SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1603.09246 · 2016

Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles

Paolo Favaro, Mehdi Noroozi

arXiv · PDF · Open in the Atlas

Code that ran

10 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 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; 3 failures of our instrument; 3 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/1603.09246.json · badge: /api/ran/1603.09246.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
facebookresearch/vissl community
Papers with Code archive
5 ran · 1 no recorded run
SharadGitHub/Self-Supervised-Autoencoder community
Papers with Code archive
2 ran · 0 no recorded run
bbrattoli/JigsawPuzzlePytorch community
Papers with Code archive
1 ran · 1 no recorded run
Confusezius/selfsupervised_learning community
Papers with Code archive
1 ran · 0 no recorded run
virtualgraham/sc_patch community
Papers with Code archive
1 ran · 0 no recorded run
kalelpark/FG-SSL community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
AlexNet Ran
ran on a synthesized input · constructed an object; no result was computed
Confusezius/selfsupervised_learning/JigsawNet/baseline_conv.py
pointer only (licence: NONE) · get_code("c8988caab7b2ad80")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Confusezius/selfsupervised_learning when it was checked. 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.

AttrDict Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("6663d33830e09bae")
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 an instance of its own class, AttrDict: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type 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
from copy import deepcopy

# The AttrDict class takes a single parameter: dictionary
# Let's create a comprehensive test dictionary with nested structures

test_dictionary = {
    "model_name": "test_model",
    "learning_rate": 0.001,
    "batch_size": 32,
    "epochs": 10,
    "nested_config": {
        "layer1": {
            "units": 64,
            "activation": "relu"
        },
        "layer2": {
            "units": 128,
            "activation": "sigmoid"
        }
    },
    "data_path": "/path/to/data",
    "seed": 42,
    "parameters": {
        "dropout": 0.5,
        "regularization": {
            "l1": 0.001,
            "l2": 0.002
        }
    }
}

ARGS = {
    "dictionary": test_dictionary
}
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 AttrDict — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a AttrDict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
Classifier Ran
ran on a synthesized input · constructed an object; no result was computed
SharadGitHub/Self-Supervised-Autoencoder/Jigsaw Task/skeleton/src/segnet.py
pointer only (licence: NONE) · get_code("80068288a687a198")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, SharadGitHub/Self-Supervised-Autoencoder when it was checked. 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.

Flatten Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("0b750ebcd399df8e")
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 an instance of its own class, Flatten: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type 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 torch.nn as nn

torch.manual_seed(0)

# The Flatten module's __init__ takes one parameter: dim
# The forward method takes feat (a tensor) and flattens it starting from self.dim
# The flops method takes x (unused) and returns 0

# Create a sample tensor to understand the expected input
# feat should be a multi-dimensional tensor
sample_feat = torch.randn(2, 4, 3, 5)  # batch_size=2, with additional dimensions

ARGS = {
    "dim": -1
}
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 Flatten — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Flatten output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
LRN Ran
ran on a synthesized input
bbrattoli/JigsawPuzzlePytorch/JigsawNetwork.py
pointer only (licence: NONE) · get_code("6b7aa1cb23b2f8e5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, bbrattoli/JigsawPuzzlePytorch when it was checked. 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.

Separable Ran
ran on a synthesized input · constructed an object; no result was computed
SharadGitHub/Self-Supervised-Autoencoder/Jigsaw Task/skeleton/src/segnet.py
pointer only (licence: NONE) · get_code("49b29ff2cd6ea499")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, SharadGitHub/Self-Supervised-Autoencoder when it was checked. 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.

VggNetwork Ran
ran on a synthesized input · constructed an object; no result was computed
virtualgraham/sc_patch/src/nine_patch/nine_patch_a.py
pointer only (licence: NONE) · get_code("4fe485f49849d35b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, virtualgraham/sc_patch when it was checked. 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_trunk_forward_outputs_module_list Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("b9b9733099ed93fd")
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. The image it ran in was not recorded. If your copy returns the same type 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 torch.nn as nn

torch.manual_seed(0)

# Create a simple feature block that processes input
class SimpleBlock(nn.Module):
    def __init__(self, in_dim, out_dim):
        super().__init__()
        self.linear = nn.Linear(in_dim, out_dim)
    
    def forward(self, x):
        return self.linear(x)

# Build feature blocks - each transforms the tensor
feature_blocks = nn.ModuleList([
    SimpleBlock(4, 5),
    SimpleBlock(5, 6),
    SimpleBlock(6, 7),
    SimpleBlock(7, 8),
])

# Define feature names corresponding to each block
all_feat_names = ["layer1", "layer2", "layer3", "layer4"]

# Create input tensor: batch_size=2, input_dim=4
feat = torch.randn(2, 4)

# Define which features we want to extract
out_feat_keys = ["layer2", "layer4"]

ARGS = {
    "feat": feat,
    "out_feat_keys": out_feat_keys,
    "feature_blocks": feature_blocks,
    "all_feat_names": all_feat_names,
}
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 — a list output has no array shape of its own; none is stated
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
parse_out_keys_arg Ran
ran; our contract draft was wrong, not the code
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("fa8d202b59c77f16")
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 tuple. The image it ran in was not recorded. If your copy returns the same type 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 torch.nn as nn

torch.manual_seed(0)

# The function takes:
# 1. out_feat_keys: List[str] - a list of feature key names to extract
# 2. all_feat_names: List[str] - a list of all available feature names in the model

# The function checks that all requested keys exist in all_feat_names
# and returns the keys plus the index of the maximum feature to extract

# Create valid test arguments
out_feat_keys = ["layer2", "layer4"]
all_feat_names = ["layer1", "layer2", "layer3", "layer4", "layer5"]

ARGS = {
    "out_feat_keys": out_feat_keys,
    "all_feat_names": all_feat_names,
}
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 tuple — a tuple output has no array shape of its own; none is stated
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
register_model_trunk Ran
ran; our contract draft was wrong, not the code
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("4e819ea98b2a0bc8")
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 function. The image it ran in was not recorded. If your copy returns the same type 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 torch.nn as nn
from typing import Callable

# Clear the global registries before setting up test arguments
MODEL_TRUNKS_NAMES = set()
MODEL_TRUNKS_REGISTRY = {}

# Create a simple model trunk class to use as test input
class SimpleModelTrunk(nn.Module):
    """A minimal model trunk for testing."""
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# The function parameter is 'name', which is a string
ARGS = {
    'name': 'test_model_trunk'
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
AlexNetJigsaw Not yet run
unverified (harvested, no recorded run)
facebookresearch/vissl/vissl/models/trunks/alexnet_jigsaw.py
code served (permissive licence) · get_code("3eb3d5e042c389a9")
How it ran

no invocation recorded.

Network Not yet run
unverified (harvested, no recorded run)
bbrattoli/JigsawPuzzlePytorch/JigsawNetwork.py
pointer only (licence: NONE) · get_code("a8e0685e21eae7d2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, bbrattoli/JigsawPuzzlePytorch when it was checked. 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.

jigsaw_train Not yet run
unverified (harvested, no recorded run)
kalelpark/FG-SSL/learner/pretrain.py
pointer only (licence: NONE) · get_code("10ab7c87e639cc4b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kalelpark/FG-SSL when it was checked. 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 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 5 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 5 later) and 8 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
AttrDict
brick of this paper's assembly: helper
2 other papers
2 later · 2 only as an assembly brick
1803.07728 (2018-03)
Unsupervised Representation Learning by Predicting Image Rotations
2106.08254 (2021-06)
BEIT: BERT Pre-Training of Image Transformers
register_model_trunk
brick of this paper's assembly: helper
2 other papers
2 later · 2 only as an assembly brick
1803.07728 (2018-03)
Unsupervised Representation Learning by Predicting Image Rotations
2106.08254 (2021-06)
BEIT: BERT Pre-Training of Image Transformers
Flatten
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
1803.07728 (2018-03)
Unsupervised Representation Learning by Predicting Image Rotations
get_trunk_forward_outputs_module_list
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
1803.07728 (2018-03)
Unsupervised Representation Learning by Predicting Image Rotations
parse_out_keys_arg
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
1803.07728 (2018-03)
Unsupervised Representation Learning by Predicting Image Rotations

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

In this paper we study the problem of image representation learning without human annotation. By following the principles of selfsupervision, we build a convolutional neural network (CNN) that can be trained to solve Jigsaw puzzles as a pretext task, which requires no manual labeling, and then later repurposed to solve object classification and detection. To maintain the compatibility across tasks we introduce the context-free network (CFN), a siamese-ennead CNN. The CFN takes image tiles as input and explicitly limits the receptive field (or context) of its early processing units to one tile at a time. We show that the CFN includes fewer parameters than AlexNet while preserving the same semantic learning capabilities. By training the CFN to solve Jigsaw puzzles, we learn both a feature mapping of object parts as well as their correct spatial arrangement. Our experimental evaluations show that the learned features capture semantically relevant content. Our proposed method for learning visual representations outperforms state of the art methods in several transfer learning benchmarks.

For agents

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

get_harvested_code_for_paper("1603.09246")
get_code_for_paper("1603.09246")
have("1603.09246")

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

curl https://syntology.ai/api/ran/1603.09246.json

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

[![Syntology run record](https://syntology.ai/api/ran/1603.09246.svg)](https://syntology.ai/paper/1603.09246)

Connect an agent — have() is free.