Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step

• URL: http://arxiv.org/abs/2410.14919v4
• Date: Tue, 24 Dec 2024 05:06:20 GMT
• Title: Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step
• Authors: Mingyuan Zhou, Huangjie Zheng, Yi Gu, Zhendong Wang, Hai Huang,
• Abstract summary: We introduce SiDA (SiD with Adversarial Loss), which improves generation quality and distillation efficiency.<n>SiDA incorporates real images and adversarial loss, allowing it to distinguish between real images and those generated by SiD.<n>SiDA converges significantly faster than its predecessor when distilled from scratch.
• Score: 64.53013367995325
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Score identity Distillation (SiD) is a data-free method that has achieved SOTA performance in image generation by leveraging only a pretrained diffusion model, without requiring any training data. However, its ultimate performance is constrained by how accurate the pretrained model captures the true data scores at different stages of the diffusion process. In this paper, we introduce SiDA (SiD with Adversarial Loss), which not only enhances generation quality but also improves distillation efficiency by incorporating real images and adversarial loss. SiDA utilizes the encoder from the generator's score network as a discriminator, allowing it to distinguish between real images and those generated by SiD. The adversarial loss is batch-normalized within each GPU and then combined with the original SiD loss. This integration effectively incorporates the average "fakeness" per GPU batch into the pixel-based SiD loss, enabling SiDA to distill a single-step generator. SiDA converges significantly faster than its predecessor when distilled from scratch, and swiftly improves upon the original model's performance during fine-tuning from a pre-distilled SiD generator. This one-step adversarial distillation method establishes new benchmarks in generation performance when distilling EDM diffusion models, achieving FID scores of 1.110 on ImageNet 64x64. When distilling EDM2 models trained on ImageNet 512x512, our SiDA method surpasses even the largest teacher model, EDM2-XXL, which achieved an FID of 1.81 using classifier-free guidance (CFG) and 63 generation steps. In contrast, SiDA achieves FID scores of 2.156 for size XS, 1.669 for S, 1.488 for M, 1.413 for L, 1.379 for XL, and 1.366 for XXL, all without CFG and in a single generation step. These results highlight substantial improvements across all model sizes. Our code is available at https://github.com/mingyuanzhou/SiD/tree/sida.

Related papers

• Autoregressive Distillation of Diffusion Transformers [18.19070958829772]
  We propose AutoRegressive Distillation (ARD), a novel approach that leverages the historical trajectory of the ODE to predict future steps. ARD offers two key benefits: 1) it mitigates exposure bias by utilizing a predicted historical trajectory that is less susceptible to accumulated errors, and 2) it leverages the previous history of the ODE trajectory as a more effective source of coarse-grained information. Our model achieves a $5times$ reduction in FID degradation compared to the baseline methods while requiring only 1.1% extra FLOPs on ImageNet-256.
  arXiv  Detail & Related papers  (2025-04-15T15:33:49Z)
• One-Step Diffusion Distillation through Score Implicit Matching [74.91234358410281]
  We present Score Implicit Matching (SIM) a new approach to distilling pre-trained diffusion models into single-step generator models. SIM shows strong empirical performances for one-step generators. By applying SIM to a leading transformer-based diffusion model, we distill a single-step generator for text-to-image generation.
  arXiv  Detail & Related papers  (2024-10-22T08:17:20Z)
• Generative Dataset Distillation Based on Diffusion Model [45.305885410046116]
  We propose a novel generative dataset distillation method based on Stable Diffusion. Specifically, we use the SDXL-Turbo model which can generate images at high speed and quality. We achieved third place in the generative track of the ECCV 2024 DD Challenge.
  arXiv  Detail & Related papers  (2024-08-16T08:52:02Z)
• Long and Short Guidance in Score identity Distillation for One-Step Text-to-Image Generation [62.30570286073223]
  Diffusion-based text-to-image generation models have shown the capacity to generate images consistent with textual descriptions. In this paper, we enhance Score identity Distillation (SiD) by developing long and short classifier-free guidance (LSG) to efficiently distill pretrained Stable Diffusion models. SiD equipped with LSG rapidly improves FID and CLIP scores, achieving state-of-the-art FID performance while maintaining a competitive CLIP score.
  arXiv  Detail & Related papers  (2024-06-03T17:44:11Z)
• Diffusion Models Are Innate One-Step Generators [2.3359837623080613]
  Diffusion Models (DMs) can generate remarkable high-quality results. DMs' layers are differentially activated at different time steps, leading to an inherent capability to generate images in a single step. Our method achieves the SOTA results on CIFAR-10, AFHQv2 64x64 (FID 1.23), FFHQ 64x64 (FID 0.85) and ImageNet 64x64 (FID 1.16) with great efficiency.
  arXiv  Detail & Related papers  (2024-05-31T11:14:12Z)
• 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)
• Diffusion Time-step Curriculum for One Image to 3D Generation [91.07638345953016]
  Score distillation sampling(SDS) has been widely adopted to overcome the absence of unseen views in reconstructing 3D objects from a textbfsingle image. We find out the crux is the overlooked indiscriminate treatment of diffusion time-steps during optimization. We propose the Diffusion Time-step Curriculum one-image-to-3D pipeline (DTC123), which involves both the teacher and student models collaborating with the time-step curriculum in a coarse-to-fine manner.
  arXiv  Detail & Related papers  (2024-04-06T09:03:18Z)
• Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generation [61.03530321578825]
  We introduce Score identity Distillation (SiD), an innovative data-free method that distills the generative capabilities of pretrained diffusion models into a single-step generator. SiD not only facilitates an exponentially fast reduction in Fr'echet inception distance (FID) during distillation but also approaches or even exceeds the FID performance of the original teacher diffusion models.
  arXiv  Detail & Related papers  (2024-04-05T12:30:19Z)
• SDXS: Real-Time One-Step Latent Diffusion Models with Image Conditions [5.100085108873068]
  We present two models, SDXS-512 and SDXS-1024, achieving inference speeds of approximately 100 FPS (30x faster than SD v1.5) and 30 FPS (60x faster than SDXL) on a single GPU. Our training approach offers promising applications in image-conditioned control, facilitating efficient image-to-image translation.
  arXiv  Detail & Related papers  (2024-03-25T11:16:23Z)
• ACT-Diffusion: Efficient Adversarial Consistency Training for One-step Diffusion Models [59.90959789767886]
  We show that optimizing consistency training loss minimizes the Wasserstein distance between target and generated distributions. By incorporating a discriminator into the consistency training framework, our method achieves improved FID scores on CIFAR10 and ImageNet 64$times$64 and LSUN Cat 256$times$256 datasets.
  arXiv  Detail & Related papers  (2023-11-23T16:49:06Z)

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.