SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2409.16288 · ECCV · 2024

Self-Supervised Any-Point Tracking by Contrastive Random Walks

Andrew Owens, Ayush Shrivastava

arXiv · PDF · Open in the Atlas

Code that ran

7 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; 4 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 1 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/2409.16288.json · badge: /api/ran/2409.16288.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
ayshrv/gmrw canonical (classifier label)
classifier basis: census
7 ran · 1 no recorded run
FunctionStatusWhere it lives
coords_grid Ran
ran, honoured the contract we drafted
ayshrv/gmrw/models/gmflow_model/geometry.py
pointer only (licence: MIT); the stored copy get_code serves, from envision-research/driverecon, is pointer only (licence: NONE) · get_code("134980aa63271fbb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from envision-research/driverecon, 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.

generate_4d_window_grid Ran
ran, honoured the contract we drafted
ayshrv/gmrw/models/gmflow_model/geometry.py
code served (permissive licence) · get_code("125f42bf3edbd6fb")
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 5, dtype family float, shape (8, 12, 5, 7, 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)

ARGS = {
    "h_min": 0.0,
    "h_max": 10.0,
    "w_min": 0.0,
    "w_max": 10.0,
    "len_h": 5,
    "len_w": 7,
    "H": 8,
    "W": 12,
    "device": torch.device("cpu")
}
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 5, dtype family float, shape (8, 12, 5, 7, 4) — shape (8, 12, 5, 7, 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
generate_window_grid Ran
ran, honoured the contract we drafted
ayshrv/gmrw/models/gmflow_model/geometry.py
code served (permissive licence) · get_code("0823d84f9697f2a7")
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 3, dtype family float, shape (4, 3, 2) 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)

ARGS = {
    "h_min": 0.0,
    "h_max": 10.0,
    "w_min": 0.0,
    "w_max": 20.0,
    "len_h": 4,
    "len_w": 3,
    "device": torch.device("cpu"),
}
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 3, dtype family float, shape (4, 3, 2) — shape (4, 3, 2) 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_emb Ran
ran, honoured the contract we drafted
ayshrv/gmrw/models/gmflow_model/position.py
code served (permissive licence) · get_code("33805df68cb6892e")
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 2, dtype family float, shape (2, 8) 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)

# The function expects sin_inp to be a tensor that has .sin() and .cos() methods
# It will stack the results and flatten them
sin_inp = torch.randn(2, 4)

ARGS = {
    "sin_inp": sin_inp
}
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 2, dtype family float, shape (2, 8) — shape (2, 8) 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
merge_splits Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ayshrv/gmrw/models/gmflow_model/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from cvlab-kaist/PF3plat, is pointer only (licence: NOASSERTION) · get_code("6dc05be22af3598b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from cvlab-kaist/PF3plat, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

single_head_full_attention Ran
ran; our contract draft was wrong, not the code
ayshrv/gmrw/models/gmflow_model/transformer.py
pointer only (licence: MIT); the stored copy get_code serves, from ispc-lab/tma, is pointer only (licence: NONE) · get_code("52b500940a8025a3")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ispc-lab/tma, 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.

split_feature Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ayshrv/gmrw/models/gmflow_model/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from cvlab-kaist/PF3plat, is pointer only (licence: NOASSERTION) · get_code("4df5b1dd5d60d6cb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from cvlab-kaist/PF3plat, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

train_args Not yet run
unverified (harvested, no recorded run)
ayshrv/gmrw/arguments.py
code served (permissive licence) · get_code("0978cc604c978eb0")
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, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 6 later) and 2 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
get_emb 11 other papers
9 dated earlier, 2 later · 2 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
2211.13220 (2022-11)
TetraDiffusion: Tetrahedral Diffusion Models for 3D Shape Generation
2303.15269 (2023-03)
Handwritten Text Generation from Visual Archetypes
and 8 more not named here
single_head_full_attention 6 other papers
3 dated earlier, 3 later
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
2303.11629 (2023-03)
TMA: Temporal Motion Aggregation for Event-based Optical Flow
2403.05327 (2024-03)
DiffSF: Diffusion Models for Scene Flow Estimation
and 3 more not named here
coords_grid 5 other papers
2 dated earlier, 3 later
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
2312.04524 (2023-12)
RAVE: Randomized Noise Shuffling for Fast and Consistent Video Editing with Diffusion Models
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
and 2 more not named here
generate_window_grid 3 other papers
1 dated earlier, 2 later
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
merge_splits 3 other papers
1 dated earlier, 2 later
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
2410.22128 (2024-10)
PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
split_feature 3 other papers
1 dated earlier, 2 later
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
2410.22128 (2024-10)
PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion

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 simple, self-supervised approach to the Tracking Any Point (TAP) problem. We train a global matching transformer to find cycle consistent tracks through video via contrastive random walks, using the transformer's attention-based global matching to define the transition matrices for a random walk on a space-time graph. The ability to perform "all pairs" comparisons between points allows the model to obtain high spatial precision and to obtain a strong contrastive learning signal, while avoiding many of the complexities of recent approaches (such as coarse-to-fine matching). To do this, we propose a number of design decisions that allow global matching architectures to be trained through self-supervision using cycle consistency. For example, we identify that transformer-based methods are sensitive to shortcut solutions, and propose a data augmentation scheme to address them. Our method achieves strong performance on the TapVid benchmarks, outperforming previous self-supervised tracking methods, such as DIFT, and is competitive with several supervised methods.

For agents

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

get_harvested_code_for_paper("2409.16288")
get_code_for_paper("2409.16288")
have("2409.16288")

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

curl https://syntology.ai/api/ran/2409.16288.json

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

[![Syntology run record](https://syntology.ai/api/ran/2409.16288.svg)](https://syntology.ai/paper/2409.16288)

Connect an agent — have() is free.