Teacher-Feature Drifting: One-Step Diffusion Distillation with Pretrained Diffusion Representations

• URL: http://arxiv.org/abs/2605.07327v1
• Date: Fri, 08 May 2026 06:33:33 GMT
• Title: Teacher-Feature Drifting: One-Step Diffusion Distillation with Pretrained Diffusion Representations
• Abstract summary: We show that a single drifting loss can be directly used to simplify one step distillation.<n>Unlike the original Drifting Model, which relies on an additional pretrained feature extractor, we use intermediate hidden states of the pretrained teacher model as the feature representation.<n>Our method achieves efficient one step generation with competitive image quality and diversity, achieving FID scores of 1.58 on ImageNet-64$times$64 and 18.4 on SDXL.
• Score: 48.0169045294604
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Sampling from pretrained diffusion and flow-matching models typically requires many forward passes to generate diverse and high-fidelity images. Existing distillation methods often rely on multiple auxiliary networks, carefully designed training stages, or complex optimization pipelines. In this work, we revisit the recently proposed Drifting Model objective and show that a single drifting loss can be directly used to simplify one step distillation. A key observation is that the pretrained diffusion teacher itself already provides a strong representation space. Unlike the original Drifting Model, which relies on an additional pretrained feature extractor, we use intermediate hidden states of the pretrained teacher model as the feature representation. This removes the need for training or introducing an extra representation network while preserving a semantically meaningful feature geometry for drifting. Furthermore, we introduce a lightweight mode coverage loss to mitigate mode collapse during distillation and encourage the student generator to cover diverse teacher-supported regions. Extensive experiments on ImageNet and SDXL demonstrate that our method achieves efficient one step generation with competitive image quality and diversity, achieving FID scores of 1.58 on ImageNet-64$\times$64 and 18.4 on SDXL, while substantially simplifying the overall distillation framework.

Related papers

• Diffusion As Self-Distillation: End-to-End Latent Diffusion In One Model [53.77953728335891]
  Latent Diffusion Models rely on a complex, three-part architecture consisting of a separate encoder, decoder, and diffusion network.<n>We propose Diffusion as Self-Distillation (DSD), a new framework with key modifications to the training objective that stabilize the latent space.<n>This approach enables, for the first time, the stable end-to-end training of a single network that simultaneously learns to encode, decode, and perform diffusion.
  arXiv  Detail & Related papers  (2025-11-18T17:58:16Z)
• Revisiting Diffusion Models: From Generative Pre-training to One-Step Generation [2.3359837623080613]
  We show that diffusion training may be viewed as a form of generative pre-training.<n>We create a one-step generation model by fine-tuning a pre-trained model with 85% of parameters frozen.
  arXiv  Detail & Related papers  (2025-06-11T03:55:26Z)
• Towards Training One-Step Diffusion Models Without Distillation [72.80423908458772]
  We introduce a family of new training methods that entirely forgo teacher score supervision.<n>We find that initializing the student model with the teacher's weights remains critical.
  arXiv  Detail & Related papers  (2025-02-11T23:02:14Z)
• Relational Diffusion Distillation for Efficient Image Generation [27.127061578093674]
  Diffusion model's high delay hinders its wide application in edge devices with scarce computing resources.<n>We propose Diffusion Distillation (RDD), a novel distillation method tailored specifically for distilling diffusion models.<n>Our proposed RDD leads to 1.47 FID decrease under 1 sampling step compared to state-of-the-art diffusion distillation methods and achieving 256x speed-up.
  arXiv  Detail & Related papers  (2024-10-10T07:40:51Z)
• Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think [72.48325960659822]
  One main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations.<n>We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders.<n>The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs.
  arXiv  Detail & Related papers  (2024-10-09T14:34:53Z)
• Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillation [19.88187051373436]
  We propose Distribution Backtracking Distillation (DisBack) to speed up the sampling speed of diffusion models.<n>DisBack achieves faster and better convergence than the existing distillation method, with FID score of 1.38 on ImageNet 64x64 dataset.
  arXiv  Detail & Related papers  (2024-08-28T17:58:17Z)
• Plug-and-Play Diffusion Distillation [14.359953671470242]
  We propose a new distillation approach for guided diffusion models. An external lightweight guide model is trained while the original text-to-image model remains frozen. We show that our method reduces the inference of classifier-free guided latent-space diffusion models by almost half.
  arXiv  Detail & Related papers  (2024-06-04T04:22:47Z)
• Improved Distribution Matching Distillation for Fast Image Synthesis [54.72356560597428]
  We introduce DMD2, a set of techniques that lift this limitation and improve DMD training. First, we eliminate the regression loss and the need for expensive dataset construction. Second, we integrate a GAN loss into the distillation procedure, discriminating between generated samples and real images.
  arXiv  Detail & Related papers  (2024-05-23T17:59:49Z)
• BOOT: Data-free Distillation of Denoising Diffusion Models with Bootstrapping [64.54271680071373]
  Diffusion models have demonstrated excellent potential for generating diverse images. Knowledge distillation has been recently proposed as a remedy that can reduce the number of inference steps to one or a few. We present a novel technique called BOOT, that overcomes limitations with an efficient data-free distillation algorithm.
  arXiv  Detail & Related papers  (2023-06-08T20:30:55Z)

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.