SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.06799 · 2021

Zero-Cost Operation Scoring in Differentiable Architecture Search

arXiv · PDF · Open in the Atlas

Code that ran

8 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 4 have no contract verdict; 3 failures of our instrument; 2 no recorded run); 2 of them were also run on the host, not in the jail, by the peer-agreement lane that set their level, measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2106.06799.json · badge: /api/ran/2106.06799.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
zerocostptnas/zerocost_operation_score canonical (classifier label)
classifier basis: census
8 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_md5 Ran
ran; our contract draft was wrong, not the code
zerocostptnas/zerocost_operation_score/nasbench201/DownsampledImageNet.py
code served (permissive licence) · get_code("ad0636a128295d5a")
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 str. 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 2026-09-05T01:26:15Z). 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 os
import tempfile
import hashlib

# Create a temporary file with some content for testing
temp_dir = tempfile.mkdtemp()
test_file = os.path.join(temp_dir, "test_file.bin")

# Write some test data to the file
with open(test_file, 'wb') as f:
    f.write(b"Hello, this is test data for MD5 calculation." * 100)

ARGS = {
    "fpath": test_file,
    "chunk_size": 1024 * 1024
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
check_md5 Ran
ran, violated the contract we drafted
zerocostptnas/zerocost_operation_score/nasbench201/DownsampledImageNet.py
code served (permissive licence) · get_code("380a00b777aad655")
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 bool, ndim 0, dtype family float, shape () 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 2026-09-04T23:17:45Z). 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 os
import sys
import hashlib
import torch
import tempfile

# Create a temporary file with known content
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test_file.bin")

# Write some test data to the file
test_data = b"Hello, this is test data for MD5 calculation!"
with open(temp_file, 'wb') as f:
    f.write(test_data)

# Calculate the MD5 hash of the test data
md5_hash = hashlib.md5(test_data).hexdigest()

# Define arguments for check_md5 function
ARGS = {
    'fpath': temp_file,
    'md5': md5_hash
}
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 bool, 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
Jocab_Score Ran
ran on a synthesized input
zerocostptnas/zerocost_operation_score/nasbench201/op_score.py
code served (permissive licence) · get_code("37c5d8f7980d3748")
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 float32, ndim 0, dtype family float, shape () as a later run recorded it. 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, 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 numpy as np
import copy

torch.manual_seed(0)

class StandInModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.num_edge = 3
        self.num_op = 2
        self.candidate_flags = [False] * self.num_edge
        self.K = torch.zeros(2, 2)
        
        self.relu1 = nn.ReLU()
        self.linear1 = nn.Linear(4, 8)
        self.relu2 = nn.ReLU()
        self.linear2 = nn.Linear(8, 5)
        self.proj_weights = None
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.relu1(x)
        x = self.linear2(x)
        x = self.relu2(x)
        return x

ori_model = StandInModel()
input_tensor = torch.randn(2, 4)
target_tensor = torch.randint(0, 5, (2,))
weights = torch.randn(3, 4)

ARGS = {
    'ori_model': ori_model,
    'input': input_tensor,
    'target': target_tensor,
    'weights': weights
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float32, 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
check_integrity Ran
ran on a synthesized input
zerocostptnas/zerocost_operation_score/nasbench201/DownsampledImageNet.py
code served (permissive licence) · get_code("d3f4daf499849800")
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 bool, 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 os
import tempfile
import hashlib

# Create a temporary file with some content
temp_dir = tempfile.mkdtemp()
test_file = os.path.join(temp_dir, "test_file.txt")

# Write test content to file
test_content = b"Hello, this is test content for MD5 checking."
with open(test_file, 'wb') as f:
    f.write(test_content)

# Calculate the MD5 hash of the test file
md5_hash = hashlib.md5(test_content).hexdigest()

# Define ARGS dictionary with all parameters for check_integrity
ARGS = {
    "fpath": test_file,
    "md5": md5_hash
}
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 bool, 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
conv_1x1_bn Ran
ran; our contract draft was wrong, not the code
zerocostptnas/zerocost_operation_score/mobilenet_search_space/retrain_architecture/model.py
pointer only (licence: MIT); the stored copy get_code serves, from zhoudaquan/rethinking_bottleneck_design, is pointer only (licence: NOASSERTION) · get_code("a0131fb70c267a9e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from zhoudaquan/rethinking_bottleneck_design, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

conv_bn Ran
ran; our contract draft was wrong, not the code
zerocostptnas/zerocost_operation_score/mobilenet_search_space/retrain_architecture/model.py
code served (permissive licence) · get_code("e087cbfa1e53486d")
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 Sequential. 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 = {
    "inp": 3,
    "oup": 16,
    "stride": 1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
hooklogdet Ran
ran on a synthesized input
zerocostptnas/zerocost_operation_score/nasbench201/op_score.py
code served (permissive licence) · get_code("8d9701991ac208d4")
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 float64, ndim 0, 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 numpy as np

# Create a small positive definite matrix for valid slogdet computation
# Using a simple symmetric positive definite matrix
K = np.array([
    [2.0, 0.5, 0.1, 0.2],
    [0.5, 3.0, 0.2, 0.1],
    [0.1, 0.2, 2.5, 0.3],
    [0.2, 0.1, 0.3, 2.0]
], dtype=np.float64)

labels = np.array([0, 1, 0, 1], dtype=np.int64)

ARGS = {
    "K": K,
    "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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
recal_bn Ran
ran on a synthesized input
zerocostptnas/zerocost_operation_score/nasbench201/op_score.py
code served (permissive licence) · get_code("f057ad99f408ee4d")
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 MinimalNetwork. 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)

# Build a minimal network with BatchNorm2d layers
class MinimalNetwork(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 4, kernel_size=3, padding=1)
        self.bn1 = nn.BatchNorm2d(4)
        self.conv2 = nn.Conv2d(4, 8, kernel_size=3, padding=1)
        self.bn2 = nn.BatchNorm2d(8)
    
    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.conv2(x)
        x = self.bn2(x)
        return x, None

network = MinimalNetwork()

# Create a minimal data loader: list of (inputs, targets) tuples
# Batch size 2, 3 input channels, 4x4 spatial dimensions
xloader = [
    (torch.randn(2, 3, 4, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 3, 4, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 3, 4, 4), torch.randint(0, 10, (2,))),
]

recalbn = 2
device = 'cpu'

ARGS = {
    'network': network,
    'xloader': xloader,
    'recalbn': recalbn,
    'device': device,
}
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 MinimalNetwork — a MinimalNetwork output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MinimalNetwork 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
distill Not yet run
unverified (harvested, no recorded run)
zerocostptnas/zerocost_operation_score/nasbench201/networks_proposal.py
code served (permissive licence) · get_code("cb6cc47167b2300d")
How it ran

no invocation recorded.

query Not yet run
unverified (harvested, no recorded run)
zerocostptnas/zerocost_operation_score/nasbench201/networks_proposal.py
code served (permissive licence) · get_code("868f9ec838a3822b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 6 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 2 the same month, 6 later) and 4 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
conv_1x1_bn 31 other papers
15 dated earlier, 15 later, 1 same year, order not determinable · 2 only as an assembly brick
1703.07402 (2017-03)
Simple Online and Realtime Tracking with a Deep Association Metric
1801.04381 (2018-01)
MobileNetV2: Inverted Residuals and Linear Bottlenecks
1811.08383 (2018-11)
TSM: Temporal Shift Module for Efficient Video Understanding
and 28 more not named here
conv_bn 18 other papers
11 dated earlier, 6 later, 1 same year, order not determinable · 1 only as an assembly brick
1611.06403 (2016-11)
Deep Outdoor Illumination Estimation
1703.07402 (2017-03)
Simple Online and Realtime Tracking with a Deep Association Metric
1801.04381 (2018-01)
MobileNetV2: Inverted Residuals and Linear Bottlenecks
and 15 more not named here
calculate_md5 13 other papers
2 dated earlier, 1 the same month, 10 later
2001.00326 (2020-01)
NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search
2003.09553 (2020-03)
Adversarial Continual Learning
2106.05390 (2021-06)
Optimizing Reusable Knowledge for Continual Learning via Metalearning
and 10 more not named here
check_md5 13 other papers
2 dated earlier, 1 the same month, 10 later
2001.00326 (2020-01)
NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search
2003.09553 (2020-03)
Adversarial Continual Learning
2106.05390 (2021-06)
Optimizing Reusable Knowledge for Continual Learning via Metalearning
and 10 more not named here
distill 2 other papers
1 dated earlier, 1 later
2006.04492 (2020-06)
Speedy Performance Estimation for Neural Architecture Search
2108.04392 (2021-08)
Rethinking Architecture Selection in Differentiable NAS
hooklogdet 1 other paper
1 later
2205.05662 (2022-05)
Deep Architecture Connectivity Matters for Its Convergence: A Fine-Grained Analysis

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

Repositories linked to this paper

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

Abstract

We formalize and analyze a fundamental component of differentiable neural architecture search (NAS): local "operation scoring" at each operation choice. We view existing operation scoring functions as inexact proxies for accuracy, and we find that they perform poorly when analyzed empirically on NAS benchmarks. From this perspective, we introduce a novel \textit{perturbation-based zero-cost operation scoring} (Zero-Cost-PT) approach, which utilizes zero-cost proxies that were recently studied in multi-trial NAS but degrade significantly on larger search spaces, typical for differentiable NAS. We conduct a thorough empirical evaluation on a number of NAS benchmarks and large search spaces, from NAS-Bench-201, NAS-Bench-1Shot1, NAS-Bench-Macro, to DARTS-like and MobileNet-like spaces, showing significant improvements in both search time and accuracy. On the ImageNet classification task on the DARTS search space, our approach improved accuracy compared to the best current training-free methods (TE-NAS) while being over 10$\times$ faster (total searching time 25 minutes on a single GPU), and observed significantly better transferability on architectures searched on the CIFAR-10 dataset with an accuracy increase of 1.8 pp. Our code is available at: https://github.com/zerocostptnas/zerocost_operation_score.

For agents

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

get_harvested_code_for_paper("2106.06799")
get_code_for_paper("2106.06799")
have("2106.06799")

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

curl https://syntology.ai/api/ran/2106.06799.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.06799.svg)](https://syntology.ai/paper/2106.06799)

Connect an agent — have() is free.