SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2111.06265 · NeurIPS · 2021

Dense Unsupervised Learning for Video Segmentation

Stefan Roth, Nikita Araslanov, Simone Schaub-Meyer

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2111.06265.json · badge: /api/ran/2111.06265.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
visinf/dense-ulearn-vos canonical (classifier label)
classifier basis: classifier_2025
6 ran · 0 no recorded run
FunctionStatusWhere it lives
BaseNet Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/dense-ulearn-vos/models/net.py
code served (permissive licence) · get_code("095e671ad6717b22")
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, BaseNet: 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 SimpleNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 8)
        self.bn1 = nn.BatchNorm2d(8)
        self.conv1 = nn.Conv2d(3, 8, 3, padding=1)
        self.linear2 = nn.Linear(8, 4)

simple_net = SimpleNet()

ARGS = {}
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 BaseNet — 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 BaseNet 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
MLP Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/dense-ulearn-vos/models/net.py
code served (permissive licence) · get_code("d7436bd8b38a68f7")
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, MLP: 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 = {
    'n_in': 4,
    'n_out': 3,
}
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 MLP — 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 MLP output the recorded ndim 1 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Net Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/dense-ulearn-vos/models/net.py
code served (permissive licence) · get_code("0329d5667de34d21")
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, Net: 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 MockBackbone(nn.Module):
    def __init__(self):
        super().__init__()
        self.fdim = 8
        self.conv1 = nn.Conv2d(3, 8, 3, padding=1)
        self.conv2 = nn.Conv2d(8, 16, 3, padding=1)
    
    def forward(self, x):
        res4 = self.conv1(x)
        res3 = self.conv2(x)
        return res4, res3

class MockConfig:
    def __init__(self):
        self.MODEL = type('obj', (object,), {'FEATURE_DIM': 4})()

cfg = MockConfig()
backbone = MockBackbone()

ARGS = {
    'cfg': cfg,
    'backbone': backbone
}
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 Net — 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 Net 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
convert_dict Ran
ran; our contract draft was wrong, not the code
visinf/dense-ulearn-vos/infer_vos.py
code served (permissive licence) · get_code("2445218fa1c1d604")
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 dict. 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)

# The function convert_dict takes a state_dict (a dictionary)
# and removes 'module.' prefix from keys.
# We need to create a realistic state_dict that might come from
# a DataParallel wrapped model.

state_dict = {
    'module.layer1.weight': torch.randn(4, 3),
    'module.layer1.bias': torch.randn(4),
    'module.layer2.weight': torch.randn(5, 4),
    'layer3.weight': torch.randn(2, 5),  # Some keys without 'module.' prefix
    'module.bn.running_mean': torch.randn(4),
}

ARGS = {
    'state_dict': state_dict
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
mask2rgb Ran
ran; our contract draft was wrong, not the code
visinf/dense-ulearn-vos/infer_vos.py
code served (permissive licence) · get_code("b46a0e1d2d74a9c9")
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 3, dtype family float, shape (2, 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

torch.manual_seed(0)

# Create a small palette callable that maps mask indices to RGB values
class SimplePalette(nn.Module):
    def __init__(self, num_classes=10):
        super().__init__()
        # Create a simple color palette: (num_classes, 4) where last dim is RGBA
        self.colors = torch.rand(num_classes, 4)
    
    def forward(self, mask):
        # mask is expected to be 2D (H, W) with class indices
        # Return (H, W, 4) with RGBA values
        h, w = mask.shape
        output = torch.zeros(h, w, 4)
        for i in range(h):
            for j in range(w):
                class_idx = int(mask[i, j].item()) % len(self.colors)
                output[i, j] = self.colors[class_idx]
        return output

# Create a small mask (2x4) with class indices
mask = torch.randint(0, 10, (2, 4))

# Create palette
palette = SimplePalette(num_classes=10)

ARGS = {
    "mask": mask,
    "palette": palette
}
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 3, dtype family float, shape (2, 4, 3) — shape (2, 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: none by the harness; any seed is in the setup block
mask_overlay Ran
ran; our contract draft was wrong, not the code
visinf/dense-ulearn-vos/infer_vos.py
code served (permissive licence) · get_code("bd7f22f1700017a4")
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 3, dtype family float, shape (2, 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

torch.manual_seed(0)

# Create a simple palette callable that maps mask indices to RGB values
class SimplePalette(nn.Module):
    def __init__(self, num_classes=10):
        super().__init__()
        # Create a color palette: num_classes x 4 (RGBA)
        self.colors = torch.rand(num_classes, 4)
    
    def forward(self, mask):
        # mask is expected to be 2D (H x W) with class indices
        # Return RGB values for each pixel
        mask_flat = mask.long().flatten()
        mask_rgb = self.colors[mask_flat]  # (H*W, 4)
        mask_rgb = mask_rgb.reshape(mask.shape[0], mask.shape[1], 4)
        return mask_rgb

# Create instances
palette = SimplePalette(num_classes=10)

# Create a small mask (2x4) with class indices
mask = torch.randint(0, 10, (2, 4))

# Create a small image (2x4x3) with values in [0, 1]
image = torch.rand(2, 4, 3)

ARGS = {
    "mask": mask,
    "image": image,
    "palette": palette,
}
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 3, dtype family float, shape (2, 4, 3) — shape (2, 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: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

We present a novel approach to unsupervised learning for video object segmentation (VOS). Unlike previous work, our formulation allows to learn dense feature representations directly in a fully convolutional regime. We rely on uniform grid sampling to extract a set of anchors and train our model to disambiguate between them on both inter-and intra-video levels. However, a naive scheme to train such a model results in a degenerate solution. We propose to prevent this with a simple regularisation scheme, accommodating the equivariance property of the segmentation task to similarity transformations. Our training objective admits efficient implementation and exhibits fast training convergence. On established VOS benchmarks, our approach exceeds the segmentation accuracy of previous work despite using significantly less training data and compute power. Code (Apache-2.0 License) available at https://github.com/visinf/dense-ulearn-vos. 35th Conference on Neural Information Processing Systems (NeurIPS 2021).

For agents

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

get_harvested_code_for_paper("2111.06265")
get_code_for_paper("2111.06265")
have("2111.06265")

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

curl https://syntology.ai/api/ran/2111.06265.json

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

[![Syntology run record](https://syntology.ai/api/ran/2111.06265.svg)](https://syntology.ai/paper/2111.06265)

Connect an agent — have() is free.