for each model forward at test time.
  arXiv  Detail & Related papers  (2026-05-13T10:28:50Z)
• 6Bit-Diffusion: Inference-Time Mixed-Precision Quantization for Video Diffusion Models [32.45415555978057]
  Diffusion transformers have demonstrated remarkable capabilities in generating videos.<n>However, their practical deployment is constrained by high memory usage and computational cost.<n>Post-Training Quantization provides a practical way to reduce memory usage and boost speed.
  arXiv  Detail & Related papers  (2026-03-19T10:45:11Z)
• Adaptive Spectral Feature Forecasting for Diffusion Sampling Acceleration [58.19554276924402]
  We propose spectral diffusion feature forecaster (Spectrum) to enable global, long-range feature reuse with tightly controlled error.<n>We achieve up to 4.79$times$ speedup on FLUX.1 and 4.67$times$ speedup on Wan2.1-14B, while maintaining much higher sample quality compared with the baselines.
  arXiv  Detail & Related papers  (2026-03-02T08:59:11Z)
• Predict to Skip: Linear Multistep Feature Forecasting for Efficient Diffusion Transformers [10.751183015853863]
  Diffusion Transformers (DiT) have emerged as a widely adopted backbone for high-fidelity image and video generation.<n>We propose textbfPrediT, a training-free acceleration framework that formulates feature prediction as a linear multistep problem.<n>Our method achieves up to $5.54times$ latency reduction across various DiT-based image and video generation models, while incurring negligible quality degradation.
  arXiv  Detail & Related papers  (2026-02-20T09:33:59Z)
• AdaCorrection: Adaptive Offset Cache Correction for Accurate Diffusion Transformers [37.38708392928324]
  Transformer Diffusion (TDis) achieve state-of-the-art in high-fidelity and image generation but suffer from expensive inference due to their iterative denoising.<n>AdaCorrection is an adaptive offset cache correction framework that maintains high generation fidelity while enabling efficient reuse across cache layers during diffusion inference.<n>Our approach achieves strong generation quality with minimal computational overhead, maintaining near-original FID while providing moderate acceleration.
  arXiv  Detail & Related papers  (2026-02-13T08:11:54Z)
• RAPID^3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformer [86.57077884971478]
  Diffusion Transformers (DiTs) excel at visual generation yet remain hampered by slow sampling.<n>We introduce RAPID3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformers.<n>It delivers image-wise acceleration with zero updates to the base generator.<n>It achieves nearly 3x faster sampling with competitive generation quality.
  arXiv  Detail & Related papers  (2025-09-26T13:20:52Z)
• Sequential Posterior Sampling with Diffusion Models [15.028061496012924]
  We propose a novel approach that models the transition dynamics to improve the efficiency of sequential diffusion posterior sampling in conditional image synthesis. We demonstrate the effectiveness of our approach on a real-world dataset of high frame rate cardiac ultrasound images. Our method opens up new possibilities for real-time applications of diffusion models in imaging and other domains requiring real-time inference.
  arXiv  Detail & Related papers  (2024-09-09T07:55:59Z)
• Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching [56.286064975443026]
  We make an interesting and somehow surprising observation: the computation of a large proportion of layers in the diffusion transformer, through a caching mechanism, can be readily removed even without updating the model parameters. We introduce a novel scheme, named Learningto-Cache (L2C), that learns to conduct caching in a dynamic manner for diffusion transformers. Experimental results show that L2C largely outperforms samplers such as DDIM and DPM-r, alongside prior cache-based methods at the same inference speed.
  arXiv  Detail & Related papers  (2024-06-03T18:49:57Z)
• Towards More Accurate Diffusion Model Acceleration with A Timestep Tuner [112.99126045081046]
  A diffusion model, which is formulated to produce an image using thousands of denoising steps, usually suffers from a slow inference speed.<n>We propose a textbftimestep tuner that helps find a more accurate integral direction for a particular interval at the minimum cost.<n>Experiments show that our plug-in design can be trained efficiently and boost the inference performance of various state-of-the-art acceleration methods.
  arXiv  Detail & Related papers  (2023-10-14T02:19:07Z)
• DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models [58.450152413700586]
  We introduce a soft absorbing state that facilitates the diffusion model in learning to reconstruct discrete mutations based on the underlying Gaussian space. We employ state-of-the-art ODE solvers within the continuous space to expedite the sampling process. Our proposed method effectively accelerates the training convergence by 4x and generates samples of similar quality 800x faster.
  arXiv  Detail & Related papers  (2023-10-09T15:29:10Z)
• ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting [70.83632337581034]
  Diffusion-based image super-resolution (SR) methods are mainly limited by the low inference speed. We propose a novel and efficient diffusion model for SR that significantly reduces the number of diffusion steps. Our method constructs a Markov chain that transfers between the high-resolution image and the low-resolution image by shifting the residual.
  arXiv  Detail & Related papers  (2023-07-23T15:10:02Z)
• Parallel Sampling of Diffusion Models [76.3124029406809]
  Diffusion models are powerful generative models but suffer from slow sampling. We present ParaDiGMS, a novel method to accelerate the sampling of pretrained diffusion models by denoising multiple steps in parallel.
  arXiv  Detail & Related papers  (2023-05-25T17:59:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.