SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.01986 · 2023

Towards Democratizing Joint-Embedding Self-Supervised Learning

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 2 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2303.01986.json · badge: /api/ran/2303.01986.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
copy not recorded community
not in the paper's declared repositories
2 ran · 0 no recorded run
facebookresearch/ffcv-ssl extension (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
exclude_bias_and_norm Ran
ran, violated the contract we drafted
this paper's copy was not recorded; identical code first harvested from facebookresearch/vicreg
pointer only · get_code("28c640db8f721c0d")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

gather_center Ran
ran; our contract draft was wrong, not the code
facebookresearch/ffcv-ssl/examples/train_ssl.py
code served (permissive licence) · get_code("fc2c59422af9d93e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch as ch
import torch.distributed as dist
import os

# Set up a minimal distributed environment for testing
# We'll mock the distributed functions since we can't actually run multi-process

# Mock distributed functions to simulate single-process behavior
class MockDist:
    _world_size = 1
    _rank = 0
    
    @staticmethod
    def get_world_size():
        return MockDist._world_size
    
    @staticmethod
    def get_rank():
        return MockDist._rank
    
    @staticmethod
    def all_gather(output, x):
        # In single process, just copy x to output[0]
        output[0].copy_(x)
    
    @staticmethod
    def all_reduce(tensor):
        # In single process, no-op
        pass
    
    @staticmethod
    def is_initialized():
        return True

# Replace dist functions with mocks
dist.get_world_size = MockDist.get_world_size
dist.get_rank = MockDist.get_rank
dist.all_gather = MockDist.all_gather
dist.all_reduce = MockDist.all_reduce
dist.is_initialized = MockDist.is_initialized

# Set random seed
ch.manual_seed(0)

# Create a small tensor for testing
# batch size 2, feature dimension 4
x = ch.randn(2, 4)

ARGS = {
    'x': x
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
off_diagonal Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from neurai-lab/tamil
pointer only; the stored copy get_code serves, from neurai-lab/tamil, is pointer only (licence: NONE) · get_code("3e30d88eaef01190")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 3 later) and 0 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
off_diagonal 23 other papers
12 dated earlier, 11 later · 5 only as an assembly brick
1908.10084 (2019-08)
Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2007.16189 (2020-07)
Self-supervised learning through the eyes of a child
and 20 more not named here
exclude_bias_and_norm 7 other papers
4 dated earlier, 3 later
2105.04906 (2021-05)
VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning
2110.09348 (2021-10)
Understanding Dimensional Collapse in Contrastive Self-supervised Learning
2112.09164 (2021-12)
High Fidelity Visualization of What Your Self-Supervised Representation Knows About
and 4 more not named here
gather_center 2 other papers
1 dated earlier, 1 later
2210.01571 (2022-10)
VICRegL: Self-Supervised Learning of Local Visual Features
2304.13850 (2023-04)
Do SSL Models Have Déjà Vu? A Case of Unintended Memorization in Self-supervised Learning

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Joint Embedding Self-Supervised Learning (JE-SSL) has seen rapid developments in recent years, due to its promise to effectively leverage large unlabeled data. The development of JE-SSL methods was driven primarily by the search for ever increasing downstream classification accuracies, using huge computational resources, and typically built upon insights and intuitions inherited from a close parent JE-SSL method. This has led unwittingly to numerous pre-conceived ideas that carried over across methods e.g. that SimCLR requires very large mini batches to yield competitive accuracies; that strong and computationally slow data augmentations are required. In this work, we debunk several such ill-formed a priori ideas in the hope to unleash the full potential of JE-SSL free of unnecessary limitations. In fact, when carefully evaluating performances across different downstream tasks and properly optimizing hyper-parameters of the methods, we most often -- if not always -- see that these widespread misconceptions do not hold. For example we show that it is possible to train SimCLR to learn useful representations, while using a single image patch as negative example, and simple Gaussian noise as the only data augmentation for the positive pair. Along these lines, in the hope to democratize JE-SSL and to allow researchers to easily make more extensive evaluations of their methods, we introduce an optimized PyTorch library for SSL.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2303.01986")
get_code_for_paper("2303.01986")
have("2303.01986")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2303.01986.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2303.01986.svg)](https://syntology.ai/paper/2303.01986)

Connect an agent — have() is free.