ddl: September 2024

This survey reviews the latest advancements in efficient diffusion model. This paper aims to provide a comprehensive overview of the methodologies, applications, and future directions in this burgeoning field.

• [Abstract]
• [Introduction]
• [Main Content]
  • [Algorithm]
    • Efficient Sampling

      • Sampling Scheduling

        • Align Your Steps: Optimizing Sampling Schedules in Diffusion Models
        • Parallel Sampling of Diffusion Models
        • Simple Hierarchical Planning with Diffusion
        • Accelerating Parallel Sampling of Diffusion Models
        • A Unified Sampling Framework for Solver Searching of Diffusion Probabilistic Models
        • PipeFusion: Displaced Patch Pipeline Parallelism for Inference of Diffusion Transformer Models
        • Deep Equilibrium Approaches to Diffusion Models
        • Learning to Efficiently Sample from Diffusion Probabilistic Models
        • On Fast Sampling of Diffusion Probabilistic Model
        • DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models

      • Data-Dependent Adaptive Priors

        • PriorGrad: Improving Conditional Denoising Diffusion Models with Data-Dependent Adaptive Prior
        • DiGress: Discrete Denoising diffusion for graph generation
        • DECOMPDIFF: Diffusion Models with Decomposed Priors for Structure-Based Drug Design
        • Leapfrog diffusion model for stochastic trajectory prediction

      • Partial Sampling

        • On distillation of guided diffusion models
        • Snapfusion: Text-to-image diffusion model on mobile devices within two seconds
        • Consistent accelerated inference via confident adaptive transformers
        • Confident adaptive language modeling
        • A Simple Early Exiting Framework for Accelerated Sampling in Diffusion Models
        • Semi-parametric neural image synthesis
        • kNN-Diffusion: Image Generation via Large-Scale Retrieval
        • Re-Imagen: Retrieval-Augmented Text-to-Image Generator
        • ReDi: efficient learning-free diffusion inference via trajectory retrieval
    • Noise Schedule
      • Strategic Noise Schedules
        • Denoising Diffusion Probabilistic Models
        • Improved Denoising Diffusion Probabilistic Models
        • Imprvoed Noise Schedule for Diffusion Training
        • A Cheaper and Better Diffusion Language Model with Soft-masked Noise
      • Adaptive Noise Schedules
        • Denoising Diffusion Implicit Models
        • ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting
        • Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment
        • Text Diffusion Model with Encoder-Decoder Transformers for Sequence-to-Sequence Generation
    • SDE and ODE Solvers
      • SDE Solver
        • Diffusion Normalizing Flow
        • Gaussian Mixture Solvers for Diffusion Models
        • Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations
        • SA-Solver: Stochastic Adams Solver for Fast Sampling of Diffusion Models
        • Diffusion Models with Deterministic Normalizing Flow Priors
      • ODE Solver
        • Denoising diffusion implicit models
        • GDDIM: GENERALIZED DENOISING DIFFUSION IMPLICIT MODELS
        • DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps
        • FAST SAMPLING OF DIFFUSION MODELS WITH EXPONENTIAL INTEGRATOR
        • Improved Techniques for Maximum Likelihood Estimation for Diffusion ODEs
        • Denoising MCMC for Accelerating Diffusion-Based Generative Models
    • SGM Optimization
    • Latent Diffusion
    • Compression
      • Quantization
        • Post-Training Quantization
          • Post-training quantization on diffusion models
          • Q-diffusion: Quantizing diffusion models
          • Leveraging early-stage robustness in diffusion models for efficient and high-quality image synthesis
          • Ptqd: Accurate post-training quantization for diffusion models
        • Quantization-Aware Training
          • Temporal dynamic quantization for diffusion models
          • Efficientdm: Efficient quantization-aware fine-tuning of low-bit diffusion models
      • Pruning
        • Structural pruning for diffusion models
        • LD-Pruner: Efficient Pruning of Latent Diffusion Models using Task-Agnostic Insights
        • LayerMerge: Neural Network Depth Compression through Layer Pruning and Merging
        • Laptop-diff: Layer pruning and normalized distillation for compressing diffusion models
      • Knowledge Distillation
        • Vector Field Distillation
          • Knowledge distillation in iterative generative models for improved sampling speed
          • Progressive distillation for fast sampling of diffusion models
          • On distillation of guided diffusion models
          • Consistency models
          • Flow straight and fast: Learning to generate and transfer data with rectified flow
          • Optimizing DDPM Sampling with Shortcut Fine-Tuning
          • Fast inference in denoising diffusion models via mmd finetuning
        • Generator Distillation
          • Nerf: Representing scenes as neural radiance fields for view synthesis
          • DreamFusion: Text-to-3D using 2D Diffusion
          • Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation
          • Diff-instruct: A universal approach for transferring knowledge from pre-trained diffusion models
          • 3d paintbrush: Local stylization of 3d shapes with cascaded score distillation
  • [System]
    • Optimized Hardware-Software Co-Design
      • Speed is all you need: On-device acceleration of large diffusion models via gpu-aware optimizations
      • SDA: Low-Bit Stable Diffusion Acceleration on Edge FPGAs
      • A 28.6 mJ/iter Stable Diffusion Processor for Text-to-Image Generation with Patch Similarity-based Sparsity Augmentation and Text-based Mixed-Precision
      • Efficient memory management for large language model serving with pagedattention
      • Flightllm: Efficient large language model inference with a complete mapping flow on fpgas
    • Parallel Computing
    • Caching Technique
  • [Application]
    • Video Generation
• [Evaluation]
• [Conclusion]

• Post-training quantization on diffusion models
• Q-diffusion: Quantizing diffusion models
• Leveraging early-stage robustness in diffusion models for efficient and high-quality image synthesis
• Ptqd: Accurate post-training quantization for diffusion models

• Temporal dynamic quantization for diffusion models
• Efficientdm: Efficient quantization-aware fine-tuning of low-bit diffusion models

• Structural pruning for diffusion models
• LD-Pruner: Efficient Pruning of Latent Diffusion Models using Task-Agnostic Insights
• LayerMerge: Neural Network Depth Compression through Layer Pruning and Merging
• Laptop-diff: Layer pruning and normalized distillation for compressing diffusion models

• Knowledge distillation in iterative generative models for improved sampling speed
• Progressive distillation for fast sampling of diffusion models
• On distillation of guided diffusion models
• Consistency models
• Flow straight and fast: Learning to generate and transfer data with rectified flow
• Optimizing DDPM Sampling with Shortcut Fine-Tuning
• Fast inference in denoising diffusion models via mmd finetuning

• Nerf: Representing scenes as neural radiance fields for view synthesis
• DreamFusion: Text-to-3D using 2D Diffusion
• Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation
• Diff-instruct: A universal approach for transferring knowledge from pre-trained diffusion models
• 3d paintbrush: Local stylization of 3d shapes with cascaded score distillation

https://arxiv.org/pdf/2312.03863

https://arxiv.org/pdf/2210.09292

https://arxiv.org/pdf/2209.00796

https://arxiv.org/pdf/2404.07771

• SDE Solvers
• ODE Solvers
• Optimized Discretization
• Truncated Diffusion
• Knowledge Distillation

1. LanguageFlow: Advancing Diffusion Language Generation with Probabilistic Flows, NAACL 24 [Paper]

   ODE-solver -> Using Recited Flow to replace ODE Dataset: E2E/NLG/ART

2. Stable Target Field for Reduced Variance Score Estimation in Diffusion Models, ICLR 23 [Paper]

   Training Process Using STD to enhance SGMs, accelerating training process Dataset: CIFAR-10

3. Learning Energy-Based Models by Cooperative Diffusion Recovery Likelihood, ICLR 24 [Paper]

   Cooperative Training Dataset: CIFAR-10/ImageNet/Celeb-A

4. Autodiffusion: Training-free optimization of time steps and architectures for automated diffusion model acceleration, ICCV 23 [paper]

5. Improving Training Efficiency of Diffusion Models via Multi-Stage Framework and Tailored Multi-Decoder Architecture, CVPR 24 [paper]

6. DreamFusion: Text-to-3D using 2D Diffusion [paper]

7. Fast Training of Diffusion Models with Masked Transformers, TMLR 24 [paper]

8. MDTv2: Masked Diffusion Transformer is a Strong Image Synthesizer, ICCV 23 [paper]

9. BK-SDM: A Lightweight, Fast, and Cheap Version of Stable Diffusion, ECCV24 [paper]

1. Align Your Steps: Optimizing Sampling Schedules in Diffusion Models, ICML 24 [Paper]

   Dataset: FFHQ/CIFAR-10/ImageNet/WebVid10M

2. Parallel Sampling of Diffusion Models, NIPS 23 [Paper]

   Dataset: LSUN/Square/PushT/Franka Kitchen

3. Simple Hierarchical Planning with Diffusion, ICLR 24 [Paper]

   Dataset: Maze2D/AntMaze/Gym-MuJoCo/FrankaKitchen

4. Accelerating Parallel Sampling of Diffusion Models, ICML 24 [Paper]

   Dataset: ImageNet

5. A Unified Sampling Framework for Solver Searching of Diffusion Probabilistic Models, ICLR 24 [Paper]

   Dataset: CIFAR-10/CelebA/ImageNet-64/LSUN-Bedroom

6. PipeFusion: Displaced Patch Pipeline Parallelism for Inference of Diffusion Transformer Models, arXiv [Paper]

   Dataset: COCO Captions 2014

7. Accelerating Guided Diffusion Sampling with Splitting Numerical Methods, ICLR 23 [Paper]

   Dataset: LSUN/FFHQ

8. Diffusion Glancing Transformer for Parallel Sequence-to-Sequence Learning, NAACL 24 [Paper]

   Dataset: QQP/MS-COCO

9. Deep Equilibrium Approaches to Diffusion Models, NIPS 22 [Paper]

   Dataset: CIFAR-10/CelebA/LSUN

10. Effective Real Image Editing with Accelerated Iterative Diffusion Inversion, ICCV 23 [Paper]

    Dataset: AFHQ/COCO

1. DecompDiff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design, ICML 23 [Paper]

   Dataset: CrossDocked2020

2. Diffusion Posterior Sampling for Linear Inverse Problem Solving: A Filtering Perspective, ICLR 24 [Paper]

   Dataset: FFHQ-1kvalidation/ImageNet-1k-validation

3. Generalized Deep 3D Shape Prior via Part-Discretized Diffusion Process, CVPR 23 [Paper],

   Dataset: ShapeNet

1. Leapfrog Diffusion Model for Stochastic Trajectory Prediction, CVPR 23 [Paper]

   Dataset: NBA/NFL/SDD/ETH-UCY

2. SnapFusion: Text-to-Image Diffusion Model on Mobile Devices within Two Seconds, NIPS 23 [Paper]

   Dataset: MS-COCO

3. ReDi: Efficient Learning-Free Diffusion Inference via Trajectory Retrieval, ICML 23 [Paper]

   Dataset: MS-COCO

4. Data-free Distillation of Diffusion Models with Bootstrapping, ICML 24 [Paper]

   Dataset: FFHQ/LSUN-Bedroom

5. A Simple Early Exiting Framework for Accelerated Sampling in Diffusion Models, ICML 24 [Paper]

   Dataset: ImageNet/CelebA

6. David helps Goliath: Inference-Time Collaboration Between Small Specialized and Large General Diffusion LMs, NAACL 24 [Paper]

   Dataset: DOLLY

7. On Distillation of Guided Diffusion Models, CVPR 23 [Paper]

   Dataset: ImageNet/CIFAR-10

8. Learning Stackable and Skippable LEGO Bricks for Efficient, Reconfigurable, and Variable-Resolution Diffusion Modeling, ICLR 24 [Paper]

   Dataset: CIFAR-10/ImageNet

9. Relay Diffusion: Unifying diffusion process across resolutions for image synthesis, ICLR 24 [Paper]

   Dataset: CelebA-HQ/ImageNet

10. Semi-Implicit Denoising Diffusion Models (SIDDMs), NIPS 23 [Paper]

    Dataset: CIFAR-10/CelebA-HQ/ImageNet

11. Directly Fine-Tuning Diffusion Models on Differentiable Rewards, ICLR 24 [Paper]

    Dataset: LAION/HPDv2

12. InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image Generation, ICLR 24 [Paper]

    Dataset: MS COCO

13. Fast Sampling of Diffusion Models via Operator Learning, ICML 23 [Paper]

    Dataset: CIFAR-10/ImageNet-64

1. Denoising Diffusion Probabilistic Models, NIPS 20 [Paper]

2. Improved Denoising Diffusion Probabilistic Models, PMLR 21 [Paper]

3. Improved Noise Schedule for Diffusion Training, arxiv [Paper]

4. A Cheaper and Better Diffusion Language Model with Soft-masked Noise, arxiv [Paper]

1. Denoising Diffusion Implicit Models, arxiv [Paper]

2. ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting, NIPS 24 [Paper]

3. Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment, arxiv [Paper]

4. Text Diffusion Model with Encoder-Decoder Transformers for Sequence-to-Sequence Generation, ACL 24 [Paper]

1. Sampling is as easy as learning the score: theory for diffusion models with minimal data assumptions, ICLR 23 [Paper]

   Dataset: CIFAR-10/ImageNet 64x64

2. Improved Techniques for Maximum Likelihood Estimation for Diffusion ODEs, ICML 23 [Paper]

   Dataset: CIFAR-10/ImageNet-32

3. Gaussian Mixture Solvers for Diffusion Models, NIPS 23 [Paper]

   Dataset: CIFAR-10/ImageNet

4. Denoising MCMC for Accelerating Diffusion-Based Generative Models, ICML 23 [Paper]

   Dataset: CIFAR11/CelebA-HQ-256/FFHQ-1024

5. DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps, NIPS 22 [Paper]

   Dataset: CIFAR-10/CelebA/ImageNet/LSUN

6. Score-Based Generative Modeling through Stochastic Differential Equations, ICLR 21 [Paper]

   Dataset: CIFAR-10/LSUN/CelebA-HQ

7. Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations, ICML 24 [Paper]

   Dataset: text8/CIFAR-10

8. Diffusion Normalizing Flow, NIPS 21 [Paper]

   Dataset: CIFAR-10/MNIST

9. On the Trajectory Regularity of ODE-based Diffusion Sampling, ICML 24 [Paper]

   Dataset: LSUN Bedroom/CIFAR-10/ImageNet-64/FFHQ

1. FP-Diffusion: Improving Score-based Diffusion Models by Enforcing the Underlying Score Fokker-Planck Equation, ICML 23 [Paper]

   Dataset:MNIST/Fashion MNIST/CIFAR-10/ImageNet32

2. Accelerating Score-Based Generative Models with Preconditioned Diffusion Sampling, ECCV 22 [Paper]

   Dataset:MNIST/CIFAR-10/LSUN/FFHQ

3. Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models, ICML 23 [Paper]

   Dataset:ImageNet/CIFAR-10/CelebA/FFHQ

4. Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models, ICLR 22 [Paper]

   Dataset: CIFAR-10/ImageNet

5. Discrete Predictor-Corrector Diffusion Models for Image Synthesis, ICLR 23 [Paper]

   Dataset: ImageNet/Places2

1. Fast Timing-Conditioned Latent Audio Diffusion, ICML 24 [Paper]

   Dataset: MusicCaps/AudioCaps

2. AudioLDM: Text-to-Audio Generation with Latent Diffusion Models, ICML 23 [Paper]

   Dataset: AudioSet/AudioCaps/Freesound/BBC Sound Effect library

3. Executing Your Commands via Motion Diffusion in Latent Space, CVPR 23 [Paper]

   Dataset: HumanML3D/KIT/AMASS/HumanAct12/UESTC

4. Efficient Video Diffusion Models via Content-Frame Motion-Latent Decomposition, ICLR 24 [Paper]

   Dataset: UCF-101/WebVid-10M/MSR-VTT

5. Mixed-Type Tabular Data Synthesis with Score-based Diffusion in Latent Space, ICLR 24 [Paper]

   Dataset: Adult/Default/Shoppers/Magic/Faults/Beijing/News

6. High-Resolution Image Synthesis With Latent Diffusion Models, CVPR 22 [Paper]

   Dataset: ImageNet/CelebA-HQ/FFHQ/LSUN-Churches/LSUN-Bedrooms

7. Hyperbolic Geometric Latent Diffusion Model for Graph Generation, ICML 24 [Paper]

   Dataset: SBM/BA/Community/Ego/Barabasi-Albert/Grid/Cora/Citeseer/Polblogs/MUTAG/IMDB-B/PROTEINS/COLLAB

8. Latent 3D Graph Diffusion, ICLR 24 [Paper]

   Dataset: ChEMBL/PubChemQC/QM9/Drugs

9. PnP Inversion: Boosting Diffusion-based Editing with 3 Lines of Code, ICLR 24 [Paper]

   Dataset: PIE-Bench

10. Cross-view Masked Diffusion Transformers for Person Image Synthesis, ICML 24 [Paper]

    Dataset: DeepFashion/ImageNet

11. Towards Consistent Video Editing with Text-to-Image Diffusion Models, NIPS 23 [Paper]

    Dataset: DAVIS

12. Video Probabilistic Diffusion Models in Projected Latent Space, CVPR 23 [Paper]

    Dataset: UCF101/SkyTimelapse

13. Conditional Image-to-Video Generation With Latent Flow Diffusion Models, CVPR 23 [Paper],

    Dataset: MUG

14. Diffusion Autoencoders: Toward a Meaningful and Decodable Representation, CVPR 22 [Paper]

    Dataset: FFHQ/CelebA-HQ

15. Adapt and Diffuse: Sample-adaptive Reconstruction via Latent Diffusion Models, ICML 24 [Paper]

    Dataset: CelebA-HQ/LSUN-Bedrooms

16. Dimensionality-Varying Diffusion Process, CVPR 23 [Paper]

    Dataset: CIFAR-10/LSUN-Bedroom/LSUN-Church/LSUN-Cat/FFHQ

17. Vector Quantized Diffusion Model for Text-to-Image Synthesis, CVPR 22 [Paper]

    Dataset: CUB-200/Oxford-102/MSCOCO

1. Bespoke Non-Stationary Solvers for Fast Sampling of Diffusion and Flow Models, ICML 24 [Paper]

   Dataset: MS-COCO/LibriSpeech/ImageNet-64

2. GENIE: Higher-Order Denoising Diffusion Solvers, NIPS 22 [Paper]

   Dataset: CIFAR-10/LSUN,/ImageNet/AFHQv2

3. Diffusion Probabilistic Model Made Slim, CVPR 23 [Paper]

   Dataset: ImageNet/MS-COCO

1. Post-Training Quantization on Diffusion Models, CVPR 23 [Paper]

   Dataset: ImageNet/CIFAR-10

2. Q-Diffusion: Quantizing Diffusion Models, ICCV 23 [Paper]

   Dataset: CIFAR-10/LSUN Bedrooms/LSUN Church-Outdoor

3. PTQD: Accurate Post-Training Quantization for Diffusion Models, NIPS 23 [Paper]

   Dataset: ImageNet/LSUN

4. Binary Latent Diffusion, CVPR 23 [Paper]

   Dataset: LSUN Churches/FFHQ/CelebA-HQ/ImageNet-1K

5. DiffFit: Unlocking Transferability of Large Diffusion Models via SimpleParameter-efficient Fine-Tuning, ICCV 23 [Paper]

   Dataset: ImageNet

6. Würstchen: An Efficient Architecture for Large-Scale Text-to-Image Diffusion Models, ICLR 24 [Paper]

   Dataset: COCO-30K

7. Leveraging Early-Stage Robustness in Diffusion Models for Efficient and High-Quality Image Synthesis, NIPS 23 [Paper]

   Dataset: LSUN

1. Structural Pruning for Diffusion Models, NIPS 23 [Paper]

   Dataset: CIFAR-10/CelebA-HQ/LSUN/ImageNet

1. Infinite Resolution Diffusion with Subsampled Mollified States, ICLR 24 [Paper]

   Dataset: FFHQ/LSUN Church/CelebA-HQ

2. Fast Ensembling with Diffusion Schrödinger Bridge, ICLR 24 [Paper]

   Dataset: CIFAR-10/CIFAR-100/TinyImageNet

3. Text Diffusion Model with Encoder-Decoder Transformers for Sequence-to-Sequence Generation, NAACL 24 [Paper]

   Dataset: QQP/Wiki-Auto/Quasar-T/CCD/IWSLT14/WMT14

1. Neural Diffusion Processes, ICML 23 [Paper]

   Dataset: MNIST/CELEBA

2. Score Regularized Policy Optimization through Diffusion Behavior, ICLR 24 [Paper]

   Benchmark: BEAR/TD3+BC/IQL

3. Efficient and Degree-Guided Graph Generation via Discrete Diffusion Modeling, ICML 23 [Paper]

   Dataset: Community/Ego/Polblogs/Cora/Road-Minnesota/PPI/QM9

4. Decomposed Diffusion Sampler for Accelerating Large-Scale Inverse Problems, ICLR 24 [Paper]

   Dataset: fastMRI knee/AAPM 256×256

5. Soft Mixture Denoising: Beyond the Expressive Bottleneck of Diffusion Models, ICLR 24 [Paper]

   Dataset: CIGAR-10/LSUN-Conference

1. DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models, CVPR24 [paper]

2. PipeFusion: Displaced Patch Pipeline Parallelism for Inference of Diffusion Transformer Models [paper]

3. DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers [paper]

4. SwiftDiffusion: Efficient Diffusion Model Serving with Add-on Modules [paper]

1. DiffusionPipe: Training Large Diffusion Models with Efficient Pipelines, Arxiv [paper]

1. Speed Is All You Need: On-Device Acceleration of Large Diffusion Models via GPU-Aware Optimizations, CVPR 23 [paper]

2. SDA: Low-Bit Stable Diffusion Acceleration on Edge FPGAs, FPL 24 [paper]

3. A 28.6 mJ/iter Stable Diffusion Processor for Text-toImage Generation with Patch Similarity-based Sparsity Augmentation and Text-based Mixed-Precision [paper]

1. Approximate Caching for Efficiently Serving Text-to-Image Diffusion Models, NSDI24 [paper]

2. Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching [paper]

3. DeepCache: Accelerating Diffusion Models for Free, CVPR 24 [paper]

1. DITTO: Diffusion Inference-Time T-Optimization for Music Generation, ICML 24 [Paper]

   Dataset: Wikifonia Lead-Sheet/MusicCaps

   Application Task: Text-to-Music

2. Inf-DiT: Upsampling Any-Resolution Image with Memory-Efficient Diffusion Transformer, arXiv [Paper]

   Dataset: HPDv2/DIV2K/LAION-5B/Datacomp

   Application Task: High-Resolution Image Generation

3. Inserting Anybody in Diffusion Models via Celeb Basis, NIPS 23 [Paper]

   Dataset: LAION/StyleGAN

   Application Task: Personalized Image Generation

4. Efficient Spatially Sparse Inference for Conditional GANs and Diffusion Models, NIPS 22 [Paper]

   Dataset: LSUN/Cityscapes

   Application Task: Image Editing

5. DiffS2UT: A Semantic Preserving Diffusion Model for Textless Direct Speech-to-Speech Translation, EMNLP 23 [Paper]

   Dataset: VoxPopuli-S2S/Europarl-ST

   Application Task: Audio-to-Audio

6. DiffIR: Efficient Diffusion Model for Image Restoration, ICCV 23 [Paper]

   Dataset: CelebA-HQ, LSUN Bedrooms, Places-Standard

   Application Task: Image Restoration

7. Wavelet Diffusion Models Are Fast and Scalable Image Generators, CVPR 23[Paper]

   Dataset: CIFAR-10/STL-10/CelebA-HQ/LSUN-Church

   Application Task: Image Generation

8. PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis, ICLR 24 [Paper]

   Dataset: LAION/SAM/JourneyDB

   Appllication Task: Text-to-Image Generation

9. Non-autoregressive Conditional Diffusion Models for Time Series Prediction, ICML 23 [Paper]

   Dataset: NorPool/Caiso/Traffic/Electricity/Weather/Exchange/ETTh1/ETTm1/Wind

   Application Task: Time Series Prediction

10. IM-3D: Iterative Multiview Diffusion and Reconstruction for High-Quality 3D Generation, ICML 24 [Paper]

    Dataset: Objaverse Application Task: 3D-Object Generation