{"url":"/sota/image-generation-on-imagenet-512x512","task":{"name":"Image Generation","url":"/task/image-generation","note":null},"dataset":{"name":"ImageNet 512x512","url":null},"category":"Computer Vision","categories":["Computer Vision","Medical","Miscellaneous","Natural Language Processing"],"category_note":null,"description":"**Image Generation** (synthesis) is the task of generating new images from an existing dataset.\r\n\r\n- **Unconditional generation** refers to generating samples unconditionally from the dataset, i.e. $p(y)$\r\n- **[Conditional image generation](/task/conditional-image-generation)** (subtask) refers to generating samples conditionally from the dataset, based on a label, i.e. $p(y|x)$.\r\n\r\nIn 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.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [StyleGAN](https://github.com/NVlabs/stylegan) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["FID","NFE","Inception score"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"FID":"lower","NFE":null,"Inception score":"higher"}},"counts":{"rows":52,"rows_with_code":48,"rows_with_paper_page":52,"rows_dated":52,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"EDM2-L + DDO (SD-VAE, 25 steps, DPM-Solver-v3)","metrics":{"FID":"1.21","NFE":"50"},"uses_additional_data":false,"paper_date":"2025-03-03","paper":"/paper/direct-discriminative-optimization-your-1","paper_url":"https://arxiv.org/abs/2503.01103v2","paper_title":"Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator","code":"https://github.com/nvlabs/ddo","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"DDT-XL/2 + UCGM-S (SD-VAE + 150 sampling steps + CFG)","metrics":{"FID":"1.24","NFE":"300"},"uses_additional_data":false,"paper_date":"2025-05-12","paper":"/paper/unified-continuous-generative-models","paper_url":"https://arxiv.org/abs/2505.07447v2","paper_title":"Unified Continuous Generative Models","code":"https://github.com/LINs-Lab/UCGM","n_code_links":1,"syntology":{"n_ran":19,"n_unverified":2,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"DDT-XL/2 + UCGM-S (SD-VAE + 100 sampling steps + CFG)","metrics":{"FID":"1.25","NFE":"200"},"uses_additional_data":false,"paper_date":"2025-05-12","paper":"/paper/unified-continuous-generative-models","paper_url":"https://arxiv.org/abs/2505.07447v2","paper_title":"Unified Continuous Generative Models","code":"https://github.com/LINs-Lab/UCGM","n_code_links":1,"syntology":{"n_ran":19,"n_unverified":2,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"EDM2-XXL Autoguidance","metrics":{"FID":"1.25"},"uses_additional_data":false,"paper_date":"2024-06-04","paper":"/paper/guiding-a-diffusion-model-with-a-bad-version","paper_url":"https://arxiv.org/abs/2406.02507v3","paper_title":"Guiding a Diffusion Model with a Bad Version of Itself","code":"https://github.com/nvlabs/edm2","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":5,"model":"DDT-XL/2(22en6de 675M + guidance interval )","metrics":{"FID":"1.28","Inception score":"305","NFE":"500"},"uses_additional_data":false,"paper_date":"2025-04-08","paper":"/paper/ddt-decoupled-diffusion-transformer-1","paper_url":"https://arxiv.org/abs/2504.05741v2","paper_title":"DDT: Decoupled Diffusion Transformer","code":"https://github.com/MCG-NJU/DDT","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"EDM2- S  Autoguidance (XS, T /16)","metrics":{"FID":"1.34"},"uses_additional_data":false,"paper_date":"2024-06-04","paper":"/paper/guiding-a-diffusion-model-with-a-bad-version","paper_url":"https://arxiv.org/abs/2406.02507v3","paper_title":"Guiding a Diffusion Model with a Bad Version of Itself","code":"https://github.com/nvlabs/edm2","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":7,"model":"SiDA-EDM2-XXL (1.5B)","metrics":{"FID":"1.366","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":8,"model":"SiDA-EDM2-XL (1.1B)","metrics":{"FID":"1.379","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":9,"model":"EDM2-XXL w/ guidance interval","metrics":{"FID":"1.40"},"uses_additional_data":false,"paper_date":"2024-04-11","paper":"/paper/applying-guidance-in-a-limited-interval","paper_url":"https://arxiv.org/abs/2404.07724v2","paper_title":"Applying Guidance in a Limited Interval Improves Sample and Distribution Quality in Diffusion Models","code":"https://github.com/vectorspacelab/omnigen2","n_code_links":2,"syntology":{"n_ran":5,"n_unverified":2,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"SiDA-EDM2-L (777M)","metrics":{"FID":"1.413","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":11,"model":"SiD2","metrics":{"FID":"1.48"},"uses_additional_data":false,"paper_date":"2024-10-25","paper":"/paper/simpler-diffusion-sid2-1-5-fid-on-imagenet512","paper_url":"https://arxiv.org/abs/2410.19324v2","paper_title":"Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"SiDA-EDM2-M (498M)","metrics":{"FID":"1.488","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":13,"model":"SiDA-EDM2-S (280M)","metrics":{"FID":"1.669","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":14,"model":"EDM2-S w/ guidance interval","metrics":{"FID":"1.68"},"uses_additional_data":false,"paper_date":"2024-04-11","paper":"/paper/applying-guidance-in-a-limited-interval","paper_url":"https://arxiv.org/abs/2404.07724v2","paper_title":"Applying Guidance in a Limited Interval Improves Sample and Distribution Quality in Diffusion Models","code":"https://github.com/vectorspacelab/omnigen2","n_code_links":2,"syntology":{"n_ran":5,"n_unverified":2,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"GMem","metrics":{"FID":"1.71"},"uses_additional_data":false,"paper_date":"2024-12-11","paper":"/paper/generative-modeling-with-explicit-memory","paper_url":"https://arxiv.org/abs/2412.08781v1","paper_title":"Generative Modeling with Explicit Memory","code":"https://github.com/lins-lab/gmem","n_code_links":1,"syntology":{"n_ran":8,"n_unverified":7,"n_samples":15,"n_pointer_only_licence":15}},{"rank_in_archive_order":16,"model":"DC-AE-f32 + USiT-2B","metrics":{"FID":"1.72"},"uses_additional_data":false,"paper_date":"2024-10-14","paper":"/paper/deep-compression-autoencoder-for-efficient","paper_url":"https://arxiv.org/abs/2410.10733v8","paper_title":"Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models","code":"https://github.com/mit-han-lab/efficientvit","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":23,"n_samples":30,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"MAR-L, Diff Loss","metrics":{"FID":"1.73"},"uses_additional_data":false,"paper_date":"2024-06-17","paper":"/paper/autoregressive-image-generation-without","paper_url":"https://arxiv.org/abs/2406.11838v3","paper_title":"Autoregressive Image Generation without Vector Quantization","code":"https://github.com/lth14/mar","n_code_links":2,"syntology":{"n_ran":8,"n_unverified":5,"n_samples":13,"n_pointer_only_licence":1}},{"rank_in_archive_order":18,"model":"SIMS","metrics":{"FID":"1.73"},"uses_additional_data":false,"paper_date":"2024-08-29","paper":"/paper/self-improving-diffusion-models-with","paper_url":"https://arxiv.org/abs/2408.16333v1","paper_title":"Self-Improving Diffusion Models with Synthetic Data","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":19,"model":"PaGoDA","metrics":{"FID":"1.80"},"uses_additional_data":false,"paper_date":"2024-05-23","paper":"/paper/pagoda-progressive-growing-of-a-one-step","paper_url":"https://arxiv.org/abs/2405.14822v2","paper_title":"PaGoDA: Progressive Growing of a One-Step Generator from a Low-Resolution Diffusion Teacher","code":"https://github.com/sony/pagoda","n_code_links":1,"syntology":{"n_ran":15,"n_unverified":2,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":20,"model":"EDM2-XXL","metrics":{"FID":"1.81","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":21,"model":"EDM2-XL","metrics":{"FID":"1.85","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":22,"model":"EDM2-L","metrics":{"FID":"1.88","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":23,"model":"SiD-EDM2-XL (1.1B)","metrics":{"FID":"1.888","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":24,"model":"SiD-EDM2-L (777M)","metrics":{"FID":"1.907","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":25,"model":"MAGVIT-v2","metrics":{"FID":"1.91","Inception score":"324.3"},"uses_additional_data":false,"paper_date":"2023-10-09","paper":"/paper/language-model-beats-diffusion-tokenizer-is","paper_url":"https://arxiv.org/abs/2310.05737v3","paper_title":"Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation","code":"https://github.com/jy0205/Pyramid-Flow","n_code_links":3,"syntology":{"n_ran":19,"n_unverified":1,"n_samples":20,"n_pointer_only_licence":0}},{"rank_in_archive_order":26,"model":"SiD-EDM2-XXL (1.5B)","metrics":{"FID":"1.969","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":27,"model":"EDM2-M","metrics":{"FID":"2.01","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":28,"model":"SiD-EDM2-M (498M)","metrics":{"FID":"2.06","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":29,"model":"TiTok-B-128","metrics":{"FID":"2.13"},"uses_additional_data":false,"paper_date":"2024-06-11","paper":"/paper/an-image-is-worth-32-tokens-for","paper_url":"https://arxiv.org/abs/2406.07550v1","paper_title":"An Image is Worth 32 Tokens for Reconstruction and Generation","code":"https://github.com/bytedance/1d-tokenizer","n_code_links":2,"syntology":null},{"rank_in_archive_order":30,"model":"SiDA-EDM2-XS (125M)","metrics":{"FID":"2.156","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":31,"model":"EDM2-S","metrics":{"FID":"2.23","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":32,"model":"DiT-XL/2 with CADS","metrics":{"FID":"2.31"},"uses_additional_data":false,"paper_date":"2023-10-26","paper":"/paper/cads-unleashing-the-diversity-of-diffusion","paper_url":"https://arxiv.org/abs/2310.17347v4","paper_title":"CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":33,"model":"StyleGAN-XL","metrics":{"FID":"2.40"},"uses_additional_data":false,"paper_date":"2022-02-01","paper":"/paper/stylegan-xl-scaling-stylegan-to-large-diverse","paper_url":"https://arxiv.org/abs/2202.00273v2","paper_title":"StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets","code":"https://github.com/autonomousvision/stylegan-xl","n_code_links":2,"syntology":{"n_ran":14,"n_unverified":5,"n_samples":19,"n_pointer_only_licence":0}},{"rank_in_archive_order":34,"model":"TiTok-L-64","metrics":{"FID":"2.49"},"uses_additional_data":false,"paper_date":"2024-06-11","paper":"/paper/an-image-is-worth-32-tokens-for","paper_url":"https://arxiv.org/abs/2406.07550v1","paper_title":"An Image is Worth 32 Tokens for Reconstruction and Generation","code":"https://github.com/bytedance/1d-tokenizer","n_code_links":2,"syntology":null},{"rank_in_archive_order":35,"model":"DiffiT","metrics":{"FID":"2.67","Inception score":"252.12"},"uses_additional_data":false,"paper_date":"2023-12-04","paper":"/paper/diffit-diffusion-vision-transformers-for","paper_url":"https://arxiv.org/abs/2312.02139v3","paper_title":"DiffiT: Diffusion Vision Transformers for Image Generation","code":"https://github.com/nvlabs/diffit","n_code_links":1,"syntology":{"n_ran":18,"n_unverified":7,"n_samples":25,"n_pointer_only_licence":25}},{"rank_in_archive_order":36,"model":"SiD-EDM2-S (280M)","metrics":{"FID":"2.707","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":37,"model":"DiT-XL/2 with SA-Solver","metrics":{"FID":"2.80"},"uses_additional_data":false,"paper_date":"2023-09-10","paper":"/paper/sa-solver-stochastic-adams-solver-for-fast","paper_url":"https://arxiv.org/abs/2309.05019v2","paper_title":"SA-Solver: Stochastic Adams Solver for Fast Sampling of Diffusion Models","code":"https://github.com/scxue/SA-Solver","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":38,"model":"DiMR-XL/3R","metrics":{"FID":"2.89"},"uses_additional_data":false,"paper_date":"2024-06-13","paper":"/paper/alleviating-distortion-in-image-generation","paper_url":"https://arxiv.org/abs/2406.09416v2","paper_title":"Alleviating Distortion in Image Generation via Multi-Resolution Diffusion Models and Time-Dependent Layer Normalization","code":"https://github.com/qihao067/DiMR","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":6,"n_samples":16,"n_pointer_only_licence":0}},{"rank_in_archive_order":39,"model":"EDM2-XS","metrics":{"FID":"2.91","NFE":"126"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/analyzing-and-improving-the-training-dynamics","paper_url":"https://arxiv.org/abs/2312.02696v2","paper_title":"Analyzing and Improving the Training Dynamics of Diffusion Models","code":"https://github.com/nvlabs/edm2","n_code_links":7,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":6}},{"rank_in_archive_order":40,"model":"GIVT-Causal-L+A","metrics":{"FID":"2.92"},"uses_additional_data":false,"paper_date":"2023-12-04","paper":"/paper/givt-generative-infinite-vocabulary","paper_url":"https://arxiv.org/abs/2312.02116v4","paper_title":"GIVT: Generative Infinite-Vocabulary Transformers","code":"https://github.com/google-research/big_vision","n_code_links":2,"syntology":null},{"rank_in_archive_order":41,"model":"DiT-XL/2","metrics":{"FID":"3.04","Inception score":"240.82"},"uses_additional_data":false,"paper_date":"2022-12-19","paper":"/paper/scalable-diffusion-models-with-transformers","paper_url":"https://arxiv.org/abs/2212.09748v2","paper_title":"Scalable Diffusion Models with Transformers","code":"https://github.com/huggingface/diffusers","n_code_links":13,"syntology":{"n_ran":15,"n_unverified":5,"n_samples":20,"n_pointer_only_licence":4}},{"rank_in_archive_order":42,"model":"MAGVIT-v2 (w/o guidance)","metrics":{"FID":"3.07","Inception score":"213.1"},"uses_additional_data":false,"paper_date":"2023-10-09","paper":"/paper/language-model-beats-diffusion-tokenizer-is","paper_url":"https://arxiv.org/abs/2310.05737v3","paper_title":"Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation","code":"https://github.com/jy0205/Pyramid-Flow","n_code_links":3,"syntology":{"n_ran":19,"n_unverified":1,"n_samples":20,"n_pointer_only_licence":0}},{"rank_in_archive_order":43,"model":"SiD-EDM2-XS (125M)","metrics":{"FID":"3.353","NFE":"1"},"uses_additional_data":false,"paper_date":"2024-10-19","paper":"/paper/adversarial-score-identity-distillation","paper_url":"https://arxiv.org/abs/2410.14919v4","paper_title":"Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step","code":"https://github.com/mingyuanzhou/sid","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":44,"model":"DPC-U","metrics":{"FID":"3.54","Inception score":"350.2"},"uses_additional_data":false,"paper_date":"2022-09-29","paper":"/paper/discrete-predictor-corrector-diffusion-models","paper_url":"https://openreview.net/forum?id=VM8batVBWvg","paper_title":"Discrete Predictor-Corrector Diffusion Models for Image Synthesis","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":45,"model":"Latent Diffusion (LDM-4-G)","metrics":{"FID":"3.60","Inception score":"247.67"},"uses_additional_data":false,"paper_date":"2021-12-20","paper":"/paper/high-resolution-image-synthesis-with-latent","paper_url":"https://arxiv.org/abs/2112.10752v2","paper_title":"High-Resolution Image Synthesis with Latent Diffusion Models","code":"https://github.com/compvis/stable-diffusion","n_code_links":41,"syntology":{"n_ran":21,"n_unverified":7,"n_samples":28,"n_pointer_only_licence":5}},{"rank_in_archive_order":46,"model":"Poly-INR","metrics":{"FID":"3.81"},"uses_additional_data":false,"paper_date":"2023-03-20","paper":"/paper/polynomial-implicit-neural-representations","paper_url":"https://arxiv.org/abs/2303.11424v1","paper_title":"Polynomial Implicit Neural Representations For Large Diverse Datasets","code":"https://github.com/rajhans0/poly_inr","n_code_links":1,"syntology":null},{"rank_in_archive_order":47,"model":"ADM-G, ADM-U","metrics":{"FID":"3.85","Inception score":"221.72"},"uses_additional_data":false,"paper_date":"2021-05-11","paper":"/paper/diffusion-models-beat-gans-on-image-synthesis","paper_url":"https://arxiv.org/abs/2105.05233v4","paper_title":"Diffusion Models Beat GANs on Image Synthesis","code":"https://github.com/openai/guided-diffusion","n_code_links":21,"syntology":{"n_ran":33,"n_unverified":17,"n_samples":50,"n_pointer_only_licence":16}},{"rank_in_archive_order":48,"model":"simple diffusion (U-Net)","metrics":{"FID":"4.28","Inception score":"171"},"uses_additional_data":false,"paper_date":"2023-01-26","paper":"/paper/simple-diffusion-end-to-end-diffusion-for","paper_url":"https://arxiv.org/abs/2301.11093v2","paper_title":"Simple diffusion: End-to-end diffusion for high resolution images","code":"https://github.com/fashn-AI/tryondiffusion","n_code_links":1,"syntology":null},{"rank_in_archive_order":49,"model":"MaskGIT (a=0.05)","metrics":{"FID":"4.46","Inception score":"342.0"},"uses_additional_data":false,"paper_date":"2022-02-08","paper":"/paper/maskgit-masked-generative-image-transformer","paper_url":"https://arxiv.org/abs/2202.04200v1","paper_title":"MaskGIT: Masked Generative Image Transformer","code":"https://github.com/lucidrains/soundstorm-pytorch","n_code_links":9,"syntology":{"n_ran":14,"n_unverified":7,"n_samples":21,"n_pointer_only_licence":4}},{"rank_in_archive_order":50,"model":"simple diffusion (U-ViT, L)","metrics":{"FID":"4.53","Inception score":"205.3"},"uses_additional_data":false,"paper_date":"2023-01-26","paper":"/paper/simple-diffusion-end-to-end-diffusion-for","paper_url":"https://arxiv.org/abs/2301.11093v2","paper_title":"Simple diffusion: End-to-end diffusion for high resolution images","code":"https://github.com/fashn-AI/tryondiffusion","n_code_links":1,"syntology":null},{"rank_in_archive_order":51,"model":"MaskGIT","metrics":{"FID":"7.32","Inception score":"156.0"},"uses_additional_data":false,"paper_date":"2022-02-08","paper":"/paper/maskgit-masked-generative-image-transformer","paper_url":"https://arxiv.org/abs/2202.04200v1","paper_title":"MaskGIT: Masked Generative Image Transformer","code":"https://github.com/lucidrains/soundstorm-pytorch","n_code_links":9,"syntology":{"n_ran":14,"n_unverified":7,"n_samples":21,"n_pointer_only_licence":4}},{"rank_in_archive_order":52,"model":"ADM-G","metrics":{"FID":"7.72","Inception score":"172.71"},"uses_additional_data":false,"paper_date":"2021-05-11","paper":"/paper/diffusion-models-beat-gans-on-image-synthesis","paper_url":"https://arxiv.org/abs/2105.05233v4","paper_title":"Diffusion Models Beat GANs on Image Synthesis","code":"https://github.com/openai/guided-diffusion","n_code_links":21,"syntology":{"n_ran":33,"n_unverified":17,"n_samples":50,"n_pointer_only_licence":16}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 7,081 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":7081,"papers_extracted_not_yet_verified":217,"boards_without_verdict":27,"papers_not_yet_extracted":2325},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":40,"rows_with_any_sample_ran":40,"distinct_papers_with_graph_line":18,"distinct_papers_with_any_sample_ran":18,"samples_over_distinct_papers":{"n_ran":225,"n_unverified":105,"n_samples":330,"n_pointer_only_licence":80,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":409,"n_unverified":166,"n_samples":575,"n_pointer_only_licence":144,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}