Papers › Ensembling Off-the-shelf Models for GAN Training

Ensembling Off-the-shelf Models for GAN Training

16 Dec 2021CVPR 2022 1arXiv:2112.09130archive 2025-07-28

Nupur Kumari, Richard Zhang, Eli Shechtman, Jun-Yan Zhu

The advent of large-scale training has produced a cornucopia of powerful visual recognition models. However, generative models, such as GANs, have traditionally been trained from scratch in an unsupervised manner. Can the collective "knowledge" from a large bank of pretrained vision models be leveraged to improve GAN training? If so, with so many models to choose from, which one(s) should be selected, and in what manner are they most effective? We find that pretrained computer vision models can significantly improve performance when used in an ensemble of discriminators. Notably, the particular subset of selected models greatly affects performance. We propose an effective selection mechanism, by probing the linear separability between real and fake samples in pretrained model embeddings, choosing the most accurate model, and progressively adding it to the discriminator ensemble. Interestingly, our method can improve GAN training in both limited data and large-scale settings. Given only 10k training samples, our FID on LSUN Cat matches the StyleGAN2 trained on 1.6M images. On the full dataset, our method improves FID by 1.5x to 2x on cat, church, and horse categories of LSUN.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

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="2112.09130")

Code

Syntology Ran 12 of 15 code samples harvested from 1 repository linked to this paper; 3 have no recorded run. Of those that ran: 5 ran · our draft was wrong; 1 ran · fixture could not drive it; 6 ran with no contract checked.

By repository: official repository: 15 samples from 1 repository, 12 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

nupurkmr9/vision-aided-gan officialmentioned in papermentioned on GitHubpytorchMIT 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

15 samples harvested; 12 ran; 0 honoured the contract we drafted; 3 have no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

5ran · our draft was wrong
1ran · fixture could not drive it
6ran
3unverified

Licence: 0 of the 15 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 nupurkmr9/vision-aided-gan. “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.

window_partition nupurkmr9/vision-aided-gan/vision_aided_loss/swintaskspecific.py official repository ran · fixture could not drive it fingerprinted MIT (permissive) · 144d10b49baeb8a6 · report
conv3x3 nupurkmr9/vision-aided-gan/vision_aided_loss/resnet.py official repository ran · our draft was wrong MIT (permissive) · fac5364e2f53c6db · report
convrelu nupurkmr9/vision-aided-gan/vision_aided_loss/face_normals.py official repository ran MIT (permissive) · bb6429dd2e9f16d3 · report
get_interp_size nupurkmr9/vision-aided-gan/vision_aided_loss/face_parsing.py official repository ran MIT (permissive) · 80a82d5149279cdf · report
get_pad_layer nupurkmr9/vision-aided-gan/vision_aided_loss/blurpool.py official repository ran · our draft was wrong MIT (permissive) · 50aa4a59fe718c61 · report
get_pad_layer_1d nupurkmr9/vision-aided-gan/vision_aided_loss/blurpool.py official repository ran MIT (permissive) · 176a539246e68879 · report
interp nupurkmr9/vision-aided-gan/vision_aided_loss/face_parsing.py official repository ran MIT (permissive) · 3e87c5d93ee4df83 · report
rand_brightness nupurkmr9/vision-aided-gan/vision_aided_loss/DiffAugment_pytorch.py official repository ran · our draft was wrong fingerprinted MIT (permissive) · 011230b2b9b8fb6f · report
rand_saturation nupurkmr9/vision-aided-gan/vision_aided_loss/DiffAugment_pytorch.py official repository ran · our draft was wrong fingerprinted MIT (permissive) · 5b0d8787e670fc63 · report
resnet18 nupurkmr9/vision-aided-gan/vision_aided_loss/resnet.py official repository ran MIT (permissive) · 46b2f3cc882fb768 · report
resnet34 nupurkmr9/vision-aided-gan/vision_aided_loss/resnet.py official repository ran MIT (permissive) · edd87b551d6f38c9 · report
window_reverse nupurkmr9/vision-aided-gan/vision_aided_loss/swintaskspecific.py official repository ran · our draft was wrong MIT (permissive) · 61bf152e6a42a184 · report
DiffAugment nupurkmr9/vision-aided-gan/vision_aided_loss/DiffAugment_pytorch.py official repository unverified MIT (permissive) · 1d566d068b0ed0af · report
get_upsampling_weight nupurkmr9/vision-aided-gan/vision_aided_loss/face_parsing.py official repository unverified MIT (permissive) · e475794319393649 · report
open_url nupurkmr9/vision-aided-gan/vision_aided_loss/utils_.py official repository unverified MIT (permissive) · 4d45510ba267046e · report

Tasks

Image Generation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Generation AFHQ Cat Vision-aided GAN FID 2.44 #1 of 8 Archive leaderboard report
Image Generation AFHQ Cat Vision-aided GAN clean-FID 2.51 ± .02 #1 of 8 Archive leaderboard report
Image Generation AFHQ Cat Vision-aided GAN clean-KID 0.46 ± .03 #1 of 8 Archive leaderboard report
Image Generation AFHQ Dog Vision-aided GAN FID 4.60 #2 of 6 Archive leaderboard report
Image Generation AFHQ Dog Vision-aided GAN clean-FID 4.73 ± .02 #2 of 6 Archive leaderboard report
Image Generation AFHQ Dog Vision-aided GAN clean-KID 0.38 ± .01 #2 of 6 Archive leaderboard report
Image Generation AFHQ Wild Vision-aided GAN FID 2.25 #1 of 5 Archive leaderboard report
Image Generation AFHQ Wild Vision-aided GAN clean-FID 2.35 ± .02 #1 of 5 Archive leaderboard report
Image Generation AFHQ Wild Vision-aided GAN clean-KID 0.38 ± .02 #1 of 5 Archive leaderboard report
Image Generation LSUN Cat 256 x 256 Vision-aided GAN Clean-FID (trainfull) 3.98 ± 0.03 #1 of 8 Archive leaderboard report
Image Generation LSUN Cat 256 x 256 Vision-aided GAN FID 3.87 #1 of 8 Archive leaderboard report
Image Generation LSUN Churches 256 x 256 Vision-aided GAN Clean-FID (trainfull) 1.72 ± 0.01 #2 of 27 Archive leaderboard report
Image Generation LSUN Churches 256 x 256 Vision-aided GAN FID 1.72 #2 of 27 Archive leaderboard report
Image Generation LSUN Horse 256 x 256 Vision-aided GAN Clean-FID (trainfull) 2.11 #1 of 6 Archive leaderboard report
Image Generation LSUN Horse 256 x 256 Vision-aided GAN FID 2.15 #1 of 6 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

Introduced by this paper: Vision-aided GAN

ConvolutionPath Length RegularizationR1 RegularizationVision-aided GANWeight Demodulation

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