One-Step Diffusion with Distribution Matching Distillation
Open accessThe DOI matches its Crossref record; 54 of the 54 references it cites were found in a registry or an index.
Abstract
Diffusion models generate high-quality images but require dozens of forward passes. We introduce Distribution Matching Distillation (DMD), a procedure to transform a diffusion model into a one-step image generator with minimal impact on image quality. We enforce the one-step image generator match the diffusion model at distribution level, by minimizing an approximate KL divergence whose gradient can be expressed as the difference between 2 score functions, one of the target distribution and the other of the synthetic distribution being produced by our one-step generator. The score functions are parameterized as two diffusion models trained separately on each distribution. Combined with a simple regression loss matching the large-scale structure of the multi-step diffusion outputs, our method outperforms all published few-step diffusion approaches, reaching 2.62 FID on ImageNet 64×64 and 11.49 FID on zero-shot COCO-30k, comparable to Stable Diffusion but orders of magnitude faster. Utilizing FP16 inference, our model can generate images at 20 FPS on modern hardware.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
Microsoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460
The unreasonable effectiveness of deep features as a perceptual metric
Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018) · arXiv · doi:10.48550/arxiv.1801.03924 · Cited by 24,812
Self-citationUnpaired image-to-image translation using cycle-consistent adversarial networks
Zhu, J.-Y., Park, T., Isola, P., & Efros, A. A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.244 · Cited by 23,602
Self-citationImage-to-Image Translation with Conditional Adversarial Networks
Isola, P., Zhu, J.-Y., Zhou, T., & Efros, A. A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.632 · Cited by 22,930
High-Resolution Image Synthesis with Latent Diffusion Models
Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01042 · Cited by 20,302
Auto-Encoding Variational Bayes
Kingma, D. P., & Welling, M. (2013) · arXiv · doi:10.48550/arxiv.1312.6114 · Cited by 20,067
Learning Transferable Visual Models From Natural Language Supervision
Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.00020 · Cited by 15,293
The Unreasonable Effectiveness of Deep Features as a Perceptual Metric
Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00068 · Cited by 14,623
Self-citationDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204
Decoupled Weight Decay Regularization
Loshchilov, I., & Hutter, F. (2017) · arXiv · doi:10.48550/arxiv.1711.05101 · Cited by 12,522
A Style-Based Generator Architecture for Generative Adversarial Networks
Karras, T., Laine, S., & Aila, T. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00453 · Cited by 11,714 · cited 1×
“Among them, GANs have shown unprecedented quality in realism [4, 26– 28, 30, 67], particularly when the GAN loss can be combined with task-specific, auxiliary regression losses to mitigate training instability, ranging from paired image translation [24, 54, 79, 89] to unpaired image editing [37, 55,…”
BackgroundDenoising Diffusion Probabilistic Models
Ho, J., Jain, A., & Abbeel, P. (2020) · arXiv · doi:10.48550/arxiv.2006.11239 · Cited by 8,712 · cited 7×
“In contrast to GANs [15] and VAEs [34], however, their sampling is a slow, iterative process that transforms a Gaussian noise sample into an intricate image by progressive denoising [21, 74].”
BackgroundMethodInfluentialGANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., & Hochreiter, S. (2017) · arXiv · doi:10.48550/arxiv.1706.08500 · Cited by 8,365
Analyzing and Improving the Image Quality of StyleGAN
Karras, T., Laine, S., Aittala, M., Hellsten, J., Lehtinen, J., & Aila, T. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00813 · Cited by 6,332
Analyzing and Improving the Image Quality of StyleGAN
Karras, T., Laine, S., Aittala, M., Hellsten, J., Lehtinen, J., & Aila, T. (2019) · arXiv · doi:10.48550/arxiv.1912.04958 · Cited by 6,332
High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs
Wang, T.-C., Liu, M.-Y., Zhu, J.-Y., Tao, A., Kautz, J., & Catanzaro, B. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00917 · Cited by 5,360
Progressive Growing of GANs for Improved Quality, Stability, and Variation
Karras, T., Aila, T., Laine, S., & Lehtinen, J. (2017) · arXiv · doi:10.48550/arxiv.1710.10196 · Cited by 4,727
Large Scale GAN Training for High Fidelity Natural Image Synthesis
Brock, A., Donahue, J., & Simonyan, K. (2018) · arXiv · doi:10.48550/arxiv.1809.11096 · Cited by 3,717 · cited 2×
“Among them, GANs have shown unprecedented quality in realism [4, 26– 28, 30, 67], particularly when the GAN loss can be combined with task-specific, auxiliary regression losses to mitigate training instability, ranging from paired image translation [24, 54, 79, 89] to unpaired image editing [37, 55,…”
BackgroundMethodDiffusion Models Beat GANs on Image Synthesis
Dhariwal, P., & Nichol, A. (2021) · arXiv · doi:10.48550/arxiv.2105.05233 · Cited by 3,362
Semantic Image Synthesis With Spatially-Adaptive Normalization
Park, T., Liu, M.-Y., Wang, T.-C., & Zhu, J.-Y. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00244 · Cited by 2,974 · cited 1×
“…them, GANs have shown unprecedented quality in realism [4, 26– 28, 30, 67], particularly when the GAN loss can be combined with task-specific, auxiliary regression losses to mitigate training instability, ranging from paired image translation [24, 54, 79, 89] to unpaired image editing [37, 55, 94].”
BackgroundSelf-citationDeep Unsupervised Learning using Nonequilibrium Thermodynamics
Sohl-Dickstein, J., Weiss, E. A., Maheswaranathan, N., & Ganguli, S. (2015) · arXiv · doi:10.48550/arxiv.1503.03585 · Cited by 2,637 · cited 3×
“Diffusion models [21, 61, 63, 64, 71, 74] have revolutionized image generation, achieving unprecedented levels of realism and diversity with a stable training procedure.”
BackgroundMethodPhotorealistic Text-to-Image Diffusion Models with Deep Language Understanding
Saharia, C., Chan, W., Saxena, S., Li, L., Whang, J., Denton, E., Ghasemipour, S. K. S., Ayan, B. K., Mahdavi, S. S., Lopes, R. G., et al. (2022) · arXiv · doi:10.48550/arxiv.2205.11487 · Cited by 2,528 · cited 3×
“Diffusion models [21, 61, 63, 64, 71, 74] have revolutionized image generation, achieving unprecedented levels of realism and diversity with a stable training procedure.”
BackgroundHierarchical Text-Conditional Image Generation with CLIP Latents
Ramesh, A., Dhariwal, P., Nichol, A., Chu, C., & Chen, M. (2022) · arXiv · doi:10.48550/arxiv.2204.06125 · Cited by 2,511 · cited 3×
“Diffusion models [21, 61, 63, 64, 71, 74] have revolutionized image generation, achieving unprecedented levels of realism and diversity with a stable training procedure.”
BackgroundDenoising Diffusion Implicit Models
Song, J., Meng, C., & Ermon, S. (2020) · arXiv · doi:10.48550/arxiv.2010.02502 · Cited by 2,507 · cited 3×
“These models operate by progressively transforming noise into coherent structures through a reverse diffusion process [72, 74].”
BackgroundMethodGenerative Adversarial Text to Image Synthesis
Reed, S., Akata, Z., Yan, X., Logeswaran, L., Schiele, B., & Lee, H. (2016) · arXiv · doi:10.48550/arxiv.1605.05396 · Cited by 2,494
Zero-shot text-to-image generation
Ramesh, A., Pavlov, M., Goh, G., Gray, S., Voss, C., Radford, A., Chen, M., & Sutskever, I. (2021) · arXiv · doi:10.48550/arxiv.2102.12092 · Cited by 2,417
Score-Based Generative Modeling through Stochastic Differential Equations
Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S., & Poole, B. (2020) · arXiv · doi:10.48550/arxiv.2011.13456 · Cited by 1,953 · cited 8×
“In contrast to GANs [15] and VAEs [34], however, their sampling is a slow, iterative process that transforms a Gaussian noise sample into an intricate image by progressive denoising [21, 74].”
BackgroundMethodInfluentialGenerative Modeling by Estimating Gradients of the Data Distribution
Song, Y., & Ermon, S. (2019) · arXiv · doi:10.48550/arxiv.1907.05600 · Cited by 1,453 · cited 2×
“Since diffusion models are known to approximate the score functions on diffused distributions [23, 73], we can interpret the denoised diffusion outputs as gradient directions for making an image “more realistic”, or if the diffusion model is learned on the fake images, “more fake”.”
BackgroundGLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models
Nichol, A., Dhariwal, P., Ramesh, A., Shyam, P., Mishkin, P., McGrew, B., Sutskever, I., & Chen, M. (2021) · arXiv · doi:10.48550/arxiv.2112.10741 · Cited by 1,357
Training Generative Adversarial Networks with Limited Data
Karras, T., Aittala, M., Hellsten, J., Laine, S., Lehtinen, J., & Aila, T. (2020) · arXiv · doi:10.48550/arxiv.2006.06676 · Cited by 1,170
Classifier-Free Diffusion Guidance
Ho, J., & Salimans, T. (2022) · arXiv · doi:10.48550/arxiv.2207.12598 · Cited by 1,126 · cited 1×
“Classifier-Free Guidance [20] is widely used to improve the image quality of text-to-image diffusion models.”
MethodLAION-5B: An open large-scale dataset for training next generation image-text models
Schuhmann, C., Beaumont, R., Vencu, R., Gordon, C., Wightman, R., Cherti, M., Coombes, T., Katta, A., Mullis, C., Wortsman, M., et al. (2022) · arXiv · doi:10.48550/arxiv.2210.08402 · Cited by 1,022
A Connection Between Score Matching and Denoising Autoencoders
Vincent, P. (2011) · Neural Computation · doi:10.1162/neco_a_00142 · Cited by 1,011 · cited 1×
“…model from the pretrained diffusion model µ base , updating parameters � during training, by minimizing a standard denoising objective [21, 77]: where L denoise is weighted according to the diffusion timestep t , using the same weighting strategy employed during the training of the base…”
Method(2005) · doi:10.5555/1046920.1088696 · Cited by 756
High-Resolution Image Synthesis with Latent Diffusion Models
Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2021) · arXiv · doi:10.48550/arxiv.2112.10752 · Cited by 737
Elucidating the Design Space of Diffusion-Based Generative Models
Karras, T., Aittala, M., Aila, T., & Laine, S. (2022) · arXiv · doi:10.48550/arxiv.2206.00364 · Cited by 702
DiffWave: A Versatile Diffusion Model for Audio Synthesis
Kong, Z., Ping, W., Huang, J., Zhao, K., & Catanzaro, B. (2020) · arXiv · doi:10.48550/arxiv.2009.09761 · Cited by 569
Training Deep Nets with Sublinear Memory Cost
Chen, T., Xu, B., Zhang, C., & Guestrin, C. (2016) · arXiv · doi:10.48550/arxiv.1604.06174 · Cited by 561
SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis
Podell, D., English, Z., Lacey, K., Blattmann, A., Dockhorn, T., Müller, J., Penna, J., & Rombach, R. (2023) · arXiv · doi:10.48550/arxiv.2307.01952 · Cited by 455
Structure and Content-Guided Video Synthesis with Diffusion Models
Esser, P., Chiu, J., Atighehchian, P., Granskog, J., & Germanidis, A. (2023) · arXiv · doi:10.48550/arxiv.2302.03011 · Cited by 399
Imagen Video: High Definition Video Generation with Diffusion Models
Ho, J., Chan, W., Saharia, C., Whang, J., Gao, R., Gritsenko, A., Kingma, D. P., Poole, B., Norouzi, M., Fleet, D. J., et al. (2022) · arXiv · doi:10.48550/arxiv.2210.02303 · Cited by 399
Make-A-Video: Text-to-Video Generation without Text-Video Data
Singer, U., Polyak, A., Hayes, T., Yin, X., An, J., Zhang, S., Hu, Q., Yang, H., Ashual, O., Gafni, O., et al. (2022) · arXiv · doi:10.48550/arxiv.2209.14792 · Cited by 395
Kingma, D. P., Salimans, T., Poole, B., & Ho, J. (2021) · arXiv · doi:10.48550/arxiv.2107.00630 · Cited by 374
Scaling Autoregressive Models for Content-Rich Text-to-Image Generation
Yu, J., Xu, Y., Koh, J. Y., Luong, T., Baid, G., Wang, Z., Vasudevan, V., Ku, A., Yang, Y., Ayan, B. K., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.10789 · Cited by 371
StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets
Sauer, A., Schwarz, K., & Geiger, A. (2022) · Special Interest Group on Computer Graphics and Interactive Techniques Conference Proceedings · doi:10.1145/3528233.3530738 · Cited by 353
Make-A-Scene: Scene-Based Text-to-Image Generation with Human Priors
Gafni, O., Polyak, A., Ashual, O., Sheynin, S., Parikh, D., & Taigman, Y. (2022) · arXiv · doi:10.48550/arxiv.2203.13131 · Cited by 332
On Aliased Resizing and Surprising Subtleties in GAN Evaluation
Parmar, G., Zhang, R., & Zhu, J.-Y. (2021) · arXiv · doi:10.48550/arxiv.2104.11222 · Cited by 293
Self-citationProgressive Distillation for Fast Sampling of Diffusion Models
Salimans, T., & Ho, J. (2022) · arXiv · doi:10.48550/arxiv.2202.00512 · Cited by 287
Large Scale Image Completion via Co-Modulated Generative Adversarial Networks
Zhao, S., Cui, J., Sheng, Y., Dong, Y., Liang, X., Chang, E. I., & Xu, Y. (2021) · arXiv · doi:10.48550/arxiv.2103.10428 · Cited by 224 · cited 1×
“…them, GANs have shown unprecedented quality in realism [4, 26– 28, 30, 67], particularly when the GAN loss can be combined with task-specific, auxiliary regression losses to mitigate training instability, ranging from paired image translation [24, 54, 79, 89] to unpaired image editing [37, 55, 94].”
BackgroundMuse: Text-To-Image Generation via Masked Generative Transformers
Chang, H., Zhang, H., Barber, J., Maschinot, A., Lezama, J., Jiang, L., Yang, M.-H., Murphy, K., Freeman, W. T., Rubinstein, M., et al. (2023) · arXiv · doi:10.48550/arxiv.2301.00704 · Cited by 157 · cited 1×
“Distribution Matching Recently, a few classes of generative models have shown success in scaling up to complex datasets by recovering samples that are corrupted by a pre-defined mechanism, such as noise injection [21, 61, 64] or token masking [5, 60, 86].”
BackgroundSelf-citationSong, Y., Dhariwal, P., Chen, M., & Sutskever, I. (2023) · arXiv · doi:10.48550/arxiv.2303.01469 · Cited by 155
Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
Liu, X., Gong, C., & Liu, Q. (2022) · arXiv · doi:10.48550/arxiv.2209.03003 · Cited by 124
LAFITE: Towards Language-Free Training for Text-to-Image Generation
Zhou, Y., Zhang, R., Chen, C., Li, C., Tensmeyer, C., Yu, T., Gu, J., Xu, J., & Sun, T. (2021) · arXiv · doi:10.48550/arxiv.2111.13792 · Cited by 38
Self-citation
Metadata
- Authors
- Yin, T., Gharbi, M., Zhang, R., Shechtman, E., Durand, F., Freeman, W. T., & Park, T.
- Venue
- 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2024-06-16
- Pages
- 6613-6623
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4402726956
- Cited by
- 163
- References
- 54
- FWCI
- 35.59
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.