SOTAVerified

Image Generation

Image Generation (synthesis) is the task of generating new images from an existing dataset.

  • Unconditional generation refers to generating samples unconditionally from the dataset, i.e. $p(y)$
  • Conditional image generation (subtask) refers to generating samples conditionally from the dataset, based on a label, i.e. $p(y|x)$.

In this section, you can find state-of-the-art leaderboards for unconditional generation. For conditional generation, and other types of image generations, refer to the subtasks.

( Image credit: StyleGAN )

Papers

Showing 201225 of 6689 papers

TitleStatusHype
Deciphering Oracle Bone Language with Diffusion ModelsCode3
DDT: Decoupled Diffusion TransformerCode3
Kiss3DGen: Repurposing Image Diffusion Models for 3D Asset GenerationCode3
DALL-Eval: Probing the Reasoning Skills and Social Biases of Text-to-Image Generation ModelsCode3
3DIS-FLUX: simple and efficient multi-instance generation with DiT renderingCode3
AP-LDM: Attentive and Progressive Latent Diffusion Model for Training-Free High-Resolution Image GenerationCode3
3DIS: Depth-Driven Decoupled Instance Synthesis for Text-to-Image GenerationCode3
CtrLoRA: An Extensible and Efficient Framework for Controllable Image GenerationCode3
Kandinsky 3: Text-to-Image Synthesis for Multifunctional Generative FrameworkCode3
On the Trajectory Regularity of ODE-based Diffusion SamplingCode3
DiLightNet: Fine-grained Lighting Control for Diffusion-based Image GenerationCode3
Behavior Generation with Latent ActionsCode3
Intelligent Grimm - Open-ended Visual Storytelling via Latent Diffusion ModelsCode3
InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image GenerationCode3
Inf-DiT: Upsampling Any-Resolution Image with Memory-Efficient Diffusion TransformerCode3
Deep Generative Models on 3D Representations: A SurveyCode3
PanoHead: Geometry-Aware 3D Full-Head Synthesis in 360degCode3
Personalized Image Generation with Deep Generative Models: A Decade SurveyCode3
Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image SynthesisCode3
AnimeGamer: Infinite Anime Life Simulation with Next Game State PredictionCode3
ControlAR: Controllable Image Generation with Autoregressive ModelsCode3
Anim-Director: A Large Multimodal Model Powered Agent for Controllable Animation Video GenerationCode3
An Image is Worth 32 Tokens for Reconstruction and GenerationCode3
Consistency Models Made EasyCode3
Consistency Flow Matching: Defining Straight Flows with Velocity ConsistencyCode3
Show:102550
← PrevPage 9 of 268Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Improved DDPMFID12.3Unverified
2ADMFID11.84Unverified
3BigGAN-deepFID8.1Unverified
4Polarity-BigGANFID6.82Unverified
5VQGAN+Transformer (k=mixed, p=1.0, a=0.005)FID6.59Unverified
6MaskGITFID6.18Unverified
7VQGAN+Transformer (k=600, p=1.0, a=0.05)FID5.2Unverified
8CDMFID4.88Unverified
9ADM-GFID4.59Unverified
10RINFID4.51Unverified
#ModelMetricClaimedVerifiedStatus
1PresGANFID52.2Unverified
2RESFLOWFID48.29Unverified
3Residual FlowFID46.37Unverified
4GLF+perceptual loss (ours)FID44.6Unverified
5ProdPoly no activation functionsFID40.45Unverified
6ProdPoly no activation functionsFID36.77Unverified
7ACGANFID35.47Unverified
8DenseFlow-74-10FID34.9Unverified
9NVAE w/ flowFID32.53Unverified
10QSNGANFID31.97Unverified
#ModelMetricClaimedVerifiedStatus
1GLIDE + CLSFID30.87Unverified
2GLIDE + CLIPFID30.46Unverified
3GLIDE + CLS-FREEFID29.22Unverified
4GLIDE + CLIP + CLS + CLS-FREEFID29.18Unverified
5PGMGANFID21.73Unverified
6CLR-GANFID20.27Unverified
7FMFID14.45Unverified
8CT (Direct Generation, NFE=1)FID13Unverified
9CT (Direct Generation, NFE=2)FID11.1Unverified
10GLIDE +CLSKID7.95Unverified