SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.06403 · CVPR · 2021

Lite-HRNet: A Lightweight High-Resolution Network

Lei Zhang, Lu Yuan, Jingdong Wang, Bin Xiao, Nong Sang, Changxin Gao, Changqian Yu

arXiv · PDF · Open in the Atlas

Code that ran

13 of 29 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 10 have no contract verdict; 3 failures of our instrument; 16 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/2104.06403.json · badge: /api/ran/2104.06403.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
zh320/realtime-semantic-segmentation-pytorch community
Papers with Code archive
9 ran · 5 no recorded run
viet-hoang-99/Lite_HRnet_vh community
Papers with Code archive
3 ran · 10 no recorded run
xiuyu0000/papers_with_examples community (classifier label)
classifier basis not recorded on the link
1 ran · 1 no recorded run
FunctionStatusWhere it lives
ConvModule Ran
ran on a synthesized input · constructed an object; no result was computed
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("7f5303fab6ab6aa5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

CrossResolutionWeightModule Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("82e1655d4701b970")
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 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 numpy as np

INIT_ARGS = {
    'channels': 4,
    'act_type': 'relu',
    'ch_reduction': 2,
    'pool_size': (3, 4)
}

# Create a list of 2 feature tensors with distinct spatial sizes
feats = [
    torch.randn(2, 4, 6, 8),  # batch=2, channels=4, H=6, W=8
    torch.randn(2, 4, 3, 4)   # batch=2, channels=4, H=3, W=4
]

ARGS = {
    'feats': feats
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
DSConvBNAct Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("7c4e84db2f49cedf")
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 4, dtype family float, shape (2, 8, 16, 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 torch
import torch.nn as nn

# Import the classes
class Activation(nn.Module):
    def __init__(self, act_type, **kwargs):
        super().__init__()
        activation_hub = {'relu': nn.ReLU, 'relu6': nn.ReLU6, 'leakyrelu': nn.LeakyReLU, 'prelu': nn.PReLU, 'celu': nn.CELU, 'elu': nn.ELU, 'hardswish': nn.Hardswish, 'hardtanh': nn.Hardtanh, 'gelu': nn.GELU, 'glu': nn.GLU, 'selu': nn.SELU, 'silu': nn.SiLU, 'sigmoid': nn.Sigmoid, 'softmax': nn.Softmax, 'tanh': nn.Tanh, 'none': nn.Identity}
        act_type = act_type.lower()
        if act_type not in activation_hub.keys():
            raise NotImplementedError(f'Unsupport activation type: {act_type}')
        self.activation = activation_hub[act_type](**kwargs)

    def forward(self, x):
        return self.activation(x)

class DWConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, act_type='relu', **kwargs):
        if isinstance(kernel_size, list) or isinstance(kernel_size, tuple):
            padding = ((kernel_size[0] - 1) // 2 * dilation, (kernel_size[1] - 1) // 2 * dilation)
        elif isinstance(kernel_size, int):
            padding = (kernel_size - 1) // 2 * dilation
        super().__init__(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, dilation=dilation, groups=in_channels, bias=False), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class PWConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, act_type='relu', bias=True, **kwargs):
        super().__init__(nn.Conv2d(in_channels, out_channels, 1, bias=bias), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class DSConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, act_type='relu', **kwargs):
        super().__init__(DWConvBNAct(in_channels, in_channels, kernel_size, stride, dilation, act_type, **kwargs), PWConvBNAct(in_channels, out_channels, act_type, **kwargs))

INIT_ARGS = {
    'in_channels': 4,
    'out_channels': 8,
    'kernel_size': 3,
    'stride': 1,
    'dilation': 1,
    'act_type': 'relu'
}

ARGS = {
    'input': torch.randn(2, 4, 16, 16)
}
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 4, dtype family float, shape (2, 8, 16, 16) — shape (2, 8, 16, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
DownsampleBlock Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("ae084989a6c34e09")
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 4, dtype family float, shape (2, 8, 4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

# Define the classes from the specification
class Activation(nn.Module):
    def __init__(self, act_type, **kwargs):
        super().__init__()
        activation_hub = {'relu': nn.ReLU, 'relu6': nn.ReLU6, 'leakyrelu': nn.LeakyReLU, 'prelu': nn.PReLU, 'celu': nn.CELU, 'elu': nn.ELU, 'hardswish': nn.Hardswish, 'hardtanh': nn.Hardtanh, 'gelu': nn.GELU, 'glu': nn.GLU, 'selu': nn.SELU, 'silu': nn.SiLU, 'sigmoid': nn.Sigmoid, 'softmax': nn.Softmax, 'tanh': nn.Tanh, 'none': nn.Identity}
        act_type = act_type.lower()
        if act_type not in activation_hub.keys():
            raise NotImplementedError(f'Unsupport activation type: {act_type}')
        self.activation = activation_hub[act_type](**kwargs)

    def forward(self, x):
        return self.activation(x)

class DWConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, act_type='relu', **kwargs):
        if isinstance(kernel_size, list) or isinstance(kernel_size, tuple):
            padding = ((kernel_size[0] - 1) // 2 * dilation, (kernel_size[1] - 1) // 2 * dilation)
        elif isinstance(kernel_size, int):
            padding = (kernel_size - 1) // 2 * dilation
        super().__init__(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, dilation=dilation, groups=in_channels, bias=False), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class PWConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, act_type='relu', bias=True, **kwargs):
        super().__init__(nn.Conv2d(in_channels, out_channels, 1, bias=bias), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class DSConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, act_type='relu', **kwargs):
        super().__init__(DWConvBNAct(in_channels, in_channels, kernel_size, stride, dilation, act_type, **kwargs), PWConvBNAct(in_channels, out_channels, act_type, **kwargs))

class DownsampleBlock(nn.Module):
    def __init__(self, in_ch, out_ch, num_block, act_type):
        super().__init__()
        assert num_block >= 1
        layers = []
        if num_block > 1:
            for i in range(num_block):
                hid_ch = in_ch if i != num_block - 1 else out_ch
                layers.append(DSConvBNAct(in_ch, hid_ch, 3, 2, act_type=act_type))
        else:
            layers.append(DSConvBNAct(in_ch, out_ch, 3, 2, act_type=act_type))
        self.conv = nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv(x)
        return x

INIT_ARGS = {
    'in_ch': 4,
    'out_ch': 8,
    'num_block': 2,
    'act_type': 'relu'
}

ARGS = {
    'x': torch.randn(2, 4, 16, 16)
}
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 4, dtype family float, shape (2, 8, 4, 4) — shape (2, 8, 4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
GlobalAvgPool2d Ran
ran on a synthesized input · constructed an object; no result was computed
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("9778e2200dd35fd6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

PWConvBNAct Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("c1585d202a6ecc57")
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 4, 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

INIT_ARGS = {
    'in_channels': 6,
    'out_channels': 8,
    'act_type': 'relu',
    'bias': True
}

# Create a sample input tensor with shape (batch_size, in_channels, height, width)
x = torch.randn(2, 6, 5, 7)

ARGS = {
    'input': x
}
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 4, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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)
RepresentationHead Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("8f3ca13d6f37ea95")
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 4, 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

# INIT_ARGS for RepresentationHead.__init__
INIT_ARGS = {
    'base_ch': 8,
    'num_class': 5,
    'num_stage': 3,
    'act_type': 'relu',
    'hid_ch': 16
}

# ARGS for RepresentationHead.forward
# feats is a list of feature maps from different stages
# Each stage i has channels = base_ch * (2^i)
# Stage 0: 8 channels, spatial size H x W
# Stage 1: 16 channels, spatial size H/2 x W/2
# Stage 2: 32 channels, spatial size H/4 x W/4

batch_size = 2
H, W = 12, 16  # Base spatial dimensions

feats = [
    torch.randn(batch_size, 8, H, W),      # Stage 0: base_ch * 2^0 = 8
    torch.randn(batch_size, 16, H//2, W//2),  # Stage 1: base_ch * 2^1 = 16
    torch.randn(batch_size, 32, H//4, W//4)   # Stage 2: base_ch * 2^2 = 32
]

ARGS = {
    'feats': feats
}
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 4, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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)
ShuffleBlock Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("4a4ad8085453946a")
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 4, dtype family float, shape (2, 8, 4, 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 torch.nn as nn

# Import the classes from the provided code
class Activation(nn.Module):
    def __init__(self, act_type, **kwargs):
        super().__init__()
        activation_hub = {'relu': nn.ReLU, 'relu6': nn.ReLU6, 'leakyrelu': nn.LeakyReLU, 'prelu': nn.PReLU, 'celu': nn.CELU, 'elu': nn.ELU, 'hardswish': nn.Hardswish, 'hardtanh': nn.Hardtanh, 'gelu': nn.GELU, 'glu': nn.GLU, 'selu': nn.SELU, 'silu': nn.SiLU, 'sigmoid': nn.Sigmoid, 'softmax': nn.Softmax, 'tanh': nn.Tanh, 'none': nn.Identity}
        act_type = act_type.lower()
        if act_type not in activation_hub.keys():
            raise NotImplementedError(f'Unsupport activation type: {act_type}')
        self.activation = activation_hub[act_type](**kwargs)

    def forward(self, x):
        return self.activation(x)

class ConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size=3, stride=1, dilation=1, groups=1, bias=False, act_type='relu', **kwargs):
        if isinstance(kernel_size, list) or isinstance(kernel_size, tuple):
            padding = ((kernel_size[0] - 1) // 2 * dilation, (kernel_size[1] - 1) // 2 * dilation)
        elif isinstance(kernel_size, int):
            padding = (kernel_size - 1) // 2 * dilation
        super().__init__(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, dilation, groups, bias), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class DWConvBNAct(nn.Sequential):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, act_type='relu', **kwargs):
        if isinstance(kernel_size, list) or isinstance(kernel_size, tuple):
            padding = ((kernel_size[0] - 1) // 2 * dilation, (kernel_size[1] - 1) // 2 * dilation)
        elif isinstance(kernel_size, int):
            padding = (kernel_size - 1) // 2 * dilation
        super().__init__(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, dilation=dilation, groups=in_channels, bias=False), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

def channel_shuffle(x, groups=2):
    batchsize, num_channels, height, width = x.size()
    channels_per_group = num_channels // groups
    x = x.view(batchsize, groups, channels_per_group, height, width)
    x = torch.transpose(x, 1, 2).contiguous()
    x = x.view(batchsize, -1, height, width)
    return x

class ShuffleBlock(nn.Module):
    def __init__(self, in_channels, out_channels, stride, act_type):
        super().__init__()
        assert stride in [1, 2]
        in_ch_l, out_ch_l = (in_channels // 2, out_channels // 2)
        in_ch_r, out_ch_r = (in_channels - in_ch_l, out_channels - out_ch_l)
        self.in_ch_l = in_ch_l
        if stride != 1 or in_ch_l != out_ch_l:
            self.left_branch = ConvBNAct(in_ch_l, out_ch_l, 1, stride, act_type=act_type)
        else:
            self.left_branch = nn.Identity()
        self.right_branch = nn.Sequential(ConvBNAct(in_ch_r, out_ch_r, 1, act_type=act_type), DWConvBNAct(out_ch_r, out_ch_r, 3, stride, act_type=act_type), ConvBNAct(out_ch_r, out_ch_r, 1, act_type=act_type))

    def forward(self, x):
        x_l = x[:, :self.in_ch_l]
        x_r = x[:, self.in_ch_l:]
        x_l = self.left_branch(x_l)
        x_r = self.right_branch(x_r)
        x = torch.cat([x_l, x_r], dim=1)
        x = channel_shuffle(x)
        return x

INIT_ARGS = {
    'in_channels': 8,
    'out_channels': 8,
    'stride': 1,
    'act_type': 'relu'
}

ARGS = {
    'x': torch.randn(2, 8, 4, 3)
}
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 4, dtype family float, shape (2, 8, 4, 3) — shape (2, 8, 4, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
SpatialWeightModule Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("50b9800a371ca617")
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 4, dtype family float, shape (2, 16, 1, 1) 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

class Activation(nn.Module):

    def __init__(self, act_type, **kwargs):
        super().__init__()
        activation_hub = {'relu': nn.ReLU, 'relu6': nn.ReLU6, 'leakyrelu': nn.LeakyReLU, 'prelu': nn.PReLU, 'celu': nn.CELU, 'elu': nn.ELU, 'hardswish': nn.Hardswish, 'hardtanh': nn.Hardtanh, 'gelu': nn.GELU, 'glu': nn.GLU, 'selu': nn.SELU, 'silu': nn.SiLU, 'sigmoid': nn.Sigmoid, 'softmax': nn.Softmax, 'tanh': nn.Tanh, 'none': nn.Identity}
        act_type = act_type.lower()
        if act_type not in activation_hub.keys():
            raise NotImplementedError(f'Unsupport activation type: {act_type}')
        self.activation = activation_hub[act_type](**kwargs)

    def forward(self, x):
        return self.activation(x)

class ConvBNAct(nn.Sequential):

    def __init__(self, in_channels, out_channels, kernel_size=3, stride=1, dilation=1, groups=1, bias=False, act_type='relu', **kwargs):
        if isinstance(kernel_size, list) or isinstance(kernel_size, tuple):
            padding = ((kernel_size[0] - 1) // 2 * dilation, (kernel_size[1] - 1) // 2 * dilation)
        elif isinstance(kernel_size, int):
            padding = (kernel_size - 1) // 2 * dilation
        super().__init__(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, dilation, groups, bias), nn.BatchNorm2d(out_channels), Activation(act_type, **kwargs))

class SpatialWeightModule(nn.Module):

    def __init__(self, channels, act_type, ch_reduction=8):
        super().__init__()
        hid_channels = channels // ch_reduction
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.fc = nn.Sequential(ConvBNAct(channels, hid_channels, 1, act_type=act_type), ConvBNAct(hid_channels, channels, 1, act_type='sigmoid'))

    def forward(self, x):
        x_avg = self.avg_pool(x)
        x_avg = self.fc(x_avg)
        return x_avg


INIT_ARGS = {
    'channels': 16,
    'act_type': 'relu',
    'ch_reduction': 8
}

ARGS = {
    'x': torch.randn(2, 16, 4, 3)
}
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 4, dtype family float, shape (2, 16, 1, 1) — shape (2, 16, 1, 1) 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)
UpsampleBlock Ran
ran on a synthesized input
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("21e87966c949f4a8")
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 4, 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

INIT_ARGS = {
    'in_ch': 8,
    'out_ch': 6,
    'scale_factor': 2,
    'act_type': 'relu'
}

ARGS = {
    'input': torch.randn(2, 8, 5, 7)
}
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 4, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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)
channel_shuffle Ran
ran; our contract draft was wrong, not the code
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("01a5b048271d7607")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

channel_shuffle Ran
ran; our contract draft was wrong, not the code
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("64e0d826e27a411e")
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 4, dtype family float, shape (2, 4, 3, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small tensor with shape (batch=2, channels=4, height=3, width=5)
# The function expects num_channels to be divisible by groups
x = torch.randn(2, 4, 3, 5)

ARGS = {
    "x": x,
    "groups": 2
}
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 4, dtype family float, shape (2, 4, 3, 5) — shape (2, 4, 3, 5) 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_ax Ran
ran; our contract draft was wrong, not the code
xiuyu0000/papers_with_examples/maskrcnn/src/infer.py
code served (permissive licence) · get_code("46bf80043a90eddf")
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 ndarray, ndim 2, dtype family float, shape (2, 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 matplotlib.pyplot as plt

ARGS = {
    "rows": 2,
    "cols": 3,
    "size": 16
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (2, 3) — shape (2, 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
CCWBlock Not yet run
unverified (harvested, no recorded run)
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("8b8d270632be0efe")
How it ran

no invocation recorded.

ConditionalChannelWeighting Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("b12bbf8bdd283fd4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

CrossResolutionWeighting Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("cf2130259d852422")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

DWConvBNAct Not yet run
unverified (harvested, no recorded run)
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("1656ec64a6120ea4")
How it ran

no invocation recorded.

DepthwiseSeparableConvModule Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("7228e155ce724b07")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

FusionBlock Not yet run
unverified (harvested, no recorded run)
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("29d5d4d24965174a")
How it ran

no invocation recorded.

IterativeHead Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("998cb1c170207bce")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

LiteHRModule Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("619eac8f9d49f5ae")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

LiteHRNet Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("a532cb9416bf3cef")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

LiteHRNet Not yet run
unverified (harvested, no recorded run)
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("8caec653733b1dc1")
How it ran

no invocation recorded.

ShuffleUnit Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("d77006bb3a2e1085")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

SpatialWeighting Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("bcd14fbd45f266e7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

StageBlock Not yet run
unverified (harvested, no recorded run)
zh320/realtime-semantic-segmentation-pytorch/models/lite_hrnet.py
code served (permissive licence) · get_code("6770080877034782")
How it ran

no invocation recorded.

Stem Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("63a7e1662cd623f7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

constant_init Not yet run
unverified (harvested, no recorded run)
viet-hoang-99/Lite_HRnet_vh/lib/models/lite_hrnet.py
pointer only (licence: NONE) · get_code("a10a9642969fb96e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, viet-hoang-99/Lite_HRnet_vh 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.

random_colors Not yet run
unverified (harvested, no recorded run)
xiuyu0000/papers_with_examples/maskrcnn/src/infer.py
code served (permissive licence) · get_code("255e1c60a7eead8b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 29 samples lifted from this paper's linked repositories, 29 carry a function identity and 0 do not. Of the 29 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 27 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
constant_init
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
2204.12484 (2022-04)
ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation
get_ax 1 other paper
1 dated earlier
1703.06870 (2017-03)
Mask R-CNN

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 present an efficient high-resolution network, Lite-HRNet, for human pose estimation. We start by simply applying the efficient shuffle block in ShuffleNet to HRNet (high-resolution network), yielding stronger performance over popular lightweight networks, such as MobileNet, ShuffleNet, and Small HRNet. We find that the heavily-used pointwise (1 × 1) convolutions in shuffle blocks become the computational bottleneck. We introduce a lightweight unit, conditional channel weighting, to replace costly pointwise (1 × 1) convolutions in shuffle blocks. The complexity of channel weighting is linear w.r.t the number of channels and lower than the quadratic time complexity for pointwise convolutions. Our solution learns the weights from all the channels and over multiple resolutions that are readily available in the parallel branches in HRNet. It uses the weights as the bridge to exchange information across channels and resolutions, compensating the role played by the pointwise (1 × 1) convolution. Lite-HRNet demonstrates superior results on human pose estimation over popular lightweight networks. Moreover, Lite-HRNet can be easily applied to semantic segmentation task in the same lightweight manner. The code and models have been publicly available at https://github.com/HRNet/Lite-HRNet.

For agents

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

get_harvested_code_for_paper("2104.06403")
get_code_for_paper("2104.06403")
have("2104.06403")

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

curl https://syntology.ai/api/ran/2104.06403.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.06403.svg)](https://syntology.ai/paper/2104.06403)

Connect an agent — have() is free.