SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.13425 · CVPR · 2021

Diverse Branch Block: Building a Convolution as an Inception-like Unit

Xiangyu Zhang, Xiaohan Ding, Jungong Han, Guiguang Ding

arXiv · PDF · Open in the Atlas

Code that ran

11 of 20 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 7 failures of our instrument; 9 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2103.13425.json · badge: /api/ran/2103.13425.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
hunto/image_classification_sota community
Papers with Code archive
9 ran · 1 no recorded run
DingXiaoH/DiverseBranchBlock marked official in Papers with Code
Papers with Code archive
2 ran · 2 no recorded run
xdhe1216/ACNet community
Papers with Code archive
0 ran · 6 no recorded run
FunctionStatusWhere it lives
BNAndPad Ran
ran on a synthesized input · constructed an object; no result was computed
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("f8ecece04e49713f")
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, BNAndPad: 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)

class BNAndPad(nn.Module):

    def __init__(self, pad_pixels, num_features, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True, last_conv_bias=None, bn=nn.BatchNorm2d):
        super(BNAndPad, self).__init__()
        self.bn = bn(num_features, eps, momentum, affine, track_running_stats)
        self.pad_pixels = pad_pixels
        self.last_conv_bias = last_conv_bias

    def forward(self, input):
        output = self.bn(input)
        if self.pad_pixels > 0:
            bias = -self.bn.running_mean
            if self.last_conv_bias is not None:
                bias += self.last_conv_bias
            pad_values = self.bn.bias.data + self.bn.weight.data * (bias / torch.sqrt(self.bn.running_var + self.bn.eps))
            ' pad '
            n, c, h, w = output.size()
            values = pad_values.view(1, -1, 1, 1)
            w_values = values.expand(n, -1, self.pad_pixels, w)
            x = torch.cat([w_values, output, w_values], dim=2)
            h = h + self.pad_pixels * 2
            h_values = values.expand(n, -1, h, self.pad_pixels)
            x = torch.cat([h_values, x, h_values], dim=3)
            output = x
        return output

    @property
    def weight(self):
        return self.bn.weight

    @property
    def bias(self):
        return self.bn.bias

    @property
    def running_mean(self):
        return self.bn.running_mean

    @property
    def running_var(self):
        return self.bn.running_var

    @property
    def eps(self):
        return self.bn.eps


ARGS = {
    'pad_pixels': 2,
    'num_features': 4,
    'eps': 1e-05,
    'momentum': 0.1,
    'affine': True,
    'track_running_stats': True,
    'last_conv_bias': torch.tensor([0.1, 0.2, 0.3, 0.4]),
    'bn': nn.BatchNorm2d
}
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 BNAndPad — 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 BNAndPad 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
BNAndPadLayer Ran
ran on a synthesized input · constructed an object; no result was computed
DingXiaoH/DiverseBranchBlock/diversebranchblock.py
code served (permissive licence) · get_code("0ef8b858982058a3")
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, BNAndPadLayer: 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 torch.nn.functional as F

class BNAndPadLayer(nn.Module):

    def __init__(self, pad_pixels, num_features, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True):
        super(BNAndPadLayer, self).__init__()
        self.bn = nn.BatchNorm2d(num_features, eps, momentum, affine, track_running_stats)
        self.pad_pixels = pad_pixels

    def forward(self, input):
        output = self.bn(input)
        if self.pad_pixels > 0:
            if self.bn.affine:
                pad_values = self.bn.bias.detach() - self.bn.running_mean * self.bn.weight.detach() / torch.sqrt(self.bn.running_var + self.bn.eps)
            else:
                pad_values = -self.bn.running_mean / torch.sqrt(self.bn.running_var + self.bn.eps)
            output = F.pad(output, [self.pad_pixels] * 4)
            pad_values = pad_values.view(1, -1, 1, 1)
            output[:, :, 0:self.pad_pixels, :] = pad_values
            output[:, :, -self.pad_pixels:, :] = pad_values
            output[:, :, :, 0:self.pad_pixels] = pad_values
            output[:, :, :, -self.pad_pixels:] = pad_values
        return output

    @property
    def weight(self):
        return self.bn.weight

    @property
    def bias(self):
        return self.bn.bias

    @property
    def running_mean(self):
        return self.bn.running_mean

    @property
    def running_var(self):
        return self.bn.running_var

    @property
    def eps(self):
        return self.bn.eps


torch.manual_seed(0)

ARGS = {
    'pad_pixels': 2,
    'num_features': 4,
    'eps': 1e-05,
    'momentum': 0.1,
    'affine': True,
    'track_running_stats': 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 BNAndPadLayer — 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 BNAndPadLayer 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
ConvBN Ran
ran on a synthesized input · constructed an object; no result was computed
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("9dfb66e9d6efe79c")
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, ConvBN: 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 = {
    "in_channels": 3,
    "out_channels": 16,
    "kernel_size": 3,
    "stride": 1,
    "padding": 1,
    "dilation": 1,
    "groups": 1,
    "bias": False,
    "bn": nn.BatchNorm2d,
}
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 ConvBN — 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 ConvBN 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
conv_bn Ran
ran; our contract draft was wrong, not the code
DingXiaoH/DiverseBranchBlock/diversebranchblock.py
code served (permissive licence) · get_code("2d0d8fe75cfed19e")
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 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 = {
    'in_channels': 3,
    'out_channels': 64,
    'kernel_size': 3,
    'stride': 1,
    'padding': 1,
    'dilation': 1,
    'groups': 1,
    'padding_mode': 'zeros'
}
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 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
transIII_1x1_kxk Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("39738c9934012420")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Analyze the function to understand tensor requirements:
# transIII_1x1_kxk takes:
# - k1, b1: first convolution kernel and bias
# - k2, b2: second convolution kernel and bias
# - groups: number of groups (default 1)
#
# From the code:
# - k1.permute(1, 0, 2, 3): k1 is 4D (out_channels, in_channels, H, W)
# - k2: used in F.conv2d(k2, k1.permute(...)), so k2 is input, k1_T is filter
# - b1.reshape(1, -1, 1, 1): b1 is 1D with size matching k1's first dim
# - b2: 1D bias, added at end
# - With groups: k1.size(0) // groups, k2.size(0) // groups

# For groups=1 case:
# k1: (out_c1, in_c1, kh, kw) - will be permuted to (in_c1, out_c1, kh, kw)
# k2: (out_c2, in_c2, kh, kw) - in_c2 should match out_c1 after permute
# b1: (out_c1,)
# b2: (out_c2,)

# Let's use small distinct dimensions:
out_c1 = 4
in_c1 = 3
kh = 3
kw = 3
out_c2 = 2

k1 = torch.randn(out_c1, in_c1, kh, kw)  # (4, 3, 3, 3)
b1 = torch.randn(out_c1)  # (4,)
k2 = torch.randn(out_c2, out_c1, kh, kw)  # (2, 4, 3, 3)
b2 = torch.randn(out_c2)  # (2,)

ARGS = {
    'k1': k1,
    'b1': b1,
    'k2': k2,
    'b2': b2,
    'groups': 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 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
transII_addbranch Ran
ran; our contract draft was wrong, not the code
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("0ebbd01db07e9acb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes kernels and biases, and sums them along dim=0
# kernels and biases should be tensors
# kernels: shape (batch, ...) where we sum over batch dimension
# biases: shape (batch, ...) where we sum over batch dimension

# Create sample tensors with batch dimension first
# Using distinct sizes: batch=2, feature_dim=4, and another dim=3 for kernels
kernels = torch.randn(2, 4, 3)  # batch=2, 4 filters, 3 kernel size
biases = torch.randn(2, 4)      # batch=2, 4 biases

ARGS = {
    "kernels": kernels,
    "biases": biases,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
transIV_depthconcat Ran
ran; our contract draft was wrong, not the code
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("e8825a840b343a27")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function expects:
# - kernels: something that can be passed to torch.cat()
# - biases: something that can be passed to torch.cat()
# Both are typically lists or tuples of tensors

kernels = [
    torch.randn(2, 4),
    torch.randn(3, 4),
    torch.randn(2, 4)
]

biases = [
    torch.randn(2),
    torch.randn(3),
    torch.randn(2)
]

ARGS = {
    "kernels": kernels,
    "biases": biases
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
transIX_bn_to_1x1 Ran
ran; our contract draft was wrong, not the code
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("f89847a21ac13751")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Create a minimal BatchNorm2d module
bn = nn.BatchNorm2d(num_features=4)

# The function accesses: bn.weight, bn.bias, bn.running_mean, bn.running_var, bn.eps
# BatchNorm2d provides all of these by default

ARGS = {
    "bn": bn,
    "in_channels": 4,
    "groups": 1
}
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 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
transI_fusebn Ran
ran; our contract draft was wrong, not the code
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("e28e80243880eb6f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# kernel: a 4D tensor used in convolution (out_channels, in_channels, height, width)
kernel = torch.randn(4, 3, 2, 2)

# bn: a BatchNorm2d module with weight, bias, running_var, running_mean, eps
bn = nn.BatchNorm2d(4)
# Initialize with non-default values to make it realistic
bn.weight.data = torch.randn(4)
bn.bias.data = torch.randn(4)
bn.running_mean.data = torch.randn(4)
bn.running_var.data = torch.abs(torch.randn(4)) + 0.1  # ensure positive
bn.eps = 1e-5

# conv_bias: either None or a 1D tensor of size matching output channels
conv_bias = torch.randn(4)

ARGS = {
    "kernel": kernel,
    "bn": bn,
    "conv_bias": conv_bias
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
transVI_multiscale Ran
ran; our contract draft was wrong, not the code
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("a803f4daf16c141d")
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, 4, 7, 7) 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)

# Create a small kernel tensor
# kernel is used with .size(2) and .size(3), so it needs at least 4 dimensions
# Format: (batch, channels, height, width)
kernel = torch.randn(2, 4, 3, 5)

# target_kernel_size is an integer used in arithmetic operations
target_kernel_size = 7

ARGS = {
    "kernel": kernel,
    "target_kernel_size": target_kernel_size,
}
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, 4, 7, 7) — shape (2, 4, 7, 7) 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
transV_avg Ran
ran, honoured the contract we drafted
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("cfc36fe1d39f88c2")
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 (8, 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 torch
import numpy as np

torch.manual_seed(0)

ARGS = {
    "channels": 8,
    "kernel_size": 3,
    "groups": 2
}
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 Tensor, ndim 4, dtype family float, shape (8, 4, 3, 3) — shape (8, 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
DBB Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("3651747acdf99893")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet 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.

DiverseBranchBlock Not yet run
unverified (harvested, no recorded run)
DingXiaoH/DiverseBranchBlock/diversebranchblock.py
code served (permissive licence) · get_code("f5bd3fa9f5408fcc")
How it ran

no invocation recorded.

DiverseBranchBlock Not yet run
unverified (harvested, no recorded run)
hunto/image_classification_sota/lib/models/utils/dbb/dbb_block.py
code served (permissive licence) · get_code("27501fd33648e839")
How it ran

no invocation recorded.

IdentityBasedConv1x1 Not yet run
unverified (harvested, no recorded run)
DingXiaoH/DiverseBranchBlock/diversebranchblock.py
code served (permissive licence) · get_code("4982742829c08cfa")
How it ran

no invocation recorded.

fusion_1x1_kxk_down Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("e9b84f15940445d7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet 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.

fusion_asym Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("128e0f508f439f6e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet 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.

fusion_dbb Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("e27f3303da04b980")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet 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.

fusion_diff_kernel_size Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("1083b50ad0a305af")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet 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.

fusion_same_kernel_size Not yet run
unverified (harvested, no recorded run)
xdhe1216/ACNet/acnet_repvgg_dbb_block_utils.py
pointer only (licence: NONE) · get_code("87e546f4e96feae1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, xdhe1216/ACNet when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 19 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
BNAndPadLayer
brick of this paper's assembly: helper
1 other paper
1 later
2501.13492 (2025-01)
Quantized Spike-driven Transformer

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 propose a universal building block of Convolutional Neural Network (ConvNet) to improve the performance without any inference-time costs. The block is named Diverse Branch Block (DBB), which enhances the representational capacity of a single convolution by combining diverse branches of different scales and complexities to enrich the feature space, including sequences of convolutions, multiscale convolutions, and average pooling. After training, a DBB can be equivalently converted into a single conv layer for deployment. Unlike the advancements of novel Con-vNet architectures, DBB complicates the training-time microstructure while maintaining the macro architecture, so that it can be used as a drop-in replacement for regular conv layers of any architecture. In this way, the model can be trained to reach a higher level of performance and then transformed into the original inference-time structure for inference. DBB improves ConvNets on image classification (up to 1.9% higher top-1 accuracy on ImageNet), object detection and semantic segmentation. The PyTorch code and models are released at https://github.com/ DingXiaoH/DiverseBranchBlock.

For agents

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

get_harvested_code_for_paper("2103.13425")
get_code_for_paper("2103.13425")
have("2103.13425")

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

curl https://syntology.ai/api/ran/2103.13425.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.13425.svg)](https://syntology.ai/paper/2103.13425)

Connect an agent — have() is free.