- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes Error-Regulated Adaptive Caching (ERAC), a framework designed to accelerate Diffusion Transformers through adaptive feature caching. The key idea is to dynamically determine when to perform a full model computation based on estimated caching errors, rather than relying on fixed caching schedules. Specifically, the authors introduce a lightweight, time-aware autoregressive error estimator trained on calibration trajectories generated using fixed-schedule caching. The estimator predicts caching errors from latent drift, temporal information, and previous error estimates. When the predicted error exceeds a predefined threshold, a full DiT computation is triggered to refresh the cached reference.
The method is evaluated on FLUX.1-dev for image generation and Wan2.1 1.3B for video generation. Experimental results demonstrate competitive quality-efficiency trade-offs compared with existing acceleration methods, particularly at higher FLOPs-based speedup ratios, using Image Reward and GenEval for images and PSNR, LPIPS, and SSIM for videos.
Critical strengths
Strengths
- The paper is well-written, with clear and fluent explanations, making it easy to follow.
- The generated images demonstrate promising visual quality.
Critical weaknesses
Weaknesses
- The paper lacks sufficient experimental completeness, particularly ablation studies and other necessary analyses that could be included in the appendix (e.g., the analyses mentioned in Weakness 2).
- Although the method employs a lightweight error predictor, the paper does not provide sufficient details about the number and types of trajectories used for training. This makes it difficult to assess the computational cost and the method's sensitivity to the amount of calibration data.
- Around Line 148, the authors argue that the proxy signals adopted by previous methods are merely correlated with the true error rather than directly representing it. I agree with this observation. However, I believe the proposed method also essentially relies on proxy signals. More importantly, as mentioned around Line 253, the authors train the error predictor using calibration trajectories generated under fixed-schedule caching, whereas inference dynamically determines whether to perform a full DiT computation based on the predicted error. Consequently, the error estimates may not accurately reflect the actual inference conditions, potentially leading to a distribution shift between training and inference. It remains unclear whether the predictor can reliably estimate the true error under such circumstances.
- Around Line 275, the authors state that the MLP "shares the same weights across all time steps", but the paper does not sufficiently justify this design choice. Different denoising stages may exhibit different error evolution patterns.
- The experiments lack latency measurements. Additionally, VBench scores should be reported for video generation, while additional metrics such as PSNR are recommended for image generation.
Errors
- The categorization in Section 2.2, Training-Free Diffusion Acceleration, appears to be incorrect. Δ-DiT should be categorized as a feature caching method rather than a sampling-step reduction method.
- Around Line 366, the reported PSNR difference of 0.91 dB does not appear to correspond to any of the baseline comparisons in the experimental results.
Questions for authors
Questions
My main concerns have been detailed in the Weaknesses section. I would appreciate further clarification and explanations from the authors regarding these issues.