SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.08955 · 2020

ResNeSt: Split-Attention Networks

arXiv · PDF · Open in the Atlas

Code that ran

28 of 48 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, 0 violated it and 17 have no contract verdict; 11 failures of our instrument; 20 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/2004.08955.json · badge: /api/ran/2004.08955.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
zhanghang1989/ResNeSt canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
zhanghang1989/PyTorch-Encoding canonical (classifier label)
classifier basis: classifier_2025
0 ran · 3 no recorded run
chongruo/detectron2-resnest community (classifier label)
classifier basis not recorded on the link
9 ran · 3 no recorded run
Burf/ResNeSt-Tensorflow2 community (classifier label)
classifier basis not recorded on the link
4 ran · 4 no recorded run
osmr/imgclsmob community (classifier label)
classifier basis not recorded on the link
4 ran · 3 no recorded run
ferna11i/detectron2_ResNeST community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
RobertHong1992/Resnest reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 2 no recorded run
ChengWeiGu/ResNeSt-Pytorch reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 1 no recorded run
copy not recorded community
not in the paper's declared repositories
0 ran · 1 no recorded run
mohitktanwr/Deep-Stem-ResNeSt-ISPRS community (classifier label)
classifier basis not recorded on the link
0 ran · 1 no recorded run
FunctionStatusWhere it lives
load_pkl_data Ran
ran; our contract draft was wrong, not the code
ChengWeiGu/ResNeSt-Pytorch/train_resneSt.py
pointer only (licence: NONE) · get_code("116a44af0a09b23d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ChengWeiGu/ResNeSt-Pytorch 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.

batched_nms Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/layers/nms.py
pointer only (licence: Apache-2.0) · get_code("0df9d3b583fbc6cb")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

batched_nms Ran
ran on a synthesized input
ferna11i/detectron2_ResNeST/detectron2/layers/nms.py
pointer only (licence: Apache-2.0) · get_code("a608182a325f07f7")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, ferna11i/detectron2_ResNeST, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

calc_block_num_params Ran
ran on a synthesized input
osmr/imgclsmob/gluon/model_stats.py
code served (permissive licence) · get_code("1e94f0e5aae421b2")
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 int64, 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 numpy as np
import torch

class Parameter:
    def __init__(self, shape, differentiable=True):
        self.shape = shape
        self._differentiable = differentiable

class Block:
    def __init__(self):
        self.params = {
            'weight1': Parameter(shape=(4, 3), differentiable=True),
            'weight2': Parameter(shape=(5, 2), differentiable=True),
            'bias': Parameter(shape=(3,), differentiable=True),
            'frozen': Parameter(shape=(2, 2), differentiable=False),
            'none_param': Parameter(shape=None, differentiable=True),
        }

block = Block()

ARGS = {
    'block': block,
}
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 int64, 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
calc_sha1 Ran
ran on a synthesized input
osmr/imgclsmob/prep_model.py
code served (permissive licence) · get_code("413e0014f62a9efc")
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 str. 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 tempfile
import os

# Create a temporary file with some content
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("Hello, this is test content for sha1 hashing.")
temp_file.close()
file_name = temp_file.name

# Define arguments for calc_sha1
ARGS = {
    'file_name': file_name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
cat Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
chongruo/detectron2-resnest/detectron2/layers/wrappers.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from xinyulyu/fgpl, is pointer only (licence: NONE) · get_code("6cbf46794989b731")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

cat Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ferna11i/detectron2_ResNeST/detectron2/layers/wrappers.py
pointer only (licence: Apache-2.0) · get_code("7664e656f28eeecc")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, ferna11i/detectron2_ResNeST, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

check_sha1 Ran
ran on a synthesized input
zhanghang1989/ResNeSt/resnest/utils.py
code served (permissive licence) · get_code("3ee0c4bb32b3b945")
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 hashlib
import tempfile
import os

# Create a temporary file with known content
temp_file = tempfile.NamedTemporaryFile(mode='wb', delete=False, suffix='.txt')
test_content = b"Hello, this is test content for sha1 hashing."
temp_file.write(test_content)
temp_file.close()

# Calculate the sha1 hash of the test content
sha1 = hashlib.sha1()
sha1.update(test_content)
expected_hash = sha1.hexdigest()

ARGS = {
    'filename': temp_file.name,
    'sha1_hash': expected_hash
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
convert_basic_c2_names Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/checkpoint/c2_model_loading.py
pointer only (licence: Apache-2.0) · get_code("e8526a516b4f9206")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

convert_c2_detectron_names Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/checkpoint/c2_model_loading.py
pointer only (licence: Apache-2.0) · get_code("52810b504f0d1066")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_checkpoint_url Ran
ran on a synthesized input
chongruo/detectron2-resnest/detectron2/model_zoo/model_zoo.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from tztztztztz/eql.detectron2, is pointer only (licence: Apache-2.0) · get_code("22f71a6d1391bbe3")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_model_file Ran
ran on a synthesized input
zhanghang1989/ResNeSt/resnest/gluon/model_store.py
code served (permissive licence) · get_code("1ef6246b268a5705")
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 str. 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 os
import tempfile
import shutil

# Mock the check_sha1 function since it's used in get_model_file
def check_sha1(file_path, sha1_hash):
    return True

# Mock the download function since it's used in get_model_file
def download(url, path, overwrite=False):
    # Create a dummy zip file at the path
    import zipfile
    os.makedirs(os.path.dirname(path), exist_ok=True)
    # Create a dummy .params file inside the zip
    file_name = os.path.basename(path).replace('.zip', '.params')
    with zipfile.ZipFile(path, 'w') as zf:
        zf.writestr(file_name, b'dummy model params')

# Inject mocks into the module namespace
import sys
current_module = sys.modules[__name__]
setattr(current_module, 'check_sha1', check_sha1)
setattr(current_module, 'download', download)

# Create a temporary directory for testing
temp_root = tempfile.mkdtemp()

# Define ARGS with valid parameters for get_model_file
ARGS = {
    'name': 'resnest50',
    'root': temp_root
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gluoncv — 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
group_conv Ran
ran on a synthesized input
Burf/ResNeSt-Tensorflow2/resnest/splat.py
code served (permissive licence) · get_code("21a473c9ae1aa4ca")
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 4, dtype family float, shape (2, 4, 4, 16) 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 test input tensor
# Shape: (batch_size, height, width, channels)
x = tf.constant(np.random.randn(2, 4, 4, 8).astype(np.float32))

# Define ARGS dict with all parameters
ARGS = {
    'x': x,
    'filters': 16,
    'kernel_size': [3, 5],
    'padding': 'same',
    'activation': 'relu',
    'name': 'test_group_conv'
}
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 4, dtype family float, shape (2, 4, 4, 16) — shape (2, 4, 4, 16) 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
interpolate Ran
ran on a synthesized input
chongruo/detectron2-resnest/detectron2/layers/wrappers.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from airsplay/py-bottom-up-attention, is pointer only (licence: Apache-2.0) · get_code("e3055ddd5ac36b0b")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

load_weight Ran
ran on a synthesized input
Burf/ResNeSt-Tensorflow2/resnest/resnest.py
code served (permissive licence) · get_code("aea5d02e03df9b82")
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 MinimalKerasModel. 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 tensorflow as tf
import numpy as np

torch.manual_seed(0)

# Create a minimal Keras model with weights that mimics ResNeSt structure
class MinimalKerasModel(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self._weights_list = []
        
        # Create some conv layer weights (4D: H, W, in_channels, out_channels)
        conv_weight = tf.Variable(np.random.randn(3, 3, 4, 8).astype(np.float32), name='conv.weight:0', trainable=True)
        self._weights_list.append(conv_weight)
        
        # Conv bias
        conv_bias = tf.Variable(np.random.randn(8).astype(np.float32), name='conv.bias:0', trainable=True)
        self._weights_list.append(conv_bias)
        
        # Downsample weight
        down_weight = tf.Variable(np.random.randn(1, 1, 8, 16).astype(np.float32), name='downsample.weight:0', trainable=True)
        self._weights_list.append(down_weight)
        
        # Downsample bias
        down_bias = tf.Variable(np.random.randn(16).astype(np.float32), name='downsample.bias:0', trainable=True)
        self._weights_list.append(down_bias)
        
        # FC layer weight
        fc_weight = tf.Variable(np.random.randn(10, 5).astype(np.float32), name='fc.weight:0', trainable=True)
        self._weights_list.append(fc_weight)
        
        # FC bias
        fc_bias = tf.Variable(np.random.randn(5).astype(np.float32), name='fc.bias:0', trainable=True)
        self._weights_list.append(fc_bias)
        
        # Group conv weight (4D with group dimension)
        group_weight = tf.Variable(np.random.randn(3, 3, 4, 16).astype(np.float32), name='group.weight:0', trainable=True)
        self._weights_list.append(group_weight)
    
    @property
    def weights(self):
        return self._weights_list

# Create a minimal Keras model instance
keras_model = MinimalKerasModel()

# Create a mock torch state dict with matching structure
torch_state_dict = {
    'conv.weight': torch.randn(8, 4, 3, 3),  # PyTorch format: out_channels, in_channels, H, W
    'conv.bias': torch.randn(8),
    'downsample.weight': torch.randn(16, 8, 1, 1),
    'downsample.bias': torch.randn(16),
    'fc.weight': torch.randn(5, 10),
    'group.weight': torch.randn(16, 4, 3, 3),
}

# Construct the arguments
ARGS = {
    'keras_model': keras_model,
    'torch_url': 'https://s3.us-west-1.wasabisys.com/resnest/torch/resnest50-528c19ca.pth',
    'group_size': 2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MinimalKerasModel — a MinimalKerasModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MinimalKerasModel 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_stage Ran
ran on a synthesized input
zhanghang1989/ResNeSt/resnest/d2/resnest.py
code served (permissive licence) · get_code("7e206a38c0ece70d")
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 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 as nn

torch.manual_seed(0)

# Define a minimal ResNetBlockBase subclass that the function expects
class ResNetBlockBase(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1):
        super().__init__()
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.stride = stride
        # Minimal implementation - just a simple conv layer
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=stride, padding=1)
    
    def forward(self, x):
        return self.conv(x)

# Create a concrete block class for testing
class TestResNetBlock(ResNetBlockBase):
    pass

ARGS = {
    'block_class': TestResNetBlock,
    'num_blocks': 3,
    'first_stride': 2,
    'in_channels': 64,
    'out_channels': 128,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, 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
nonzero_tuple Ran
ran on a synthesized input
ferna11i/detectron2_ResNeST/detectron2/layers/wrappers.py
pointer only (licence: Apache-2.0) · get_code("eb03e701fa2c3245")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, ferna11i/detectron2_ResNeST, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

pad_masks Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/layers/mask_ops.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from thinklab-sjtu/bench2drivezoo, is pointer only (licence: NOASSERTION) · get_code("406f84cf799ed34d")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

paste_mask_in_image_old Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/layers/mask_ops.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from thinklab-sjtu/bench2drivezoo, is pointer only (licence: NOASSERTION) · get_code("d03592674030077f")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

paste_masks_in_image Ran
ran on a synthesized input
ferna11i/detectron2_ResNeST/detectron2/layers/mask_ops.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from DIG-Beihang/ALLOW, is pointer only (licence: Apache-2.0) · get_code("5837267fa2aed31c")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, ferna11i/detectron2_ResNeST, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

paste_masks_in_image Ran
ran; our contract draft was wrong, not the code
chongruo/detectron2-resnest/detectron2/layers/mask_ops.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Jamie725/RGBT-detection, is pointer only (licence: Apache-2.0) · get_code("13b3f33534eea02a")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

post_process Ran
ran on a synthesized input
osmr/imgclsmob/prep_model.py
code served (permissive licence) · get_code("8f535a7e6722bc55")
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 os
import re
import hashlib
import tempfile
import shutil

# Create temporary directory for test files
temp_dir = tempfile.mkdtemp()
dst_dir = tempfile.mkdtemp()

# Create a dummy model file
model_file_path = os.path.join(temp_dir, "model.pth")
with open(model_file_path, 'wb') as f:
    f.write(b"dummy model content for testing")

# Create a dummy log file with proper format containing err-top5 metric
log_file_path = os.path.join(temp_dir, "training.log")
log_content = """epoch 1: loss=2.5
epoch 2: loss=2.3, err-top5=15.42
epoch 3: loss=2.1, err-top5=14.87"""

with open(log_file_path, 'w') as f:
    f.write(log_content)

ARGS = {
    "dst_dir_path": dst_dir,
    "model_name": "resnet50",
    "model_file_path": model_file_path,
    "log_file_path": log_file_path,
    "dst_model_file_ext": "pth",
    "log_line_num": 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
prepare_dst_model Ran
ran on a synthesized input
osmr/imgclsmob/convert_models.py
code served (permissive licence) · get_code("16e0587d2f35fcdf")
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 numpy as np
import torch
import torch.nn as nn

torch.manual_seed(0)

# Mock classes for different frameworks
class MockGluonNet:
    def _collect_params_with_prefix(self):
        return {'layer1.weight': None, 'layer1.bias': None}

class MockPyTorchNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv = nn.Conv2d(3, 16, kernel_size=3)
        self.bn = nn.BatchNorm2d(16)
        self.linear = nn.Linear(16, 10)
    
    def forward(self, x):
        return self.linear(self.bn(self.conv(x)))

class MockChainerNet:
    def namedparams(self):
        return [('layer1/W', np.zeros((3, 4))), ('layer1/b', np.zeros(4))]

class MockKerasLayer:
    def __init__(self, name, weights_list):
        self.name = name
        self.weights = weights_list

class MockKerasWeight:
    def __init__(self, name):
        self.name = name

class MockKerasNet:
    def __init__(self):
        w1 = MockKerasWeight('weight1:0')
        w2 = MockKerasWeight('weight2:0')
        self.weights = [w1, w2]
        self.layers = [
            MockKerasLayer('layer1', [w1]),
            MockKerasLayer('layer2', [w2])
        ]

class MockTF2Net:
    def __init__(self):
        self.in_size = (224, 224)
        self.data_format = 'channels_last'
        self.weights = []
    
    def __call__(self, x, x_len=None):
        return x

# Build mock prepare_model functions
def mock_prepare_gluon(model_name, use_pretrained, pretrained_model_file_path, dtype, tune_layers, classes, in_channels, ctx):
    return MockGluonNet()

def mock_prepare_pytorch(model_name, use_pretrained, pretrained_model_file_path, use_cuda, use_data_parallel, num_classes, in_channels):
    return MockPyTorchNet()

def mock_prepare_chainer(model_name, use_pretrained, pretrained_model_file_path):
    return MockChainerNet()

def mock_prepare_keras(model_name, use_pretrained, pretrained_model_file_path):
    return MockKerasNet()

def mock_prepare_tensorflow(model_name, use_pretrained, pretrained_model_file_path):
    return MockGluonNet()

def mock_prepare_tf2(model_name, use_pretrained, pretrained_model_file_path):
    return MockTF2Net()

# Patch the imports
import sys
from unittest.mock import MagicMock

sys.modules['gluon'] = MagicMock()
sys.modules['gluon.utils'] = MagicMock(prepare_model=mock_prepare_gluon)
sys.modules['pytorch'] = MagicMock()
sys.modules['pytorch.utils'] = MagicMock(prepare_model=mock_prepare_pytorch)
sys.modules['chainer_'] = MagicMock()
sys.modules['chainer_.utils'] = MagicMock(prepare_model=mock_prepare_chainer)
sys.modules['keras_'] = MagicMock()
sys.modules['keras_.utils'] = MagicMock(prepare_model=mock_prepare_keras)
sys.modules['tensorflow_'] = MagicMock()
sys.modules['tensorflow_.utils'] = MagicMock(prepare_model=mock_prepare_tensorflow)
sys.modules['tensorflow2'] = MagicMock()
sys.modules['tensorflow2.utils'] = MagicMock(prepare_model=mock_prepare_tf2)

ARGS = {
    'dst_fwk': 'pytorch',
    'dst_model': 'resnet18',
    'src_fwk': 'pytorch',
    'ctx': None,
    'use_cuda': False,
    'num_classes': 10,
    'in_channels': 3,
    'model_type': 'image'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
chainer_, gluon, keras_, numpy, pytorch, tensorflow, tensorflow2, tensorflow_, 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
resnest50 Ran
ran on a synthesized input
Burf/ResNeSt-Tensorflow2/resnest/resnest.py
code served (permissive licence) · get_code("b55f89f2f2fe736d")
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 Functional. 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 tensorflow as tf
import numpy as np

# Mock the ResNet function since it's referenced but not defined
def ResNet(img_input, depths, classes, include_top, radix, group_size, block_width, stem_width, deep_stem, avg_down, avd, avd_first):
    """Mock ResNet function that returns a tensor"""
    x = tf.keras.layers.GlobalAveragePooling2D()(img_input)
    x = tf.keras.layers.Dense(classes, activation='softmax')(x)
    return x

# Patch the ResNet into the module scope
import sys
current_module = sys.modules[__name__]
current_module.ResNet = ResNet

# Define ARGS for resnest50
ARGS = {
    'input_tensor': None,
    'input_shape': (224, 224, 3),
    'classes': 1000,
    'include_top': True,
    'weights': None,  # Use None to avoid attempting to download weights
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Functional — a Functional output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Functional 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
short_hash Ran
ran; our contract draft was wrong, not the code
RobertHong1992/Resnest/core/resnest.py
pointer only (licence: NONE) · get_code("979123ace229d553")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, RobertHong1992/Resnest 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.

short_hash Ran
ran; our contract draft was wrong, not the code
zhanghang1989/ResNeSt/resnest/gluon/model_store.py
code served (permissive licence) · get_code("b06092cd4eb27a10")
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 str. 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

_model_sha1 = {name: checksum for checksum, name in [('bcfefe1dd1dd1ef5cfed5563123c1490ea37b42e', 'resnest50'), ('5da943b3230f071525a98639945a6b3b3a45ac95', 'resnest101'), ('0c5d117df664ace220aa6fc2922c094bb079d381', 'resnest200'), ('11ae7f5da2bcdbad05ba7e84f9b74383e717f3e3', 'resnest269'), ('5e16dbe56f1fba8e1bc2faddd91f874bfbd74193', 'resnest50_fast_1s1x64d'), ('85eb779a5e313d74b5e5390dae02aa8082a0f469', 'resnest50_fast_2s1x64d'), ('3f215532c6d8e07a10df116309993d4479fc3e4b', 'resnest50_fast_4s1x64d'), ('af3514c2ec757a3a9666a75b82f142ed47d55bee', 'resnest50_fast_1s2x40d'), ('2db13245aa4967cf5e8617cb4911880dd41628a4', 'resnest50_fast_2s2x40d'), ('b24d515797832e02da4da9c8a15effd5e44cfb56', 'resnest50_fast_4s2x40d'), ('7318153ddb5e542a20cc6c58192f3c6209cff9ed', 'resnest50_fast_1s4x24d')]}

def short_hash(name):
    if name not in _model_sha1:
        raise ValueError('Pretrained model for {name} is not available.'.format(name=name))
    return _model_sha1[name][:8]

ARGS = {
    'name': 'resnest50'
}
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 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
split_attention_block Ran
ran on a synthesized input
Burf/ResNeSt-Tensorflow2/resnest/splat.py
code served (permissive licence) · get_code("8fc30ed2b4a12eb8")
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 4, dtype family float, shape (2, 4, 3, 16) 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

# Set random seed for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Create a small input tensor
# batch_size=2, height=4, width=3, channels=8
x = tf.random.normal([2, 4, 3, 8])

# Define arguments for split_attention_block
ARGS = {
    'x': x,
    'n_filter': 16,
    'kernel_size': 3,
    'stride_size': 1,
    'dilation': 1,
    'group_size': 2,
    'radix': 2,
    'dropout_rate': 0.1,
    'expansion': 4,
    'prefix': 'test'
}
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 4, dtype family float, shape (2, 4, 3, 16) — shape (2, 4, 3, 16) 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
unzip Ran
ran on a synthesized input
zhanghang1989/ResNeSt/resnest/utils.py
code served (permissive licence) · get_code("79a96a9c7fd65ce5")
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 str. 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 os
from pathlib import Path
import tempfile
import zipfile

# Create a temporary directory and a test zip file
temp_dir = tempfile.mkdtemp()
zip_path = os.path.join(temp_dir, "test.zip")

# Create a simple zip file with some test content
with zipfile.ZipFile(zip_path, 'w') as zf:
    zf.writestr("folder1/file1.txt", "content1")
    zf.writestr("folder1/file2.txt", "content2")

# Define the root directory where the zip will be extracted
extract_root = os.path.join(temp_dir, "extract")
os.makedirs(extract_root, exist_ok=True)

# Construct ARGS dict with all parameters
ARGS = {
    'zip_file_path': zip_path,
    'root': extract_root
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
ResNet Not yet run
unverified (harvested, no recorded run)
Burf/ResNeSt-Tensorflow2/resnest/resnet.py
code served (permissive licence) · get_code("8209cb3497010910")
How it ran

no invocation recorded.

accuracy Not yet run
unverified (harvested, no recorded run)
this paper's copy was not recorded; identical code first harvested from boone891214/MEST
pointer only; the stored copy get_code serves, from boone891214/MEST, is pointer only (licence: NONE) · get_code("913d82af53065418")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

batched_nms_rotated Not yet run
unverified (harvested, no recorded run)
ferna11i/detectron2_ResNeST/detectron2/layers/nms.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from DIG-Beihang/ALLOW, is pointer only (licence: Apache-2.0) · get_code("65735d51d64903a4")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, ferna11i/detectron2_ResNeST, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

batched_nms_rotated Not yet run
unverified (harvested, no recorded run)
chongruo/detectron2-resnest/detectron2/layers/nms.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from nvakhoa/superclass-fsis, is pointer only (licence: Apache-2.0) · get_code("4845964b76bba6c5")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

calc_block_num_params2 Not yet run
unverified (harvested, no recorded run)
osmr/imgclsmob/gluon/model_stats.py
code served (permissive licence) · get_code("45160e9713d39dfa")
How it ran

no invocation recorded.

change_output_layers Not yet run
unverified (harvested, no recorded run)
ChengWeiGu/ResNeSt-Pytorch/train_resneSt.py
pointer only (licence: NONE) · get_code("b74fa21880286e7a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ChengWeiGu/ResNeSt-Pytorch 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.

convert_pt2pt Not yet run
unverified (harvested, no recorded run)
osmr/imgclsmob/convert_models.py
code served (permissive licence) · get_code("cce15073ab760551")
How it ran

no invocation recorded.

download Not yet run
unverified (harvested, no recorded run)
zhanghang1989/ResNeSt/resnest/utils.py
code served (permissive licence) · get_code("eb12eb65d87b258c")
How it ran

no invocation recorded.

get_config_file Not yet run
unverified (harvested, no recorded run)
chongruo/detectron2-resnest/detectron2/model_zoo/model_zoo.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from tztztztztz/eql.detectron2, is pointer only (licence: Apache-2.0) · get_code("10d4b4b249a78f42")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_model_file Not yet run
unverified (harvested, no recorded run)
mohitktanwr/Deep-Stem-ResNeSt-ISPRS/resnest/gluon/model_store.py
code served (permissive licence) · get_code("2cd24f5ed4aae254")
How it ran

no invocation recorded.

nms_rotated Not yet run
unverified (harvested, no recorded run)
chongruo/detectron2-resnest/detectron2/layers/nms.py
pointer only (licence: Apache-2.0) · get_code("d3fa37a000ad294d")
How it ran

The body of this sample is withheld: Apache-2.0, as GitHub detects it on this paper's repository, chongruo/detectron2-resnest, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

prepare_src_model Not yet run
unverified (harvested, no recorded run)
osmr/imgclsmob/convert_models.py
code served (permissive licence) · get_code("df0046e7135787f9")
How it ran

no invocation recorded.

resnest101 Not yet run
unverified (harvested, no recorded run)
RobertHong1992/Resnest/core/resnest.py
pointer only (licence: NONE) · get_code("ab6e7a32f07c96d1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, RobertHong1992/Resnest 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.

resnest101 Not yet run
unverified (harvested, no recorded run)
zhanghang1989/PyTorch-Encoding/encoding/models/backbone/resnest.py
code served (permissive licence) · get_code("7b2de4494236346e")
How it ran

no invocation recorded.

resnest200 Not yet run
unverified (harvested, no recorded run)
zhanghang1989/PyTorch-Encoding/encoding/models/backbone/resnest.py
code served (permissive licence) · get_code("e9769d275b2aeaab")
How it ran

no invocation recorded.

resnest50 Not yet run
unverified (harvested, no recorded run)
RobertHong1992/Resnest/core/resnest.py
pointer only (licence: NONE) · get_code("d78c4213925c9d95")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, RobertHong1992/Resnest 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.

resnest50 Not yet run
unverified (harvested, no recorded run)
zhanghang1989/PyTorch-Encoding/encoding/models/backbone/resnest.py
code served (permissive licence) · get_code("6b650e26ef7115fd")
How it ran

no invocation recorded.

resnest_block Not yet run
unverified (harvested, no recorded run)
Burf/ResNeSt-Tensorflow2/resnest/resnet.py
code served (permissive licence) · get_code("2196103dc7afecbd")
How it ran

no invocation recorded.

resnest_module Not yet run
unverified (harvested, no recorded run)
Burf/ResNeSt-Tensorflow2/resnest/resnet.py
code served (permissive licence) · get_code("4d6569a50c288057")
How it ran

no invocation recorded.

rsoftmax Not yet run
unverified (harvested, no recorded run)
Burf/ResNeSt-Tensorflow2/resnest/splat.py
code served (permissive licence) · get_code("afc4cda18f695c8a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 48 samples lifted from this paper's linked repositories, 48 carry a function identity and 0 do not. Of the 47 distinct functions with an identity, 18 are also present in other papers whose code we hold (14 with at least one paper dated earlier, 8 the same month, 16 later) and 29 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
accuracy 80 other papers
31 dated earlier, 4 the same month, 45 later
1307.0414 (2013-07)
Challenges in Representation Learning: A report on three machine learning contests
1412.6553 (2014-12)
Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decomposition
1503.02531 (2015-03)
Distilling the Knowledge in a Neural Network
and 77 more not named here
convert_basic_c2_names 45 other papers
10 dated earlier, 1 the same month, 34 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1901.08164 (2019-01)
Decoupled Greedy Learning of CNNs
and 42 more not named here
convert_c2_detectron_names 29 other papers
9 dated earlier, 1 the same month, 19 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1901.08164 (2019-01)
Decoupled Greedy Learning of CNNs
and 26 more not named here
pad_masks 22 other papers
5 dated earlier, 1 the same month, 16 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
and 19 more not named here
paste_mask_in_image_old 22 other papers
5 dated earlier, 1 the same month, 16 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
and 19 more not named here
cat 18 other papers
4 dated earlier, 14 later · 3 only as an assembly brick
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
2002.10111 (2020-02)
SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation
and 15 more not named here
nms_rotated 14 other papers
5 dated earlier, 1 the same month, 8 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
and 11 more not named here
batched_nms 12 other papers
6 dated earlier, 1 the same month, 5 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
and 9 more not named here
batched_nms_rotated 12 other papers
6 dated earlier, 1 the same month, 5 later
1703.06870 (2017-03)
Mask R-CNN
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
and 9 more not named here
short_hash 11 other papers
8 dated earlier, 3 later
1803.08904 (2018-03)
Context Encoding for Semantic Segmentation
1809.02983 (2018-09)
Dual Attention Network for Scene Segmentation
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
and 8 more not named here
cat 6 other papers
6 later
2203.15175 (2022-03)
Unified Transformer Tracker for Object Tracking
2301.02657 (2023-01)
TarViS: A Unified Approach for Target-based Video Segmentation
2311.14671 (2023-11)
SEGIC: Unleashing the Emergent Correspondence for In-Context Segmentation
and 3 more not named here
paste_masks_in_image 5 other papers
2 dated earlier, 3 later
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
2003.05176 (2020-03)
Equalization Loss for Long-Tailed Object Recognition
2104.02904 (2021-04)
Multimodal Object Detection via Probabilistic Ensembling
and 2 more not named here
interpolate 4 other papers
3 dated earlier, 1 later
1804.08864 (2018-04)
Learning to See the Invisible: End-to-End Trainable Amodal Instance Segmentation
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
2003.05176 (2020-03)
Equalization Loss for Long-Tailed Object Recognition
and 1 more not named here
paste_masks_in_image 3 other papers
3 later
2007.08921 (2020-07)
Boundary-preserving Mask R-CNN
Xu_H2FA_R-CNN_Holistic_and_Hierarchical_Feature_Alignment_for_Cross-Domain_Weakly_CVPR_2022_paper (2022 (year only))
H 2 FA R-CNN: Holistic and Hierarchical Feature Alignment for Cross-domain Weakly Supervised Object Detection
batched_nms_rotated 2 other papers
2 later
Xu_H2FA_R-CNN_Holistic_and_Hierarchical_Feature_Alignment_for_Cross-Domain_Weakly_CVPR_2022_paper (2022 (year only))
H 2 FA R-CNN: Holistic and Hierarchical Feature Alignment for Cross-domain Weakly Supervised Object Detection
get_config_file 2 other papers
2 dated earlier
1908.07836 (2019-08)
PubLayNet: largest dataset ever for document layout analysis
2003.05176 (2020-03)
Equalization Loss for Long-Tailed Object Recognition
get_checkpoint_url 1 other paper
1 dated earlier
2003.05176 (2020-03)
Equalization Loss for Long-Tailed Object Recognition
short_hash 1 other paper
1 later
2007.05201 (2020-07)
ROSE: A Retinal OCT-Angiography Vessel Segmentation Dataset and New Model

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

More are linked than shown here.

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

Abstract

It is well known that featuremap attention and multi-path representation are important for visual recognition. In this paper, we present a modularized architecture, which applies the channel-wise attention on different network branches to leverage their success in capturing cross-feature interactions and learning diverse representations. Our design results in a simple and unified computation block, which can be parameterized using only a few variables. Our model, named ResNeSt, outperforms EfficientNet in accuracy and latency trade-off on image classification. In addition, ResNeSt has achieved superior transfer learning results on several public benchmarks serving as the backbone, and has been adopted by the winning entries of COCO-LVIS challenge. The source code for complete system and pretrained models are publicly available.

For agents

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

get_harvested_code_for_paper("2004.08955")
get_code_for_paper("2004.08955")
have("2004.08955")

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

curl https://syntology.ai/api/ran/2004.08955.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.08955.svg)](https://syntology.ai/paper/2004.08955)

Connect an agent — have() is free.