Papers › Let Go of Your Labels with Unsupervised Transfer

Let Go of Your Labels with Unsupervised Transfer

11 Jun 2024International Conference on Machine Learning 2024 6arXiv:2406.07236archive 2025-07-28

Artyom Gadetsky, Yulun Jiang, Maria Brbic

Foundation vision-language models have enabled remarkable zero-shot transferability of the pre-trained representations to a wide range of downstream tasks. However, to solve a new task, zero-shot transfer still necessitates human guidance to define visual categories that appear in the data. Here, we show that fully unsupervised transfer emerges when searching for the labeling of a dataset that induces maximal margin classifiers in representation spaces of different foundation models. We present TURTLE, a fully unsupervised method that effectively employs this guiding principle to uncover the underlying labeling of a downstream dataset without any supervision and task-specific representation learning. We evaluate TURTLE on a diverse benchmark suite of 26 datasets and show that it achieves new state-of-the-art unsupervised performance. Furthermore, TURTLE, although being fully unsupervised, outperforms zero-shot transfer baselines on a wide range of datasets. In particular, TURTLE matches the average performance of CLIP zero-shot on 26 datasets by employing the same representation space, spanning a wide range of architectures and model sizes. By guiding the search for the underlying labeling using the representation spaces of two foundation models, TURTLE surpasses zero-shot transfer and unsupervised prompt tuning baselines, demonstrating the surprising power and effectiveness of unsupervised transfer.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 4 code samples harvested from 1 repository linked to this paper: 3 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 4 samples from 1 repository, 3 ran (1 with no instrument failure; 2 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 3 samples that ran has no recorded date.

mlbio-epfl/turtle officialpytorchSyntology: 4 samples, 3 ran (2 instrument-failed) · 1 unverified 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

4 samples harvested: 3 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
2ran · our draft was wrong
1unverified

Licence: 4 of the 4 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 mlbio-epfl/turtle. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

mlbio-epfl/turtle official repository 4 samples, 3 ran (2 instrument-failed) · 1 unverified
_parse_args run_turtle.py ran · our draft was wrong no licence file found · pointer only · 806241343c4b7e7b · report fixture: signature args:list/1/str · returned Namespace · level recorded 2026-09-22
get_cluster_acc run_turtle.py ran · our draft was wrong no licence file found · pointer only · 89923b3abde90046 · report fixture: signature y_pred:ndarray/1/int, y_true:ndarray/1/int, return_matching:bool/0/bool · returned tuple · level recorded 2026-09-22
get_features in 57 papers precompute_representations.py ran · honoured contract no licence file found · pointer only · b300ef9fae8f3a37 · report fixture: signature dataloader:DataLoader/0/?, model:callable/0/?, device:device/0/? · returned ndarray · level recorded, date not recorded
run in 143 papers run_turtle.py unverified no licence file found · pointer only · b22aeb3079826cd8 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for TURTLE (graph method: TURTLE). 4 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

  1. run entry · level 0 · unverified · b22aeb3079826cd8
  2. _parse_args helper · level 1 · ran · our draft was wrong · 806241343c4b7e7b
  3. get_cluster_acc metric · level 1 · ran · our draft was wrong · 89923b3abde90046
  4. seed_everything helper · level 0 · unverified · 6a85cbb1a7488a1b

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2406.07236") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Image ClusteringUnsupervised Image Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Clustering Birdsnap TURTLE (CLIP + DINOv2) Accuracy 68.1 #1 of 1 Archive leaderboard report
Image Clustering CIFAR-10 TURTLE (CLIP + DINOv2) ARI 0.989 #1 of 40 Archive leaderboard report
Image Clustering CIFAR-10 TURTLE (CLIP + DINOv2) Accuracy 0.995 #1 of 40 Archive leaderboard report
Image Clustering CIFAR-10 TURTLE (CLIP + DINOv2) NMI 0.985 #1 of 40 Archive leaderboard report
Image Clustering CIFAR-100 TURTLE (CLIP + DINOv2) ARI 0.834 #1 of 30 Archive leaderboard report
Image Clustering CIFAR-100 TURTLE (CLIP + DINOv2) Accuracy 0.898 #1 of 30 Archive leaderboard report
Image Clustering CIFAR-100 TURTLE (CLIP + DINOv2) NMI 0.915 #1 of 30 Archive leaderboard report
Image Clustering CLEVR Counts TURTLE (CLIP + DINOv2) Accuracy 24.0 #1 of 1 Archive leaderboard report
Image Clustering Caltech-101 TURTLE (CLIP + DINOv2) Accuracy 89.8 #1 of 1 Archive leaderboard report
Image Clustering Country211 TURTLE (CLIP + DINOv2) Accuracy 11.1 #1 of 1 Archive leaderboard report
Image Clustering DTD TURTLE (CLIP + DINOv2) Accuracy 57.3 #2 of 2 Archive leaderboard report
Image Clustering EuroSAT TURTLE (CLIP + DINOv2) Accuracy 96.6 #1 of 1 Archive leaderboard report
Image Clustering FER2013 TURTLE (CLIP + DINOv2) Accuracy 36.2 #1 of 1 Archive leaderboard report
Image Clustering FGVC Aircraft TURTLE (CLIP + DINOv2) Accuracy 36.5 #1 of 1 Archive leaderboard report
Image Clustering Flowers-102 TURTLE (CLIP + DINOv2) Accuracy 99.6 #1 of 1 Archive leaderboard report
Image Clustering Food-101 TURTLE (CLIP + DINOv2) Accuracy 92.2 #1 of 1 Archive leaderboard report
Image Clustering GTSRB TURTLE (CLIP + DINOv2) Accuracy 48.4 #1 of 1 Archive leaderboard report
Image Clustering Hateful Memes TURTLE (CLIP + DINOv2) Accuracy 54.2 #1 of 1 Archive leaderboard report
Image Clustering ImageNet TURTLE (CLIP + DINOv2) ARI 62.5 #1 of 12 Archive leaderboard report
Image Clustering ImageNet TURTLE (CLIP + DINOv2) Accuracy 72.9 #1 of 12 Archive leaderboard report
Image Clustering ImageNet TURTLE (CLIP + DINOv2) NMI 88.2 #1 of 12 Archive leaderboard report
Image Clustering KITTI TURTLE (CLIP + DINOv2) Accuracy 39.4 #1 of 1 Archive leaderboard report
Image Clustering Kinetics-700 TURTLE (CLIP + DINOv2) Accuracy 43.0 #1 of 1 Archive leaderboard report
Image Clustering MNIST TURTLE (CLIP + DINOv2) Accuracy 97.8 #1 of 3 Archive leaderboard report
Image Clustering Oxford-IIIT Pets TURTLE (CLIP + DINOv2) Accuracy 92.3 #1 of 1 Archive leaderboard report
Image Clustering PCam TURTLE (CLIP + DINOv2) Accuracy 52.0 #1 of 1 Archive leaderboard report
Image Clustering RESISC45 TURTLE (CLIP + DINOv2) Accuracy 89.6 #1 of 1 Archive leaderboard report
Image Clustering Rendered SST2 TURTLE (CLIP + DINOv2) Accuracy 51.6 #1 of 1 Archive leaderboard report
Image Clustering STL-10 TURTLE (CLIP + DINOv2) ARI 0.994 #1 of 29 Archive leaderboard report
Image Clustering STL-10 TURTLE (CLIP + DINOv2) Accuracy 0.997 #1 of 29 Archive leaderboard report
Image Clustering STL-10 TURTLE (CLIP + DINOv2) NMI 0.993 #1 of 29 Archive leaderboard report
Image Clustering SUN397 TURTLE (CLIP + DINOv2) Accuracy 67.9 #1 of 1 Archive leaderboard report
Image Clustering Stanford Cars TURTLE (CLIP + DINOv2) Accuracy 0.646 #1 of 5 Archive leaderboard report
Image Clustering UCF101 TURTLE (CLIP + DINOv2) Accuracy 82.3 #1 of 2 Archive leaderboard report
Unsupervised Image Classification CIFAR-10 TURTLE (CLIP + DINOv2) Accuracy 99.5 #1 of 9 Archive leaderboard report
Unsupervised Image Classification ImageNet TURTLE (CLIP + DINOv2) ARI 62.5 #1 of 9 Archive leaderboard report
Unsupervised Image Classification ImageNet TURTLE (CLIP + DINOv2) Accuracy (%) 72.9 #1 of 9 Archive leaderboard report
Unsupervised Image Classification MNIST TURTLE (CLIP + DINOv2) Accuracy 97.8 #3 of 10 Archive leaderboard report
Unsupervised Image Classification STL-10 TURTLE (CLIP + DINOv2) Accuracy 99.7 #1 of 9 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

CLIP

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