SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1901.09005 · 2019

Revisiting Self-Supervised Visual Representation Learning

arXiv · PDF · Open in the Atlas

Code that ran

9 of 15 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, 0 violated it and 7 have no contract verdict; 2 failures of our instrument; 6 no recorded run), measured 2026-11-21; 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/1901.09005.json · badge: /api/ran/1901.09005.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
google/revisiting-self-supervised canonical (classifier label)
classifier basis: census
5 ran · 3 no recorded run
rickyHong/Puzzle-tensorflow-latest-repl community (classifier label)
classifier basis not recorded on the link
4 ran · 3 no recorded run
FunctionStatusWhere it lives
adaptive_pool Ran
ran on a synthesized input
rickyHong/Puzzle-tensorflow-latest-repl/utils.py
pointer only (licence: Apache-2.0) · get_code("a758b93f9c0566d4")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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.

concat_dicts Ran
ran on a synthesized input
rickyHong/Puzzle-tensorflow-latest-repl/utils.py
pointer only (licence: Apache-2.0) · get_code("4fdb7cc6673fd054")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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.

fixed_padding Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
google/revisiting-self-supervised/models/resnet.py
code served (permissive licence) · get_code("650804f0bf60d727")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 4, dtype family float, shape (2, 6, 5, 1) 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 tensorflow as tf
import numpy as np

# Create a small tensor for testing
# Shape: [batch=2, height=4, width=3, channels=1]
x = tf.constant(np.random.randn(2, 4, 3, 1).astype(np.float32))

# kernel_size parameter
kernel_size = 3

ARGS = {
    "x": x,
    "kernel_size": kernel_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 6, 5, 1) — shape (2, 6, 5, 1) 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
generate_sharded_filenames Ran
ran on a synthesized input
google/revisiting-self-supervised/datasets.py
code served (permissive licence) · get_code("67c0ff3ae27aa20b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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

ARGS = {
    "filename": "output@5"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
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
get_resize_small Ran
ran; our contract draft was wrong, not the code
rickyHong/Puzzle-tensorflow-latest-repl/preprocess.py
pointer only (licence: Apache-2.0) · get_code("f2ead8be8bb7d21c")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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.

get_shape_as_list Ran
ran on a synthesized input
google/revisiting-self-supervised/models/resnet.py
code served (permissive licence) · get_code("f3176d55b81052d5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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
import tensorflow as tf

torch.manual_seed(0)

# Create a TensorFlow tensor with a known shape
x = tf.constant([[1.0, 2.0, 3.0, 4.0],
                 [5.0, 6.0, 7.0, 8.0]])

ARGS = {
    'x': x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — a list output has no array shape of its own; none is stated
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
image_grid Ran
ran on a synthesized input
google/revisiting-self-supervised/self_supervision/patch_utils.py
code served (permissive licence) · get_code("8f93014aa9fd5d3f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 4, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 tensorflow as tf
import numpy as np

# Create a small batch of images
# Shape: (B, N, H, W, C) where:
# B = batch size = 2
# N = number of patches = 6 (for a 2x3 grid)
# H = height = 4
# W = width = 3
# C = channels = 1

batch_size = 2
num_patches = 6
height = 4
width = 3
channels = 1

# Create dummy image data
images_np = np.random.randn(batch_size, num_patches, height, width, channels).astype(np.float32)
images = tf.constant(images_np)

ny = 2  # vertical number of images
nx = 3  # horizontal number of images
padding = 0

ARGS = {
    "images": images,
    "ny": ny,
    "nx": nx,
    "padding": padding,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
repeat Ran
ran on a synthesized input
google/revisiting-self-supervised/self_supervision/exemplar.py
code served (permissive licence) · get_code("0e9af15ef7691776")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 1, dtype family int, shape (12,) 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 tensorflow as tf
import numpy as np

# Create a small TensorFlow tensor to pass as x
x = tf.constant([1, 2, 3, 4], dtype=tf.int32)

# Create a times value (scalar integer)
times = 3

# Construct the ARGS dictionary
ARGS = {
    "x": x,
    "times": times
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 1, dtype family int, shape (12,) — shape (12,) 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
str2intlist Ran
ran on a synthesized input
rickyHong/Puzzle-tensorflow-latest-repl/utils.py
pointer only (licence: Apache-2.0) · get_code("420d7fad13c47138")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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.

batch_norm Not yet run
unverified (harvested, no recorded run)
google/revisiting-self-supervised/models/resnet.py
code served (permissive licence) · get_code("d29f9171cd72ee14")
How it ran

no invocation recorded.

convbnrelu Not yet run
unverified (harvested, no recorded run)
rickyHong/Puzzle-tensorflow-latest-repl/models/vggnet.py
pointer only (licence: Apache-2.0) · get_code("1b8fc1e32b1ff09f")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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. No invocation recorded.

crop Not yet run
unverified (harvested, no recorded run)
google/revisiting-self-supervised/self_supervision/patch_model_preprocess.py
code served (permissive licence) · get_code("8fc347f41c9ad9e5")
How it ran

no invocation recorded.

get_lr Not yet run
unverified (harvested, no recorded run)
rickyHong/Puzzle-tensorflow-latest-repl/trainer.py
pointer only (licence: Apache-2.0) · get_code("baf26176db8e0350")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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. No invocation recorded.

image_to_patches Not yet run
unverified (harvested, no recorded run)
google/revisiting-self-supervised/self_supervision/patch_model_preprocess.py
code served (permissive licence) · get_code("d6adc3a769d03202")
How it ran

no invocation recorded.

vgg19 Not yet run
unverified (harvested, no recorded run)
rickyHong/Puzzle-tensorflow-latest-repl/models/vggnet.py
pointer only (licence: Apache-2.0) · get_code("5f17db2cd3b2b06d")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, rickyHong/Puzzle-tensorflow-latest-repl, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with 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. No invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 6 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 6 later) and 9 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
batch_norm 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-Supervised Learning
crop 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-Supervised Learning
fixed_padding 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-Supervised Learning
generate_sharded_filenames 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-Supervised Learning
get_resize_small 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-Supervised Learning
get_shape_as_list 1 other paper
1 later
1905.03670 (2019-05)
S4L: Self-Supervised Semi-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 2026-11-21.

Repositories linked to this paper

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

Abstract

Unsupervised visual representation learning remains a largely unsolved problem in computer vision research. Among a big body of recently proposed approaches for unsupervised learning of visual representations, a class of self-supervised techniques achieves superior performance on many challenging benchmarks. A large number of the pretext tasks for self-supervised learning have been studied, but other important aspects, such as the choice of convolutional neural networks (CNN), has not received equal attention. Therefore, we revisit numerous previously proposed self-supervised models, conduct a thorough large scale study and, as a result, uncover multiple crucial insights. We challenge a number of common practices in selfsupervised visual representation learning and observe that standard recipes for CNN design do not always translate to self-supervised representation learning. As part of our study, we drastically boost the performance of previously proposed techniques and outperform previously published state-of-the-art results by a large margin.

For agents

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

get_harvested_code_for_paper("1901.09005")
get_code_for_paper("1901.09005")
have("1901.09005")

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

curl https://syntology.ai/api/ran/1901.09005.json

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

[![Syntology run record](https://syntology.ai/api/ran/1901.09005.svg)](https://syntology.ai/paper/1901.09005)

Connect an agent — have() is free.