SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.13253 · ICLR · 2022

Learning Versatile Neural Architectures by Propagating Network Codes

Zhe Wang, Ping Luo, Mingyu Ding, Yuqi Huo, Haoyu Lu, Linjie Yang, Zhiwu Lu, Jingdong Wang, Cls Predictor, Seg Predictor

arXiv · PDF · Open in the Atlas

Code that ran

13 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 6 failures of our instrument; 1 no recorded run), measured 2026-11-22; 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/2103.13253.json · badge: /api/ran/2103.13253.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
dingmyu/NCP canonical (classifier label)
classifier basis: classifier_2025
13 ran · 1 no recorded run
FunctionStatusWhere it lives
BasicBlock Ran
ran on a synthesized input · constructed an object; no result was computed
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("7653ec0dc3bc04c8")
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, BasicBlock: 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 = {
    "inplanes": 4,
    "planes": 8,
    "stride": 1,
    "active_fn": nn.ReLU,
}
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 BasicBlock — 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 BasicBlock 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
ConvBNReLU Ran
ran on a synthesized input · constructed an object; no result was computed
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("a64ad8762fe201c4")
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, ConvBNReLU: 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)

# Define ARGS dict with all parameters for ConvBNReLU.__init__
ARGS = {
    'in_planes': 3,
    'out_planes': 16,
    'kernel_size': 3,
    'stride': 1,
    'groups': 1,
    'active_fn': nn.ReLU,
    'batch_norm_kwargs': {},
    'dilation': 1,
    'padding': 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 ConvBNReLU — 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 ConvBNReLU 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
FuseModule Ran
ran on a synthesized input · constructed an object; no result was computed
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("0aa384df5d22c220")
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, FuseModule: 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 functools

torch.manual_seed(0)

# Define helper functions and classes that are used
def conv3x3(in_planes, out_planes, stride=1, dilation=1):
    """3x3 convolution with padding."""
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=dilation, dilation=dilation, bias=False)

class Identity(nn.Module):
    """Module proxy for null op."""
    def forward(self, x):
        return x

def get_active_fn(name):
    """Select activation function."""
    active_fn = {
        'nn.ReLU6': functools.partial(nn.ReLU6, inplace=True),
        'nn.ReLU': functools.partial(nn.ReLU, inplace=True)
    }[name]
    return active_fn

class ConvBNReLU(nn.Sequential):
    """Convolution-BatchNormalization-ActivateFn."""
    def __init__(self, in_planes, out_planes, kernel_size=3, stride=1, groups=1, active_fn=None, batch_norm_kwargs=None, dilation=1, padding=None, expand_ratio=None, kernel_sizes=None, **kwargs):
        if batch_norm_kwargs is None:
            batch_norm_kwargs = {}
        if not padding:
            padding = (kernel_size - 1) // 2
        super(ConvBNReLU, self).__init__(
            nn.Conv2d(in_planes, out_planes, kernel_size, stride, padding, dilation=dilation, groups=groups, bias=False),
            nn.BatchNorm2d(out_planes, **batch_norm_kwargs),
            active_fn() if active_fn is not None else Identity()
        )

class BasicBlock(nn.Module):
    expansion = 1
    def __init__(self, inplanes, planes, stride=1, active_fn=None, **kwargs):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = active_fn() if active_fn is not None else nn.ReLU(inplace=True)
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        self.downsample = None
        self.stride = stride
        if self.stride != 1 or inplanes != planes:
            self.downsample = nn.Sequential(
                nn.Conv2d(inplanes, planes, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(planes)
            )

    def forward(self, x):
        residual = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        if self.downsample is not None:
            residual = self.downsample(x)
        out += residual
        out = self.relu(out)
        return out

def get_block_wrapper(block_str):
    """Wrapper for network block."""
    if block_str == 'ConvBNReLU':
        return ConvBNReLU
    elif block_str == 'BasicBlock':
        return BasicBlock
    else:
        raise ValueError('Unknown type of blocks.')

# Construct arguments for FuseModule
ARGS = {
    'in_branches': 2,
    'out_branches': 2,
    'block': get_block_wrapper('BasicBlock'),
    'in_channels': [16, 32],
    'out_channels': [16, 32],
    'expand_ratio': 6,
    'kernel_sizes': [3],
    'batch_norm_kwargs': None,
    'active_fn': get_active_fn('nn.ReLU6'),
    'use_hr_format': False,
    'only_fuse_neighbor': True,
    'directly_downsample': 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 FuseModule — 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 FuseModule 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 on a synthesized input
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("eb4511eaa68b7bd7")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import nn

class Identity(nn.Module):
    """Module proxy for null op."""

    def forward(self, x):
        return x

INIT_ARGS = {}

ARGS = {
    'x': torch.randn(2, 4)
}

if __name__ == '__main__':
    module = Identity(**INIT_ARGS)
    output = module(**ARGS)
    print(output)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
ParallelModule Ran
ran on a synthesized input · constructed an object; no result was computed
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("965c95dc0619b08a")
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, ParallelModule: 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 functools

torch.manual_seed(0)

class Identity(nn.Module):
    """Module proxy for null op."""
    def forward(self, x):
        return x

def conv3x3(in_planes, out_planes, stride=1, dilation=1):
    """3x3 convolution with padding."""
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=dilation, dilation=dilation, bias=False)

class BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, inplanes, planes, stride=1, active_fn=None, **kwargs):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = active_fn()
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        self.downsample = None
        self.stride = stride
        if self.stride != 1 or inplanes != planes:
            self.downsample = nn.Sequential(nn.Conv2d(inplanes, planes, kernel_size=1, stride=stride, bias=False), nn.BatchNorm2d(planes))

    def forward(self, x):
        residual = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        if self.downsample is not None:
            residual = self.downsample(x)
        out += residual
        out = self.relu(out)
        return out

class ConvBNReLU(nn.Sequential):
    """Convolution-BatchNormalization-ActivateFn."""

    def __init__(self, in_planes, out_planes, kernel_size=3, stride=1, groups=1, active_fn=None, batch_norm_kwargs=None, dilation=1, padding=None, **kwargs):
        if batch_norm_kwargs is None:
            batch_norm_kwargs = {}
        if not padding:
            padding = (kernel_size - 1) // 2
        super(ConvBNReLU, self).__init__(nn.Conv2d(in_planes, out_planes, kernel_size, stride, padding, dilation=dilation, groups=groups, bias=False), nn.BatchNorm2d(out_planes, **batch_norm_kwargs), active_fn() if active_fn is not None else Identity())

def get_active_fn(name):
    """Select activation function."""
    active_fn = {'nn.ReLU6': functools.partial(nn.ReLU6, inplace=True), 'nn.ReLU': functools.partial(nn.ReLU, inplace=True)}[name]
    return active_fn

def get_block_wrapper(block_str):
    """Wrapper for network block."""
    if block_str == 'ConvBNReLU':
        return ConvBNReLU
    elif block_str == 'BasicBlock':
        return BasicBlock
    else:
        raise ValueError('Unknown type of blocks.')

ARGS = {
    'num_branches': 2,
    'block': get_block_wrapper('BasicBlock'),
    'num_blocks': [2, 2],
    'num_channels': [4, 5],
    'expand_ratio': 6,
    'kernel_sizes': [3],
    'batch_norm_kwargs': None,
    'active_fn': get_active_fn('nn.ReLU6')
}
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 ParallelModule — 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 ParallelModule 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
_make_divisible Ran
ran, honoured the contract we drafted
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("733638be3c212532")
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 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
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# - v: a numeric value (int or float) representing channels
# - divisor: a numeric value used to make v divisible
# - min_value: optional numeric value, defaults to divisor if None

# The function performs arithmetic operations on these values
# and returns a modified channel count

ARGS = {
    "v": 35,
    "divisor": 8,
    "min_value": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, 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
conv3x3 Ran
ran; our contract draft was wrong, not the code
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("52b42ce8214983fb")
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 Conv2d. 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 = {
    'in_planes': 3,
    'out_planes': 64,
    'stride': 1,
    'dilation': 1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
denormalize Ran
ran; our contract draft was wrong, not the code
dingmyu/NCP/tools/ncp.py
code served (permissive licence) · get_code("01cd9fa7bdf70ca2")
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 1, dtype family float, shape (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 numpy as np

torch.manual_seed(0)

# Create a small normalized tensor (batch_size=1, features=4)
normalized_tensor = torch.randn(1, 4)

# Create a config object with the required attributes
class Config:
    class Data:
        def __init__(self):
            self.input_mean = np.array([1.0, 2.0, 3.0, 4.0])
            self.input_std = np.array([0.5, 0.5, 0.5, 0.5])
    
    def __init__(self):
        self.data = self.Data()

cfg = Config()

ARGS = {
    "normalized_data": normalized_tensor,
    "cfg": cfg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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
get_active_fn Ran
ran; our contract draft was wrong, not the code
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("9ce0aa3acf7a46fd")
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 partial. 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 functools
import torch
from torch import nn

# Call this function with a valid activation function name
# The function expects a string that is a key in the dict
# Valid options are: 'nn.ReLU6' or 'nn.ReLU'

ARGS = {
    'name': 'nn.ReLU'
}
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 partial — a partial output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a partial 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
get_block_wrapper Ran
ran; our contract draft was wrong, not the code
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("2448b05701889c45")
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 type. 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)

class Identity(nn.Module):
    """Module proxy for null op."""
    def forward(self, x):
        return x

def conv3x3(in_planes, out_planes, stride=1, dilation=1):
    """3x3 convolution with padding."""
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=dilation, dilation=dilation, bias=False)

class BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, inplanes, planes, stride=1, active_fn=None, **kwargs):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = active_fn()
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        self.downsample = None
        self.stride = stride
        if self.stride != 1 or inplanes != planes:
            self.downsample = nn.Sequential(nn.Conv2d(inplanes, planes, kernel_size=1, stride=stride, bias=False), nn.BatchNorm2d(planes))

    def forward(self, x):
        residual = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        if self.downsample is not None:
            residual = self.downsample(x)
        out += residual
        out = self.relu(out)
        return out

class ConvBNReLU(nn.Sequential):
    """Convolution-BatchNormalization-ActivateFn."""

    def __init__(self, in_planes, out_planes, kernel_size=3, stride=1, groups=1, active_fn=None, batch_norm_kwargs=None, dilation=1, padding=None, **kwargs):
        if batch_norm_kwargs is None:
            batch_norm_kwargs = {}
        if not padding:
            padding = (kernel_size - 1) // 2
        super(ConvBNReLU, self).__init__(nn.Conv2d(in_planes, out_planes, kernel_size, stride, padding, dilation=dilation, groups=groups, bias=False), nn.BatchNorm2d(out_planes, **batch_norm_kwargs), active_fn() if active_fn is not None else Identity())

def get_block_wrapper(block_str):
    """Wrapper for netowork block."""
    if block_str == 'ConvBNReLU':
        return ConvBNReLU
    elif block_str == 'BasicBlock':
        return BasicBlock
    else:
        raise ValueError('Unknown type of blocks.')

ARGS = {
    'block_str': 'ConvBNReLU'
}
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 type — a type output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a type 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
normalize Ran
ran; our contract draft was wrong, not the code
dingmyu/NCP/tools/ncp.py
code served (permissive licence) · get_code("fe7038d962bb928a")
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 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a small stand-in config class with nested data attribute
class DataConfig:
    def __init__(self):
        self.input_mean = torch.tensor([0.5, 0.3, 0.2, 0.1])
        self.input_std = torch.tensor([0.2, 0.15, 0.25, 0.1])

class Config:
    def __init__(self):
        self.data = DataConfig()

# Create test data: batch_size=2, feature_dim=4
data = torch.randn(2, 4)

# Create config
cfg = Config()

# Set device
device = torch.device('cpu')

ARGS = {
    'data': data,
    'cfg': cfg,
    'device': device
}
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
resize Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("e8fbff5c3395ac23")
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, 8, 10) 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 torch.nn.functional as F

torch.manual_seed(0)

# Create a small input tensor for interpolation
# Shape: (batch=2, channels=3, height=4, width=5)
input_tensor = torch.randn(2, 3, 4, 5)

# Define ARGS dict with all parameters
ARGS = {
    'input': input_tensor,
    'size': (8, 10),
    'scale_factor': None,
    'mode': 'nearest',
    'align_corners': 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 4, dtype family float, shape (2, 3, 8, 10) — shape (2, 3, 8, 10) 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
valid_epoch Ran
ran, honoured the contract we drafted
dingmyu/NCP/tools/train_predictor.py
code served (permissive licence) · get_code("2cc6dcc541f32757")
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 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 logging
import numpy as np

torch.manual_seed(0)

# Set up logging
logging.basicConfig(level=logging.INFO)

# Create a minimal model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

model = SimpleModel()

# Create a minimal dataloader (list of batches)
dataloader = [
    [torch.randn(2, 4), torch.randn(2, 3)],
    [torch.randn(2, 4), torch.randn(2, 3)],
]

# Create a criterion that returns per-sample losses
criterion = nn.MSELoss(reduction='none')

# Device
device = torch.device('cpu')

# Epoch number
epoch = 1

# Create a minimal config object
class Config:
    class Data:
        combined_metrics = ['metric1', 'metric2', 'metric3']
        output_std = np.array([1.0, 1.0, 1.0])
        output_mean = np.array([0.0, 0.0, 0.0])
    
    data = Data()

cfg = Config()

ARGS = {
    'model': model,
    'dataloader': dataloader,
    'criterion': criterion,
    'device': device,
    'epoch': epoch,
    'cfg': cfg,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, 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
MultiResolutionNet Not yet run
unverified (harvested, no recorded run)
dingmyu/NCP/models/supernet.py
code served (permissive licence) · get_code("c808f48b3238adbb")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 4 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 4 later) and 10 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
conv3x3
brick of this paper's assembly: helper
18 other papers
7 dated earlier, 11 later
1612.00593 (2016-12)
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
1901.04780 (2019-01)
DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion
1905.09646 (2019-05)
Spatial Group-wise Enhance: Improving Semantic Feature Learning in Convolutional Networks
and 15 more not named here
_make_divisible
brick of this paper's assembly: helper
11 other papers
5 dated earlier, 6 later · 7 only as an assembly brick
1704.04861 (2017-04)
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
1801.04381 (2018-01)
MobileNetV2: Inverted Residuals and Linear Bottlenecks
1802.02611 (2018-02)
Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
and 8 more not named here
Identity
brick of this paper's assembly: helper
10 other papers
2 dated earlier, 1 the same month, 7 later · 8 only as an assembly brick
1511.06434 (2015-11)
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
1802.03268 (2018-02)
Efficient Neural Architecture Search via Parameter Sharing
2103.02376 (2021-03)
Event-based Synthetic Aperture Imaging with a Hybrid Network
and 7 more not named here
resize
brick of this paper's assembly: helper
1 other paper
1 later
2106.06560 (2021-06)
HR-NAS: Searching Efficient High-Resolution Neural Architectures with Lightweight Transformers

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

This work explores how to design a single neural network capable of adapting to multiple heterogeneous vision tasks, such as image segmentation, 3D detection, and video recognition. This goal is challenging because both network architecture search (NAS) spaces and methods in different tasks are inconsistent. We solve this challenge from both sides. We first introduce a unified design space for multiple tasks and build a multitask NAS benchmark (NAS-Bench-MR) on many widely used datasets, including ImageNet, Cityscapes, KITTI, and HMDB51. We further propose Network Coding Propagation (NCP), which back-propagates gradients of neural predictors to directly update architecture codes along the desired gradient directions to solve various tasks. In this way, optimal architecture configurations can be found by NCP in our large search space in seconds. Unlike prior arts of NAS that typically focus on a single task, NCP has several unique benefits. (1) NCP transforms architecture optimization from data-driven to architecture-driven, enabling joint search an architecture among multitasks with different data distributions. (2) NCP learns from network codes but not original data, enabling it to update the architecture efficiently across datasets. (3) In addition to our NAS-Bench-MR, NCP performs well on other NAS benchmarks, such as NAS-Bench-201. (4) Thorough studies of NCP on inter-, cross-, and intra-tasks highlight the importance of cross-task neural architecture design, i.e., multitask neural architectures and architecture transferring between different tasks. Code is available at github.com/dingmyu/NCP 1 .

For agents

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

get_harvested_code_for_paper("2103.13253")
get_code_for_paper("2103.13253")
have("2103.13253")

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

curl https://syntology.ai/api/ran/2103.13253.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.13253.svg)](https://syntology.ai/paper/2103.13253)

Connect an agent — have() is free.