SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_25201 · AAAI · 2023

InstanceFormer: An Online Video Instance Segmentation Framework

Rajat Koner, Volker Tresp, Thomas Seidl, Matthias Schubert, Suprosanna Shit, Sahand Sharifzadeh, Tanveer Hannan

Open in the Atlas

Code that ran

5 of 8 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 4 failures of our instrument; 3 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/aaai_25201.json · badge: /api/ran/aaai_25201.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
rajatkoner08/InstanceFormer canonical (classifier label)
classifier basis: census
5 ran · 3 no recorded run
FunctionStatusWhere it lives
aligned_bilinear Ran
ran; our contract draft was wrong, not the code
rajatkoner08/InstanceFormer/models/online_segmentation.py
code served (permissive licence) · get_code("29dbbaafa4be18e1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 8, 10) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create a 4D tensor with shape (batch, channels, height, width)
# Using distinct sizes: batch=2, channels=3, height=4, width=5
tensor = torch.randn(2, 3, 4, 5)

# Factor must be >= 1 and an integer
factor = 2

ARGS = {
    "tensor": tensor,
    "factor": factor,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 3, 8, 10) — shape (2, 3, 8, 10) 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
compute_locations Ran
ran, honoured the contract we drafted
rajatkoner08/InstanceFormer/models/online_segmentation.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from bo-miao/refhuman, is pointer only (licence: NONE) · get_code("ed1a308adb65f645")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from bo-miao/refhuman, and GitHub detected no licence in that repository 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.

parse_dynamic_params Ran
ran; our contract draft was wrong, not the code
rajatkoner08/InstanceFormer/models/online_segmentation.py
code served (permissive licence) · get_code("46f34dc9f9a94b67")
How it ran

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

import torch

torch.manual_seed(0)

# Analyze the function to understand parameter requirements:
# - params: 2D tensor, size(0) = num_insts, size(1) = sum(weight_nums) + sum(bias_nums)
# - channels: int, used for reshaping
# - weight_nums: list of ints, specifies weight dimensions for each layer
# - bias_nums: list of ints, specifies bias dimensions for each layer

# Set up small concrete values
num_insts = 2
channels = 4
num_layers = 2

# Define weight and bias numbers for 2 layers
# Layer 0 (not last): weight has some size, bias has some size
# Layer 1 (last): weight has some size, bias has some size
weight_nums = [12, 8]  # weight dimensions for each layer
bias_nums = [4, 1]     # bias dimensions for each layer

# params tensor must have:
# - dim 0: num_insts = 2
# - dim 1: sum(weight_nums) + sum(bias_nums) = 12 + 8 + 4 + 1 = 25
total_params = sum(weight_nums) + sum(bias_nums)
params = torch.randn(num_insts, total_params)

ARGS = {
    "params": params,
    "channels": channels,
    "weight_nums": weight_nums,
    "bias_nums": bias_nums,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
window_partition Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rajatkoner08/InstanceFormer/models/swin_transformer.py
code served (permissive licence) · get_code("f9fd6241d935f07b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (12, 4, 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

torch.manual_seed(0)

# Create a tensor with shape (B, H, W, C)
# B = 2 (batch size)
# H = 8 (height, must be divisible by window_size)
# W = 12 (width, must be divisible by window_size)
# C = 4 (channels)
# window_size = 4

B = 2
H = 8
W = 12
C = 4
window_size = 4

x = torch.randn(B, H, W, C)

ARGS = {
    'x': x,
    'window_size': window_size
}
Setup block drafted by
a language model, Claude Sonnet 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 (12, 4, 4, 4) — shape (12, 4, 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: none by the harness; any seed is in the setup block
window_reverse Ran
ran; our contract draft was wrong, not the code
rajatkoner08/InstanceFormer/models/swin_transformer.py
code served (permissive licence) · get_code("fb32094c6dbece71")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 8, 12, 6) 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)

# Define parameters
# We need windows of shape (num_windows*B, window_size, window_size, C)
# where num_windows = (H // window_size) * (W // window_size)

B = 2
H = 8
W = 12
window_size = 4
C = 6

# Calculate num_windows
num_windows_h = H // window_size  # 8 // 4 = 2
num_windows_w = W // window_size  # 12 // 4 = 3
num_windows = num_windows_h * num_windows_w  # 2 * 3 = 6

# Total number of windows across batch
total_windows = num_windows * B  # 6 * 2 = 12

# Create windows tensor
windows = torch.randn(total_windows, window_size, window_size, C)

ARGS = {
    'windows': windows,
    'window_size': window_size,
    'H': H,
    'W': W
}
Setup block drafted by
a language model, Claude Sonnet 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, 8, 12, 6) — shape (2, 8, 12, 6) 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_memories Not yet run
unverified (harvested, no recorded run)
rajatkoner08/InstanceFormer/models/segmentation_helper.py
code served (permissive licence) · get_code("1f846ff23aab600f")
How it ran

no invocation recorded.

resnext101_64x4d Not yet run
unverified (harvested, no recorded run)
rajatkoner08/InstanceFormer/models/x101_64d.py
code served (permissive licence) · get_code("1230270fac641da9")
How it ran

no invocation recorded.

unwrap Not yet run
unverified (harvested, no recorded run)
rajatkoner08/InstanceFormer/engine.py
code served (permissive licence) · get_code("49c287ddc66dfc8e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 5 later) and 3 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
window_reverse 227 other papers
74 dated earlier, 78 later, 75 same year, order not determinable · 34 only as an assembly brick
1608.03983 (2016-08)
SGDR: Stochastic Gradient Descent with Warm Restarts
1711.10925 (2017-11)
Deep Image Prior
2003.13678 (2020-03)
Designing Network Design Spaces
and 224 more not named here
window_partition 223 other papers
69 dated earlier, 76 later, 78 same year, order not determinable · 32 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
1711.10925 (2017-11)
Deep Image Prior
2003.13678 (2020-03)
Designing Network Design Spaces
and 220 more not named here
aligned_bilinear 17 other papers
5 dated earlier, 4 later, 8 same year, order not determinable · 1 only as an assembly brick
2201.00487 (2022-01)
Language as Queries for Referring Video Object Segmentation
2203.12827 (2022-03)
Sparse Instance Activation for Real-Time Instance Segmentation
2206.03789 (2022-06)
Language-Bridged Spatial-Temporal Interaction for Referring Video Object Segmentation
and 14 more not named here
compute_locations 12 other papers
2 dated earlier, 5 later, 5 same year, order not determinable · 1 only as an assembly brick
2201.00487 (2022-01)
Language as Queries for Referring Video Object Segmentation
2207.14012 (2022-07)
Video Mask Transfiner for High-Quality Video Instance Segmentation
2303.06674 (2023-03)
Universal Instance Perception as Object Discovery and Retrieval
and 9 more not named here
parse_dynamic_params 9 other papers
3 dated earlier, 3 later, 3 same year, order not determinable
2004.01177 (2020-04)
Tracking Objects as Points
2201.00487 (2022-01)
Language as Queries for Referring Video Object Segmentation
2207.14012 (2022-07)
Video Mask Transfiner for High-Quality Video Instance Segmentation
and 6 more not named here

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

Recent transformer-based offline video instance segmentation (VIS) approaches achieve encouraging results and significantly outperform online approaches. However, their reliance on the whole video and the immense computational complexity caused by full Spatio-temporal attention limit them in real-life applications such as processing lengthy videos. In this paper, we propose a single-stage transformerbased efficient online VIS framework named InstanceFormer, which is especially suitable for long and challenging videos. We propose three novel components to model short-term and long-term dependency and temporal coherence. First, we propagate the representation, location, and semantic information of prior instances to model short-term changes. Second, we propose a novel memory cross-attention in the decoder, which allows the network to look into earlier instances within a certain temporal window. Finally, we employ a temporal contrastive loss to impose coherence in the representation of an instance across all frames. Memory attention and temporal coherence are particularly beneficial to long-range dependency modeling, including challenging scenarios like occlusion. The proposed InstanceFormer outperforms previous online benchmark methods by a large margin across multiple datasets. Most importantly, InstanceFormer surpasses offline approaches for challenging and long datasets such as YouTube-VIS-2021 and OVIS. Code is available at https: //github.com/rajatkoner08/InstanceFormer.

For agents

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

get_harvested_code_for_paper("aaai_25201")
get_code_for_paper("aaai_25201")
have("aaai_25201")

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

curl https://syntology.ai/api/ran/aaai_25201.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_25201.svg)](https://syntology.ai/paper/aaai_25201)

Connect an agent — have() is free.