SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2001.07685 · NeurIPS · 2020

FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence

Colin Raffel, Kihyuk Sohn, David Berthelot, Chun-Liang Li, Zizhao Zhang, Nicholas Carlini, Ekin Cubuk, Alex Kurakin, Han Zhang

arXiv · PDF · Open in the Atlas

Code that ran

53 of 74 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (13 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 2 violated it and 20 have no contract verdict; 28 failures of our instrument; 21 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, 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/2001.07685.json · badge: /api/ran/2001.07685.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
kekmodel/FixMatch-pytorch reimplementation (classifier label)
classifier basis: classifier_2025
15 ran · 4 no recorded run
CoinCheung/fixmatch-pytorch reimplementation (classifier label)
classifier basis: classifier_2025
9 ran · 1 no recorded run
S-HuaBomb/FixMatch-Paddle community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
gomezzz/distmsmatch community
Papers with Code archive
3 ran · 1 no recorded run
valencebond/FixMatch_pytorch reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
CoinCheung/fixmatch community
Papers with Code archive
3 ran · 0 no recorded run
SaraAmd/Semi-Supervised-Learning community
Papers with Code archive
3 ran · 0 no recorded run
gomezzz/MSMatch community
Papers with Code archive
2 ran · 1 no recorded run
amazon-research/exponential-moving-average-normalization community
Papers with Code archive
2 ran · 0 no recorded run
LeeDoYup/FixMatch-pytorch reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
miquelmarti/DenseFixMatch community
Papers with Code archive
1 ran · 4 no recorded run
johnchenresearch/SSL community (classifier label)
classifier basis not recorded on the link
1 ran · 4 no recorded run
ptrckhmmr/learning-to-defer-with-limited-expert-predictions community
Papers with Code archive
1 ran · 2 no recorded run
AhmadQasim/FixMatch community
Papers with Code archive
1 ran · 1 no recorded run
OFRIN/Tensorflow_FixMatch community
Papers with Code archive
1 ran · 0 no recorded run
fbuchert/fixmatch-pytorch community
Papers with Code archive
1 ran · 0 no recorded run
Zhongying-Deng/NorMatch community
Papers with Code archive
0 ran · 2 no recorded run
Celiali/FixMatch community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
mish Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/models/wideresnet.py
code served (permissive licence) · get_code("48dececd48e901d2")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 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. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T23:37:00Z). 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
import torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

# The mish function takes a single parameter 'x' which should be a tensor
# It applies: x * tanh(softplus(x))
# So x needs to be a tensor that supports these operations

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 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
AutoContrast Ran
ran; our contract draft was wrong, not the code
S-HuaBomb/FixMatch-Paddle/dataset/randaugment.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from MaxH1996/FedXDS, is pointer only (licence: NONE) · get_code("c6e7c3451c85404f")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from MaxH1996/FedXDS, and GitHub detected no licence in that repository 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.

AverageMeter Ran
ran on a synthesized input · constructed an object; no result was computed
ptrckhmmr/learning-to-defer-with-limited-expert-predictions/Semi-Supervised/Train_FixMatch.py
pointer only (licence: NONE) · get_code("2c90584bcf844d28")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ptrckhmmr/learning-to-defer-with-limited-expert-predictions 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.

BasicBlockPreAct Ran
ran on a synthesized input · constructed an object; no result was computed
valencebond/FixMatch_pytorch/models/model.py
code served (permissive licence) · get_code("735c6ea4d3188cf3")
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, BasicBlockPreAct: 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
from torch.nn import BatchNorm2d

torch.manual_seed(0)

ARGS = {
    'in_chan': 3,
    'out_chan': 16,
    'drop_rate': 0.1,
    'stride': 1,
    'pre_res_act': False,
}
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 BasicBlockPreAct — 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 BasicBlockPreAct 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
BasicBlockPreAct Ran
ran on a synthesized input · constructed an object; no result was computed
CoinCheung/fixmatch/model.py
code served (permissive licence) · get_code("8d3a0f01f7775c34")
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, BasicBlockPreAct: 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
from torch.nn import BatchNorm2d

torch.manual_seed(0)

ARGS = {
    'in_chan': 3,
    'out_chan': 16,
    'drop_rate': 0.1,
    'stride': 1,
    'pre_res_act': False,
}
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 BasicBlockPreAct — 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 BasicBlockPreAct 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
Brightness Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("f82e12f0d695e01b")
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 Image. 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 PIL
from PIL import Image
import numpy as np

# Create a small test image
img_array = np.random.randint(0, 256, (64, 64, 3), dtype=np.uint8)
test_img = Image.fromarray(img_array)

ARGS = {
    "img": test_img,
    "v": 5,
    "max_v": 2.0,
    "bias": 0.1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
Brightness Ran
ran; our contract draft was wrong, not the code
S-HuaBomb/FixMatch-Paddle/dataset/randaugment.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from MaxH1996/FedXDS, is pointer only (licence: NONE) · get_code("0d1a8d6ce2015fc7")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from MaxH1996/FedXDS, and GitHub detected no licence in that repository 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.

Color Ran
ran; our contract draft was wrong, not the code
S-HuaBomb/FixMatch-Paddle/dataset/randaugment.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from MaxH1996/FedXDS, is pointer only (licence: NONE) · get_code("32cb3f5760fa593e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from MaxH1996/FedXDS, and GitHub detected no licence in that repository 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.

ConsistencyRegularization Ran
ran on a synthesized input · constructed an object; no result was computed
SaraAmd/Semi-Supervised-Learning/ssl_lib/algs/consistency.py
pointer only (licence: NONE) · get_code("3476a989bf303448")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, SaraAmd/Semi-Supervised-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.

Contrast Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("3a6db236d683df14")
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 Image. 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 PIL.Image
import PIL.ImageEnhance

# Create a small test image using PIL
# The function expects a PIL Image object
img = PIL.Image.new('RGB', (10, 10), color='red')

ARGS = {
    'img': img,
    'v': 5,
    'max_v': 2.0,
    'bias': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
Equalize Ran
ran on a synthesized input
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("b6210b47d844e910")
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 Image. 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 PIL
from PIL import Image
import numpy as np

# Create a small PIL Image for testing
# The function uses PIL.ImageOps.equalize which works on PIL Images
img_array = np.random.randint(0, 256, (4, 3), dtype=np.uint8)
img = Image.fromarray(img_array, mode='L')

ARGS = {
    'img': img
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 2 and dtype family int 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
FixMatch Ran
ran on a synthesized input · constructed an object; no result was computed
amazon-research/exponential-moving-average-normalization/models/fixmatch.py
pointer only (licence: NONE) · get_code("f97994492ca55ef4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, amazon-research/exponential-moving-average-normalization 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.

FixMatch Ran
ran on a synthesized input · constructed an object; no result was computed
LeeDoYup/FixMatch-pytorch/models/fixmatch/fixmatch.py
code served (permissive licence) · get_code("95de992b4cf43fd7")
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, FixMatch: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.optim import SGD
from torch.optim.lr_scheduler import StepLR

torch.manual_seed(0)

# Define a simple backbone network
class SimpleNet(nn.Module):
    def __init__(self, num_classes):
        super(SimpleNet, self).__init__()
        self.fc1 = nn.Linear(4, 8)
        self.fc2 = nn.Linear(8, num_classes)
    
    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = self.fc2(x)
        return x

# Net builder function that returns a model
def net_builder(num_classes):
    return SimpleNet(num_classes)

# Simple logger stand-in
class SimpleLogger:
    def info(self, msg):
        pass

# Create ARGS dict
ARGS = {
    'net_builder': net_builder,
    'num_classes': 3,
    'ema_m': 0.999,
    'T': 0.95,
    'p_cutoff': 0.95,
    'lambda_u': 1.0,
    'hard_label': True,
    't_fn': None,
    'p_fn': None,
    'it': 0,
    'num_eval_iter': 1000,
    'tb_log': None,
    'logger': SimpleLogger()
}
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 FixMatch — 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 FixMatch 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
FixMatch Ran
ran on a synthesized input · constructed an object; no result was computed
gomezzz/distmsmatch/MSMatch/models/fixmatch/FixMatch.py
code served (permissive licence) · get_code("18865f52ae76d33f")
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, FixMatch: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

class SimpleNet(nn.Module):
    def __init__(self, num_classes=4, in_channels=3):
        super(SimpleNet, self).__init__()
        self.conv1 = nn.Conv2d(in_channels, 8, kernel_size=3, padding=1)
        self.fc = nn.Linear(8 * 32 * 32, num_classes)
    
    def forward(self, x):
        x = self.conv1(x)
        x = x.view(x.size(0), -1)
        x = self.fc(x)
        return x

class SimpleLogger:
    def info(self, msg):
        print(msg)

def net_builder(num_classes, in_channels):
    return SimpleNet(num_classes=num_classes, in_channels=in_channels)

ARGS = {
    'net_builder': net_builder,
    'num_classes': 4,
    'in_channels': 3,
    'ema_m': 0.999,
    'T': 1.0,
    'p_cutoff': 0.95,
    'lambda_u': 1.0,
    'hard_label': True,
    'tb_log': None,
    'logger': SimpleLogger(),
    'device': 'cpu',
    'rank': None
}
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 FixMatch — 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 FixMatch 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
Identity Ran
ran, violated the contract we drafted
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("0847c9a3c04c392e")
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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 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
import torch.nn as nn

torch.manual_seed(0)

# The Identity function takes img as first positional argument and **kwarg for any additional keyword arguments
# It simply returns img unchanged
# So we need to provide an image tensor and can optionally provide other keyword arguments

img = torch.randn(2, 3, 4, 4)  # batch=2, channels=3, height=4, width=4

ARGS = {
    'img': img
}
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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 4) — shape (2, 3, 4, 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
OhemCrossEntropy2D Ran
ran on a synthesized input
miquelmarti/DenseFixMatch/ssl_methods/fixmatch.py
pointer only (licence: NONE) · get_code("657ea9d40ca81745")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, miquelmarti/DenseFixMatch 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.

Posterize Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("99a25ff737f67a63")
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 Image. 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 PIL
from PIL import Image
import numpy as np

# Create a small PIL Image for testing
# Posterize requires an 8-bit image, so we'll create a simple RGB image
img_array = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)
test_img = Image.fromarray(img_array, mode='RGB')

# Define arguments for the Posterize function
ARGS = {
    'img': test_img,
    'v': 5,
    'max_v': 10,
    'bias': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
RandAugmentMC Ran
ran on a synthesized input
AhmadQasim/FixMatch/dataset/randaugment.py
code served (permissive licence) · get_code("4e833a8fd83394d8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Image. 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 random
import numpy as np
import PIL
import PIL.Image
import PIL.ImageOps
import PIL.ImageEnhance
import PIL.ImageDraw

# Create a small test image
test_img = PIL.Image.new('RGB', (32, 32), color=(128, 128, 128))

INIT_ARGS = {
    'n': 2,
    'm': 5
}

ARGS = {
    'img': test_img
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
ResNet Ran
ran on a synthesized input · constructed an object; no result was computed
amazon-research/exponential-moving-average-normalization/models/fixmatch.py
pointer only (licence: NONE) · get_code("88b25a8a9dc5f114")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, amazon-research/exponential-moving-average-normalization 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.

Rotate Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("db8a01288db8d302")
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 Image. 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 random
from PIL import Image

# Create a small PIL Image object that has a rotate method
# The rotate method is called on img with an integer angle parameter
img = Image.new('RGB', (4, 3), color='red')

ARGS = {
    'img': img,
    'v': 5,
    'max_v': 20,
    'bias': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
Sharpness Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("6d5a3435d35550b8")
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 Image. 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 PIL
from PIL import Image
import numpy as np

# Create a small test image
img_array = np.random.randint(0, 256, (10, 10, 3), dtype=np.uint8)
test_img = Image.fromarray(img_array)

ARGS = {
    "img": test_img,
    "v": 5,
    "max_v": 2.0,
    "bias": 0.5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
ShearX Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("c069bca31b82aa2c")
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 Image. 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 random
import PIL
from PIL import Image

# Create a small PIL Image for testing
img = Image.new('RGB', (100, 100), color='red')

ARGS = {
    'img': img,
    'v': 5,
    'max_v': 10,
    'bias': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
ShearY Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("47ddf827c80f0e32")
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 Image. 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 random
import PIL
from PIL import Image
import torch

# Create a simple PIL Image for testing
# The function calls img.transform() and img.size, so we need a valid PIL Image
test_image = Image.new('RGB', (100, 100), color='red')

ARGS = {
    'img': test_image,
    'v': 3,
    'max_v': 5,
    'bias': 0.5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
TranslateX Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("623b2503761b36c6")
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 Image. 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 random
import PIL
from PIL import Image
import numpy as np

# Create a small test image
img_array = np.random.randint(0, 256, (100, 120, 3), dtype=np.uint8)
test_img = Image.fromarray(img_array)

ARGS = {
    "img": test_img,
    "v": 5,
    "max_v": 10,
    "bias": 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
TranslateY Ran
ran; our contract draft was wrong, not the code
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("c34cf3418c8a3bb9")
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 Image. 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 random
import PIL
from PIL import Image

# Create a small test image
img = Image.new('RGB', (100, 80))

# Define ARGS dict with all parameters for TranslateY
ARGS = {
    'img': img,
    'v': 5,
    'max_v': 20,
    'bias': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
WideResNet Ran
ran on a synthesized input · constructed an object; no result was computed
OFRIN/Tensorflow_FixMatch/core/backbone/wideresnet.py
pointer only (licence: NONE) · get_code("b16158758833e962")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, OFRIN/Tensorflow_FixMatch 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.

WideResnet Ran
ran on a synthesized input · constructed an object; no result was computed
valencebond/FixMatch_pytorch/models/model.py
code served (permissive licence) · get_code("cee3484c4606263e")
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, WideResnet: 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
import math

torch.manual_seed(0)

# Define the ARGS dict for WideResnet
ARGS = {
    'n_classes': 10,
    'k': 1,
    'n': 28,
}
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 WideResnet — 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 WideResnet 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
WideResnet Ran
ran on a synthesized input · constructed an object; no result was computed
CoinCheung/fixmatch/model.py
code served (permissive licence) · get_code("b8e244cd7c2d12ef")
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, WideResnet: 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)

ARGS = {
    'n_classes': 10,
    'k': 2,
    'n': 28,
}
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 WideResnet — 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 WideResnet 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
WideResnetBackbone Ran
ran on a synthesized input · constructed an object; no result was computed
valencebond/FixMatch_pytorch/models/model.py
code served (permissive licence) · get_code("6986978358272b46")
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, WideResnetBackbone: 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
import math

torch.manual_seed(0)

ARGS = {
    'k': 1,
    'n': 28,
    'drop_rate': 0.0,
}
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 WideResnetBackbone — 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 WideResnetBackbone 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
WideResnetBackbone Ran
ran on a synthesized input · constructed an object; no result was computed
CoinCheung/fixmatch/model.py
code served (permissive licence) · get_code("03b0c8a89d29142b")
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, WideResnetBackbone: 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)

ARGS = {
    'k': 1,
    'n': 28,
    'drop_rate': 0,
}
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 WideResnetBackbone — 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 WideResnetBackbone 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
_float_parameter Ran
ran, honoured the contract we drafted
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("c0d914f7e98a27ad")
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 float, ndim 0, dtype family float, shape () 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

torch.manual_seed(0)

PARAMETER_MAX = 10

v = 5
max_v = 20

ARGS = {
    "v": v,
    "max_v": max_v,
}
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 float, ndim 0, dtype family float, shape () — shape () 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
_int_parameter Ran
ran, honoured the contract we drafted
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("895dd68f0abd14a2")
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 int, ndim 0, dtype family int, shape () 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

PARAMETER_MAX = 10

def _int_parameter(v, max_v):
    return int(v * max_v / PARAMETER_MAX)

# Setup code to construct valid arguments
v = 5
max_v = 20

ARGS = {
    "v": v,
    "max_v": max_v,
}
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 int, ndim 0, dtype family int, shape () — shape () 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
autocontrast_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("78ca1debd9c4ed87")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 5, 3) 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 cv2
import numpy as np

# Create a small test image with 3 channels (BGR format for OpenCV)
# Using uint8 dtype as expected by cv2 functions
img = np.random.randint(0, 256, size=(4, 5, 3), dtype=np.uint8)

ARGS = {
    "img": img,
    "cutoff": 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 5, 3) — shape (4, 5, 3) 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
calculate_gain Ran
ran on a synthesized input
S-HuaBomb/FixMatch-Paddle/models/initializers.py
code served (permissive licence) · get_code("96901d00bcfc5b79")
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 float, ndim 0, dtype family float, shape () 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
import torch.nn as nn
import math

# Test with 'leaky_relu' nonlinearity and a negative_slope parameter
ARGS = {
    'nonlinearity': 'leaky_relu',
    'param': 0.2
}
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 float, ndim 0, dtype family float, shape () — shape () 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
ce_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gomezzz/MSMatch/models/fixmatch/fixmatch.py
code served (permissive licence) · get_code("9723f46ec1689a37")
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) and returned Tensor, ndim 1, dtype family float, shape (2,) 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
import torch.nn.functional as F

torch.manual_seed(0)

# Test case 1: use_hard_labels=True
# logits: [batch_size, num_classes]
# targets: [batch_size] with integer class indices
logits_hard = torch.randn(2, 4)
targets_hard = torch.tensor([0, 2], dtype=torch.long)

# Test case 2: use_hard_labels=False
# logits: [batch_size, num_classes]
# targets: [batch_size, num_classes] with soft labels (probabilities)
logits_soft = torch.randn(2, 4)
targets_soft = F.softmax(torch.randn(2, 4), dim=-1)

# We'll use the hard labels case for the main ARGS
ARGS = {
    'logits': logits_hard,
    'targets': targets_hard,
    'use_hard_labels': True,
    'reduction': 'none'
}
Setup block drafted by
a language model, Claude Haiku 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 1, dtype family float, shape (2,) — shape (2,) 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
consistency_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LeeDoYup/FixMatch-pytorch/models/fixmatch/fixmatch.py
code served (permissive licence) · get_code("4a5f8edc558790c2")
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, 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
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for the function
batch_size = 2
num_classes = 4

# logits_w: weak model logits [batch_size, num_classes]
logits_w = torch.randn(batch_size, num_classes)

# logits_s: strong model logits [batch_size, num_classes]
logits_s = torch.randn(batch_size, num_classes)

# Define ARGS dictionary with all parameters
ARGS = {
    'logits_w': logits_w,
    'logits_s': logits_s,
    'name': 'ce',
    'T': 1.0,
    'p_cutoff': 0.0,
    'use_hard_labels': True,
}
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, ndim 1, dtype family float — a tuple 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
consistency_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gomezzz/distmsmatch/MSMatch/models/fixmatch/FixMatch.py
code served (permissive licence) · get_code("5efbd3f44ce7dcf5")
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, 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
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
num_classes = 4

# logits_w: weak augmentation logits [batch_size, num_classes]
logits_w = torch.randn(batch_size, num_classes)

# logits_s: strong augmentation logits [batch_size, num_classes]
logits_s = torch.randn(batch_size, num_classes)

# Test with 'ce' loss and use_hard_labels=True
ARGS = {
    'logits_w': logits_w,
    'logits_s': logits_s,
    'name': 'ce',
    'T': 1.0,
    'p_cutoff': 0.0,
    'use_hard_labels': True
}
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, ndim 1, dtype family float — a tuple 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
consistency_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gomezzz/MSMatch/models/fixmatch/fixmatch.py
code served (permissive licence) · get_code("cc357ff67b149646")
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) and returned tuple, 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
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
num_classes = 4

# logits_w: weak model logits [batch_size, num_classes]
logits_w = torch.randn(batch_size, num_classes)

# logits_s: strong model logits [batch_size, num_classes]
logits_s = torch.randn(batch_size, num_classes)

# Test with different parameter combinations
ARGS = {
    'logits_w': logits_w,
    'logits_s': logits_s,
    'name': 'ce',
    'T': 1.0,
    'p_cutoff': 0.0,
    'use_hard_labels': True,
}
Setup block drafted by
a language model, Claude Haiku 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 tuple, ndim 1, dtype family float — a tuple 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
contrast_func Ran
ran, violated the contract we drafted
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("3cbade2baaefb1c2")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 numpy as np

# Create a small test image as a numpy array
# The function expects img to be indexable with integer arrays
# and to support mean() and axis parameter
# Shape should be (height, width, channels) based on axis=(0,1) usage
img = np.random.randint(0, 256, size=(4, 3, 3), dtype=np.uint8)

# factor is a scalar that controls contrast adjustment
factor = 1.5

ARGS = {
    "img": img,
    "factor": factor,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
cross_entropy_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gomezzz/distmsmatch/MSMatch/models/fixmatch/FixMatch.py
code served (permissive licence) · get_code("774112e52d064668")
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 Tensor, ndim 1, dtype family float, shape (2,) 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
import torch.nn.functional as F

torch.manual_seed(0)

# Case 1: use_hard_labels=True
logits_hard = torch.randn(2, 4)  # batch_size=2, num_classes=4
targets_hard = torch.tensor([0, 2], dtype=torch.long)  # batch_size=2, class indices

# Case 2: use_hard_labels=False
logits_soft = torch.randn(2, 4)  # batch_size=2, num_classes=4
targets_soft = torch.randn(2, 4)  # batch_size=2, num_classes=4
targets_soft = targets_soft / targets_soft.sum(dim=1, keepdim=True)  # normalize to valid probabilities

# Test with hard labels
ARGS = {
    'logits': logits_hard,
    'targets': targets_hard,
    'use_hard_labels': True,
    'reduction': 'none'
}
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 Tensor, ndim 1, dtype family float, shape (2,) — shape (2,) 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
cutout_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("8a873e1fbc0db2dd")
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) and returned ndarray, ndim 3, dtype family int, shape (8, 6, 3) 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 numpy as np

# Create a small test image (H=8, W=6, C=3 for RGB)
img = np.random.randint(0, 256, size=(8, 6, 3), dtype=np.uint8)

# Define pad_size
pad_size = 4

# Define replace color (RGB tuple)
replace = (128, 128, 128)

ARGS = {
    "img": img,
    "pad_size": pad_size,
    "replace": replace,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (8, 6, 3) — shape (8, 6, 3) 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
equalize_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("4c53169f63a1fcef")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 cv2
import numpy as np

# Create a small test image with 3 channels (BGR format for cv2)
# Using uint8 dtype as required by cv2 functions
img = np.random.randint(0, 256, size=(4, 3, 3), dtype=np.uint8)

ARGS = {
    "img": img
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
fixmatch_augment_pool Ran
ran on a synthesized input
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("e34c59d1eb6d2adb")
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 2, 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 random
import PIL
from PIL import Image, ImageOps, ImageEnhance
import numpy as np

# Create a small test image
def create_test_image():
    """Create a small PIL Image for testing."""
    arr = np.random.randint(0, 256, (32, 32, 3), dtype=np.uint8)
    return Image.fromarray(arr, mode='RGB')

# Test each function individually to ensure they work
test_img = create_test_image()

ARGS = {}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, 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_weak_augmentation Ran
ran; our contract draft was wrong, not the code
fbuchert/fixmatch-pytorch/augmentation/augmentations.py
code served (permissive licence) · get_code("159f99b81bc70ebc")
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) and returned Compose. 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 torchvision import transforms

torch.manual_seed(0)

ARGS = {
    'img_size': 32,
    'padding': 4,
    'padding_mode': 'reflect'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Compose — a Compose output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Compose 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
safe_mean Ran
ran on a synthesized input
johnchenresearch/SSL/imagenet/training.py
code served (permissive licence) · get_code("e220f362daa9561f")
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 0, dtype family float, shape () 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 with some loss values
losses = tf.constant([1.0, 2.0, 3.0, 4.0], dtype=tf.float32)

ARGS = {
    "losses": losses
}
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 0, dtype family float, shape () — shape () 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
sharpening Ran
ran; our contract draft was wrong, not the code
SaraAmd/Semi-Supervised-Learning/ssl_lib/algs/consistency.py
pointer only (licence: NONE) · get_code("c52f15cee62d6c7a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, SaraAmd/Semi-Supervised-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.

sharpness_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("8d81faa31f390340")
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) and returned ndarray, ndim 3, dtype family int, shape (5, 5, 3) 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 cv2
import numpy as np

# Create a small test image with shape (height, width, channels)
# Using uint8 dtype as expected by the function
img = np.random.randint(0, 256, size=(5, 5, 3), dtype=np.uint8)

# Test with factor = 0.5 (intermediate case that exercises the full logic)
factor = 0.5

ARGS = {
    "img": img,
    "factor": factor,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (5, 5, 3) — shape (5, 5, 3) 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
shear_x_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("19aacbdbdaa4747b")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 cv2
import numpy as np

# Create a small test image (4x3 RGB image)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define shear factor
factor = 0.2

# Define fill color (BGR format for OpenCV)
fill = (0, 0, 0)

ARGS = {
    "img": img,
    "factor": factor,
    "fill": fill
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
shear_y_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("6514adf4f71311fc")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 cv2
import numpy as np

# Create a small test image (4x3 RGB image)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define the shear factor
factor = 0.2

# Define the fill color (BGR format for OpenCV)
fill = (0, 0, 0)

ARGS = {
    "img": img,
    "factor": factor,
    "fill": fill
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
tempereture_softmax Ran
ran, honoured the contract we drafted
SaraAmd/Semi-Supervised-Learning/ssl_lib/algs/consistency.py
pointer only (licence: NONE) · get_code("75653ad24ddcda46")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, SaraAmd/Semi-Supervised-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.

translate_x_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("adebf590ca9fbf8b")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 cv2
import numpy as np

# Create a small test image (4x3 RGB image)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define offset for horizontal translation
offset = 10

# Define fill color (BGR format for OpenCV)
fill = (0, 0, 0)

ARGS = {
    "img": img,
    "offset": offset,
    "fill": fill
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
translate_y_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CoinCheung/fixmatch-pytorch/randaugment.py
code served (permissive licence) · get_code("24b6eed27679e385")
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) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) 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 cv2
import numpy as np

# Create a small test image (4x3 RGB image)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define offset for y translation
offset = 2

# Define fill color (BGR format for cv2)
fill = (0, 0, 0)

ARGS = {
    "img": img,
    "offset": offset,
    "fill": fill
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) 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
x_u_split Ran
ran on a synthesized input
S-HuaBomb/FixMatch-Paddle/dataset/cifar.py
code served (permissive licence) · get_code("26e39e7c49b447ad")
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 math
import numpy as np

class Args:
    def __init__(self):
        self.num_labeled = 100
        self.num_classes = 10
        self.batch_size = 32
        self.eval_step = 5
        self.expand_labels = True

# Create labels array with 10 classes, enough samples per class
# Total samples = 1000 to have plenty for sampling
np.random.seed(0)
num_samples = 1000
labels = np.array([i % 10 for i in range(num_samples)])
np.random.shuffle(labels)

args = Args()

ARGS = {
    'args': args,
    'labels': labels
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
AutoContrast Not yet run
unverified (harvested, no recorded run)
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("338c2ab566a96a45")
How it ran

no invocation recorded.

CTAugment Not yet run
unverified (harvested, no recorded run)
Celiali/FixMatch/augmentations/ctaugment.py
code served (permissive licence) · get_code("d177a39181ef07d4")
How it ran

no invocation recorded.

Color Not yet run
unverified (harvested, no recorded run)
kekmodel/FixMatch-pytorch/dataset/randaugment.py
code served (permissive licence) · get_code("fbe48c6053d5de3b")
How it ran

no invocation recorded.

CutoutAbs Not yet run
unverified (harvested, no recorded run)
AhmadQasim/FixMatch/dataset/randaugment.py
code served (permissive licence) · get_code("f4c8609e68ec56f7")
How it ran

no invocation recorded.

FixMatch Not yet run
unverified (harvested, no recorded run)
miquelmarti/DenseFixMatch/ssl_methods/fixmatch.py
pointer only (licence: NONE) · get_code("2a1f6b2401f82e41")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, miquelmarti/DenseFixMatch 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.

FixMatch Not yet run
unverified (harvested, no recorded run)
gomezzz/MSMatch/models/fixmatch/fixmatch.py
code served (permissive licence) · get_code("bdc089534fcd34d6")
How it ran

no invocation recorded.

MultiLoss Not yet run
unverified (harvested, no recorded run)
miquelmarti/DenseFixMatch/ssl_methods/fixmatch.py
pointer only (licence: NONE) · get_code("980cb93dec539bfb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, miquelmarti/DenseFixMatch 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.

Showing the first 60. An agent gets every one from get_harvested_code_for_paper("2001.07685").

Also present in other papers' code

Of 74 samples lifted from this paper's linked repositories, 74 carry a function identity and 0 do not. Of the 69 distinct functions with an identity, 21 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 20 later) and 48 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
AverageMeter
brick of this paper's assembly: helper
30 other papers
1 dated earlier, 29 later · 25 only as an assembly brick
1904.12848 (2019-04)
Unsupervised Data Augmentation for Consistency Training
2002.10025 (2020-02)
Triple Wins: Boosting Accuracy, Robustness and Efficiency Together by Enabling Input-Adaptive Inference
2006.08572 (2020-06)
Flexible Dataset Distillation: Learn Labels Instead of Images
and 27 more not named here
AutoContrast
brick of this paper's assembly: data_transform
19 other papers
19 later
2003.10580 (2020-03)
Meta Pseudo Labels
2104.07713 (2021-04)
Contrastive Learning with Stronger Augmentations
2105.14148 (2021-05)
OpenMatch: Open-set Consistency Regularization for Semi-supervised Learning with Outliers
and 16 more not named here
Brightness
brick of this paper's assembly: data_transform
19 other papers
19 later
2003.10580 (2020-03)
Meta Pseudo Labels
2104.07713 (2021-04)
Contrastive Learning with Stronger Augmentations
2105.14148 (2021-05)
OpenMatch: Open-set Consistency Regularization for Semi-supervised Learning with Outliers
and 16 more not named here
Color
brick of this paper's assembly: data_transform
18 other papers
18 later
2003.10580 (2020-03)
Meta Pseudo Labels
2104.07713 (2021-04)
Contrastive Learning with Stronger Augmentations
2105.14148 (2021-05)
OpenMatch: Open-set Consistency Regularization for Semi-supervised Learning with Outliers
and 15 more not named here
equalize_func
brick of this paper's assembly: data_transform
15 other papers
15 later
Li_Align_and_Prompt_Video-and-Language_Pre-Training_With_Entity_Prompts_CVPR_2022_paper (2022 (year only))
Align and Prompt: Video-and-Language Pre-training with Entity Prompts
2207.05333 (2022-07)
IDEA: Increasing Text Diversity via Online Multi-Label Recognition for Vision-Language Pre-training
2302.06605 (2023-02)
UniAdapter: Unified Parameter-Efficient Transfer Learning for Cross-modal Modeling
and 12 more not named here
mish 15 other papers
1 dated earlier, 14 later
1908.00821 (2019-08)
Learning Lightweight Lane Detection CNNs by Self Attention Distillation
2105.14148 (2021-05)
OpenMatch: Open-set Consistency Regularization for Semi-supervised Learning with Outliers
2110.03006 (2021-10)
Unsupervised Selective Labeling for More Effective Semi-Supervised Learning
and 12 more not named here
autocontrast_func
brick of this paper's assembly: data_transform
12 other papers
12 later
Li_Align_and_Prompt_Video-and-Language_Pre-Training_With_Entity_Prompts_CVPR_2022_paper (2022 (year only))
Align and Prompt: Video-and-Language Pre-training with Entity Prompts
2207.05333 (2022-07)
IDEA: Increasing Text Diversity via Online Multi-Label Recognition for Vision-Language Pre-training
2302.06605 (2023-02)
UniAdapter: Unified Parameter-Efficient Transfer Learning for Cross-modal Modeling
and 9 more not named here
build_wideresnet 6 other papers
6 later
2110.03006 (2021-10)
Unsupervised Selective Labeling for More Effective Semi-Supervised Learning
2208.11870 (2022-08)
Fix-A-Step: Semi-supervised Learning from Uncurated Unlabeled Data
2307.11030 (2023-07)
Cluster-aware Semi-supervised Learning: Relational Knowledge Distillation Provably Learns Clustering
and 3 more not named here
entropy_from_logits 6 other papers
3 dated earlier, 3 later
1911.09785 (2019-11)
ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
1912.00594 (2019-12)
Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
1912.08766 (2019-12)
RealMix: Towards Realistic Semi-Supervised Deep Learning Algorithms
and 3 more not named here
entropy_penalty 6 other papers
3 dated earlier, 3 later
1911.09785 (2019-11)
ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
1912.00594 (2019-12)
Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
1912.08766 (2019-12)
RealMix: Towards Realistic Semi-Supervised Deep Learning Algorithms
and 3 more not named here
smart_shape 6 other papers
3 dated earlier, 3 later
1911.09785 (2019-11)
ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
1912.00594 (2019-12)
Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
1912.08766 (2019-12)
RealMix: Towards Realistic Semi-Supervised Deep Learning Algorithms
and 3 more not named here
ce_loss
brick of this paper's assembly: helper
1 other paper
1 later
2309.03598 (2023-09)
Enhancing Sample Utilization through Sample Adaptive Augmentation in Semi-Supervised Learning
contrast_func
brick of this paper's assembly: data_transform
1 other paper
1 later
2504.18397 (2025-04)
Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization
get_hparams 1 other paper
1 later
2106.16209 (2021-06)
A data-centric approach for improving ambiguous labels with combined semi-supervised classification and clustering
get_weak_augmentation
brick of this paper's assembly: entry
1 other paper
1 dated earlier
1905.02249 (2019-05)
MixMatch: A Holistic Approach to Semi-Supervised Learning
safe_mean 1 other paper
1 later
2106.16209 (2021-06)
A data-centric approach for improving ambiguous labels with combined semi-supervised classification and clustering
sharpness_func
brick of this paper's assembly: data_transform
1 other paper
1 later
2504.18397 (2025-04)
Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization
shear_x_func
brick of this paper's assembly: data_transform
1 other paper
1 later
2504.18397 (2025-04)
Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization
shear_y_func
brick of this paper's assembly: data_transform
1 other paper
1 later
2504.18397 (2025-04)
Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization
translate_x_func
brick of this paper's assembly: data_transform
1 other paper
1 later
2504.18397 (2025-04)
Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization

Shown: the 20 functions present in the most other papers. 1 more shared functions are not listed here.

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

Semi-supervised learning (SSL) provides an effective means of leveraging unlabeled data to improve a model's performance. This domain has seen fast progress recently, at the cost of requiring more complex methods. In this paper we propose FixMatch, an algorithm that is a significant simplification of existing SSL methods. FixMatch first generates pseudo-labels using the model's predictions on weaklyaugmented unlabeled images. For a given image, the pseudo-label is only retained if the model produces a high-confidence prediction. The model is then trained to predict the pseudo-label when fed a strongly-augmented version of the same image. Despite its simplicity, we show that FixMatch achieves state-of-the-art performance across a variety of standard semi-supervised learning benchmarks, including 94.93% accuracy on CIFAR-10 with 250 labels and 88.61% accuracy with 40 -just 4 labels per class. We carry out an extensive ablation study to tease apart the experimental factors that are most important to FixMatch's success. The code is available at https://github.com/google-research/fixmatch.

For agents

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

get_harvested_code_for_paper("2001.07685")
get_code_for_paper("2001.07685")
have("2001.07685")

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

curl https://syntology.ai/api/ran/2001.07685.json

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

[![Syntology run record](https://syntology.ai/api/ran/2001.07685.svg)](https://syntology.ai/paper/2001.07685)

Connect an agent — have() is free.