SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.05210 · NeurIPS · 2021

Rethinking Space-Time Networks with Improved Memory Coverage for Efficient Video Object Segmentation

Ho Cheng, Yu-Wing Tai, Chi-Keung Tang, Decoder Decoder

arXiv · PDF · Open in the Atlas

Code that ran

6 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures of our instrument; 4 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/2106.05210.json · badge: /api/ran/2106.05210.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
limingxing00/rde-vos-cvpr2022 community
Papers with Code archive
6 ran · 1 no recorded run
alipga/AMM_VOS community
Papers with Code archive
0 ran · 2 no recorded run
hkchengrex/STCN marked official in Papers with Code
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
ASPP3D Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("812c10c240586d4c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, ASPP3D: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Define ARGS dict with all parameters for ASPP3D
ARGS = {
    'in_plane': 16,
    'out_plane': 32,
    'reduction': 4
}
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 ASPP3D — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ASPP3D 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
NONLocalBlock3D Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("c5daadfa598c8697")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, NONLocalBlock3D: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# NONLocalBlock3D requires:
# - in_channels: int (number of input channels)
# - inter_channels: int or None (intermediate channels, optional)
# - sub_sample: bool (whether to subsample)
# - bn_layer: bool (whether to use batch norm layer)

ARGS = {
    'in_channels': 16,
    'inter_channels': 8,
    'sub_sample': True,
    'bn_layer': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type NONLocalBlock3D — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a NONLocalBlock3D 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
SEBasicBlock Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("01ec14b944f4999c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, SEBasicBlock: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a small downsample module that matches the expected interface
class SimpleDownsample(nn.Module):
    def __init__(self, in_channels, out_channels):
        super(SimpleDownsample, self).__init__()
        self.conv = nn.Conv3d(in_channels, out_channels, kernel_size=1)
        self.norm = nn.InstanceNorm3d(out_channels)
    
    def forward(self, x):
        return self.norm(self.conv(x))

# Define arguments for SEBasicBlock
ARGS = {
    'inplanes': 4,
    'planes': 4,
    'stride': 1,
    'downsample': None,
    'reduction': 8
}
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 SEBasicBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SEBasicBlock 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
SELayerS Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("9bae3aa52ec50d1a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, SELayerS: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# SELayerS expects:
# - channel: int (number of channels)
# - reduction: int (optional, default=16)

# Looking at the forward method:
# - x has shape (b, c, d, h, w) where it's a 5D tensor
# - avg_pool reduces to (b, 2*c, 2, 1, 1)
# - fc processes (b, 2*c) and outputs (b, 2*c)
# - final output is x * y.expand_as(x)

# Create a small valid channel size
channel = 4

# Create a small 5D input tensor for testing
# Shape: (batch=2, channels=4, depth=3, height=5, width=6)
x = torch.randn(2, channel, 3, 5, 6)

ARGS = {
    'channel': channel,
    'reduction': 16
}
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 SELayerS — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SELayerS 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
_ASPPModule3D Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("698c6d37906b6b77")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, _ASPPModule3D: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class _ASPPModule3D(nn.Module):

    def __init__(self, inplanes, planes, kernel_size, padding, dilation):
        super(_ASPPModule3D, self).__init__()
        self.atrous_conv = nn.Conv3d(inplanes, planes, kernel_size=kernel_size, stride=1, padding=padding, dilation=dilation, bias=False)

    def forward(self, x):
        x = self.atrous_conv(x)
        return torch.nn.functional.relu(x, inplace=True)


ARGS = {
    'inplanes': 16,
    'planes': 32,
    'kernel_size': 3,
    'padding': 2,
    'dilation': 2
}
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 _ASPPModule3D — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a _ASPPModule3D 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
_NonLocalBlockND Ran
ran on a synthesized input · constructed an object; no result was computed
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("42805181cd60008e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, _NonLocalBlockND: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "in_channels": 16,
    "inter_channels": None,
    "dimension": 3,
    "sub_sample": True,
    "bn_layer": False,
}
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 _NonLocalBlockND — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a _NonLocalBlockND 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
MemoryReader Not yet run
unverified (harvested, no recorded run)
hkchengrex/STCN/model/network.py
code served (permissive licence) · get_code("6618c26355cb868a")
How it ran

no invocation recorded.

MemoryReader Not yet run
unverified (harvested, no recorded run)
alipga/AMM_VOS/model/network.py
pointer only (licence: NONE) · get_code("fa9203b36ccb7cee")
How it ran

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

MyDropout Not yet run
unverified (harvested, no recorded run)
alipga/AMM_VOS/model/network.py
pointer only (licence: NONE) · get_code("aed6d31ae6227bce")
How it ran

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

SAM Not yet run
unverified (harvested, no recorded run)
limingxing00/rde-vos-cvpr2022/model/modules.py
code served (permissive licence) · get_code("2a010015f43898d9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 10 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

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

Repositories linked to this paper

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

Abstract

This paper presents a simple yet effective approach to modeling space-time correspondences in the context of video object segmentation. Unlike most existing approaches, we establish correspondences directly between frames without reencoding the mask features for every object, leading to a highly efficient and robust framework. With the correspondences, every node in the current query frame is inferred by aggregating features from the past in an associative fashion. We cast the aggregation process as a voting problem and find that the existing inner-product affinity leads to poor use of memory with a small (fixed) subset of memory nodes dominating the votes, regardless of the query. In light of this phenomenon, we propose using the negative squared Euclidean distance instead to compute the affinities. We validate that every memory node now has a chance to contribute, and experimentally show that such diversified voting is beneficial to both memory efficiency and inference accuracy. The synergy of correspondence networks and diversified voting works exceedingly well, achieves new state-of-the-art results on both DAVIS and YouTubeVOS datasets while running significantly faster at 20+ FPS for multiple objects without bells and whistles.

For agents

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

get_harvested_code_for_paper("2106.05210")
get_code_for_paper("2106.05210")
have("2106.05210")

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

curl https://syntology.ai/api/ran/2106.05210.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.05210.svg)](https://syntology.ai/paper/2106.05210)

Connect an agent — have() is free.