Papers › Taming Transformers for High-Resolution Image Synthesis

Taming Transformers for High-Resolution Image Synthesis

17 Dec 2020CVPR 2021 1arXiv:2012.09841archive 2025-07-28

Patrick Esser, Robin Rombach, Björn Ommer

Designed to learn long-range interactions on sequential data, transformers continue to show state-of-the-art results on a wide variety of tasks. In contrast to CNNs, they contain no inductive bias that prioritizes local interactions. This makes them expressive, but also computationally infeasible for long sequences, such as high-resolution images. We demonstrate how combining the effectiveness of the inductive bias of CNNs with the expressivity of transformers enables them to model and thereby synthesize high-resolution images. We show how to (i) use CNNs to learn a context-rich vocabulary of image constituents, and in turn (ii) utilize transformers to efficiently model their composition within high-resolution images. Our approach is readily applied to conditional synthesis tasks, where both non-spatial information, such as object classes, and spatial information, such as segmentations, can control the generated image. In particular, we present the first results on semantically-guided synthesis of megapixel images with transformers and obtain the state of the art among autoregressive models on class-conditional ImageNet. Code and pretrained models can be found at https://github.com/CompVis/taming-transformers .

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2012.09841")

Code

Syntology Ran 6 of 6 code samples harvested from 2 repositories linked to this paper; 0 have no recorded run. Of those that ran: 1 ran · honoured contract; 1 ran · violated contract; 3 ran · our draft was wrong; 1 ran · fixture could not drive it.

By repository: official repository: 2 samples from 1 repository, 2 ran; community (archive-listed): 3 samples from 1 repository, 3 ran; 1 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

13 repositories listed; official and paper-mentioned ones first.

CompVis/taming-transformers officialmentioned in papermentioned on GitHubpytorch report
YvanG/VQGAN-CLIP mentioned on GitHubpytorch report
dome272/VQGAN mentioned on GitHubpytorch report
dome272/vqgan-pytorch mentioned on GitHubpytorch report
hyn2028/llm-cxr mentioned on GitHubpytorchApache-2.0 report
joanrod/ocr-vqgan mentioned on GitHubpytorch report
joh-fischer/PlantLDM mentioned on GitHubpytorch report
samb-t/unleashing-transformers mentioned on GitHubpytorch report
v-iashin/SpecVQGAN mentioned on GitHubpytorch report
xiaoiker/meta_dpm mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

6 samples harvested; 6 ran; 1 honoured the contract we drafted; 0 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · honoured contract
1ran · violated contract
3ran · our draft was wrong
1ran · fixture could not drive it

Licence: 4 of the 6 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; each sample names its own and says which. Some samples are identical code Syntology first harvested from another repository; for those, this paper's copy is not located and its licence is not recorded. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

bchw_to_st CompVis/taming-transformers/scripts/sample_conditional.py official repository ran · our draft was wrong fingerprinted MIT (permissive) · 6ca974beb2276abf · report
single_image_to_torch CompVis/taming-transformers/scripts/sample_conditional.py official repository ran · fixture could not drive it fingerprinted MIT (permissive) · d2bff574a971c670 · report
adopt_weight joanrod/ocr-vqgan/taming/modules/losses/vqperceptual.py community (archive-listed) ran · honoured contract fingerprinted no licence file found · pointer only · 7439a9d4c6be04b0 · report
hinge_d_loss joanrod/ocr-vqgan/taming/modules/losses/vqperceptual.py community (archive-listed) ran · our draft was wrong fingerprinted no licence file found · pointer only · 22a83df891d750c6 · report
vanilla_d_loss joanrod/ocr-vqgan/taming/modules/losses/vqperceptual.py community (archive-listed) ran · our draft was wrong fingerprinted no licence file found · pointer only · 263c54290c553ddd · report
disabled_train identical code first harvested elsewhere ran · violated contract licence of this copy not recorded · 4cb732f513d69dfd · report

Tasks

DeepFake DetectionImage GenerationImage OutpaintingImage ReconstructionImage-to-Image TranslationInductive BiasText-to-Image GenerationVocal Bursts Intensity Prediction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
DeepFake Detection FakeAVCeleb VQGAN AP 55.0 #9 of 13 Archive leaderboard report
DeepFake Detection FakeAVCeleb VQGAN ROC AUC 51.8 #9 of 13 Archive leaderboard report
Image Generation CelebA 256x256 VQGAN FID 10.2 #13 of 17 Archive leaderboard report
Image Generation CelebA-HQ 256x256 VQGAN+Transformer FID 10.2 #13 of 19 Archive leaderboard report
Image Generation FFHQ 256 x 256 VQGAN+Transformer FID 9.6 #35 of 51 Archive leaderboard report
Image Generation ImageNet 256x256 VQGAN+Transformer (k=600, p=1.0, a=0.05) FID 5.2 #88 of 94 Archive leaderboard report
Image Generation ImageNet 256x256 VQGAN+Transformer (k=mixed, p=1.0, a=0.005) FID 6.59 #90 of 94 Archive leaderboard report
Image Outpainting LHQC Taming Block-FID (Right Extend) 22.53 #4 of 4 Archive leaderboard report
Image Outpainting LHQC Taming Block-FID (Down Extend) 26.38 #4 of 4 Archive leaderboard report
Image Outpainting LHQC Taming Block-FID (Left Extend) - #4 of 4 Archive leaderboard report
Image Outpainting LHQC Taming Block-FID (Up Extend) - #4 of 4 Archive leaderboard report
Image Reconstruction ImageNet Taming-VQGAN (16x16) FID 3.64 #15 of 15 Archive leaderboard report
Image Reconstruction ImageNet Taming-VQGAN (16x16) LPIPS 0.177 #15 of 15 Archive leaderboard report
Image Reconstruction ImageNet Taming-VQGAN (16x16) PSNR 19.93 #15 of 15 Archive leaderboard report
Image Reconstruction ImageNet Taming-VQGAN (16x16) SSIM 0.542 #15 of 15 Archive leaderboard report
Image Reconstruction Ultra-High Resolution Image Reconstruction Benchmark VQGAN (16x16) PSNR 22.91 #5 of 6 Archive leaderboard report
Image Reconstruction Ultra-High Resolution Image Reconstruction Benchmark VQGAN (16x16) rFID 5.95 #5 of 6 Archive leaderboard report
Image-to-Image Translation ADE20K Labels-to-Photos VQGAN+Transformer FID 35.5 #15 of 16 Archive leaderboard report
Image-to-Image Translation COCO-Stuff Labels-to-Photos VQGAN+Transformer FID 22.4 #9 of 15 Archive leaderboard report
Text-to-Image Generation Conceptual Captions VQ-GAN FID 28.86 #5 of 5 Archive leaderboard report
Text-to-Image Generation LHQC Taming Block-FID 38.89 #3 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections