SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.00874 · 2024

Medical SAM 2: Segment medical images as video via Segment Anything Model 2

arXiv · PDF · Open in the Atlas

Code that ran

9 of 11 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, 1 violated it and 0 have no contract verdict; 7 failures of our instrument; 2 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/2408.00874.json · badge: /api/ran/2408.00874.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
medicinetoken/medical-sam2 canonical (classifier label)
classifier basis: census
9 ran · 2 no recorded run
FunctionStatusWhere it lives
compute_axial_cis Ran
ran, violated the contract we drafted
medicinetoken/medical-sam2/sam2_train/modeling/position_encoding.py
code served (permissive licence) · get_code("7ceaad42da42edd5")
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 (12, 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 = {
    "dim": 8,
    "end_x": 4,
    "end_y": 3,
    "theta": 10000.0,
}
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 (12, 4) — shape (12, 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
do_pool Ran
ran; our contract draft was wrong, not the code
medicinetoken/medical-sam2/sam2_train/modeling/backbones/hieradet.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from mark12ding/sam2long, is pointer only (licence: NOASSERTION) · get_code("d3df9ada3e05fa55")
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 mark12ding/sam2long, 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.

get_1d_sine_pe Ran
ran; our contract draft was wrong, not the code
medicinetoken/medical-sam2/sam2_train/modeling/sam2_utils.py
code served (permissive licence) · get_code("a6c3974188a5ffb0")
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, 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)

# pos_inds: a 1D tensor of position indices
pos_inds = torch.arange(2, dtype=torch.int64)

# dim: embedding dimension (should be even for sine/cosine split)
dim = 4

# temperature: temperature parameter for positional encoding (default 10000)
temperature = 10000

ARGS = {
    "pos_inds": pos_inds,
    "dim": dim,
    "temperature": temperature,
}
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, 4) — shape (2, 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
get_activation_fn Ran
ran; our contract draft was wrong, not the code
medicinetoken/medical-sam2/sam2_train/modeling/sam2_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from CGCL-codes/UAP-SAM2, is pointer only (licence: NONE) · get_code("f2e72d81f39d65cd")
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 CGCL-codes/UAP-SAM2, 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.

init_t_xy Ran
ran, honoured the contract we drafted
medicinetoken/medical-sam2/sam2_train/modeling/position_encoding.py
code served (permissive licence) · get_code("bb060d29525f90eb")
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, ndim 2, 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

torch.manual_seed(0)

ARGS = {
    'end_x': 4,
    'end_y': 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 tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
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
reshape_for_broadcast Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
medicinetoken/medical-sam2/sam2_train/modeling/position_encoding.py
code served (permissive licence) · get_code("f8f09f78b474ee7c")
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 (1, 3, 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)

# The function expects:
# - freqs_cis: a 2D tensor with shape (x.shape[-2], x.shape[-1])
# - x: a tensor with ndim >= 2

# Let's create x with shape (batch, seq_len, hidden_dim) = (2, 3, 4)
# This means x.shape[-2] = 3 and x.shape[-1] = 4
x = torch.randn(2, 3, 4)

# freqs_cis should have shape (3, 4) to match (x.shape[-2], x.shape[-1])
freqs_cis = torch.randn(3, 4)

ARGS = {
    "freqs_cis": freqs_cis,
    "x": x,
}
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 (1, 3, 4) — shape (1, 3, 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
select_closest_cond_frames Ran
ran; our contract draft was wrong, not the code
medicinetoken/medical-sam2/sam2_train/modeling/sam2_utils.py
code served (permissive licence) · get_code("4ee20f090e58c425")
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, ndim 1, 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 numpy as np

torch.manual_seed(0)

# Construct cond_frame_outputs as a dictionary mapping frame indices to output values
# The function uses cond_frame_outputs as a dict with integer keys (frame indices)
# and arbitrary values. We'll use tensors as stand-in output values.
cond_frame_outputs = {
    0: torch.randn(2, 4),
    5: torch.randn(2, 4),
    10: torch.randn(2, 4),
    15: torch.randn(2, 4),
    20: torch.randn(2, 4),
    25: torch.randn(2, 4),
}

# frame_idx is the current frame index we're selecting conditioning frames for
frame_idx = 12

# max_cond_frame_num is the maximum number of conditioning frames to select
# Use a value that requires selection logic (not -1 and not >= len(cond_frame_outputs))
max_cond_frame_num = 3

ARGS = {
    "frame_idx": frame_idx,
    "cond_frame_outputs": cond_frame_outputs,
    "max_cond_frame_num": max_cond_frame_num,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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_partition Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
medicinetoken/medical-sam2/sam2_train/modeling/backbones/utils.py
code served (permissive licence) · get_code("9b4931a54ea2c814")
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
import torch.nn.functional as F

torch.manual_seed(0)

# Create a small input tensor with shape [B, H, W, C]
# B=2 (batch), H=5 (height), W=7 (width), C=4 (channels)
x = torch.randn(2, 5, 7, 4)

# Window size - using a value that requires padding
window_size = 3

ARGS = {
    "x": x,
    "window_size": window_size,
}
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_unpartition Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
medicinetoken/medical-sam2/sam2_train/modeling/backbones/utils.py
code served (permissive licence) · get_code("4c236e8990640f4e")
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, 5, 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)

# The function performs window unpartition on a tensor
# windows: [B * num_windows, window_size, window_size, C]
# window_size: int
# pad_hw: (Hp, Wp) - padded height and width
# hw: (H, W) - original height and width

# Let's construct valid arguments:
# Choose window_size = 2
window_size = 2

# Choose padded dimensions: Hp=4, Wp=6 (distinct sizes)
Hp, Wp = 4, 6

# Choose original dimensions: H=3, W=5 (less than or equal to padded)
H, W = 3, 5

# Number of windows in each dimension
num_windows_h = Hp // window_size  # 4 // 2 = 2
num_windows_w = Wp // window_size  # 6 // 2 = 3
total_windows = num_windows_h * num_windows_w  # 6

# Batch size
B = 2

# Channel dimension
C = 4

# Total shape: [B * num_windows, window_size, window_size, C]
# = [2 * 6, 2, 2, 4] = [12, 2, 2, 4]
windows = torch.randn(B * total_windows, window_size, window_size, C)

ARGS = {
    "windows": windows,
    "window_size": window_size,
    "pad_hw": (Hp, Wp),
    "hw": (H, W),
}
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, 5, 4) — shape (2, 3, 5, 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
build_sam2 Not yet run
unverified (harvested, no recorded run)
medicinetoken/medical-sam2/sam2_train/build_sam.py
code served (permissive licence) · get_code("31a4fffe09fed1c5")
How it ran

no invocation recorded.

build_sam2_video_predictor Not yet run
unverified (harvested, no recorded run)
medicinetoken/medical-sam2/sam2_train/build_sam.py
code served (permissive licence) · get_code("989075a0ddf071e4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 10 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 10 the same month, 9 later) and 1 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
compute_axial_cis 13 other papers
2 dated earlier, 4 the same month, 7 later
2302.01872 (2023-02)
MOSE: A New Dataset for Video Object Segmentation in Complex Scenes
2306.01567 (2023-06)
Segment Anything in High Quality
2408.02635 (2024-08)
Interactive 3D Medical Image Segmentation with SAM 2
and 10 more not named here
init_t_xy 13 other papers
2 dated earlier, 4 the same month, 7 later
2302.01872 (2023-02)
MOSE: A New Dataset for Video Object Segmentation in Complex Scenes
2306.01567 (2023-06)
Segment Anything in High Quality
2408.02635 (2024-08)
Interactive 3D Medical Image Segmentation with SAM 2
and 10 more not named here
reshape_for_broadcast 13 other papers
2 dated earlier, 4 the same month, 7 later
2302.01872 (2023-02)
MOSE: A New Dataset for Video Object Segmentation in Complex Scenes
2306.01567 (2023-06)
Segment Anything in High Quality
2408.02635 (2024-08)
Interactive 3D Medical Image Segmentation with SAM 2
and 10 more not named here
get_activation_fn 10 other papers
2 dated earlier, 2 the same month, 6 later
2404.07445 (2024-04)
Multi-view Aggregation Network for Dichotomous Image Segmentation
2407.10749 (2024-07)
SEED: A Simple and Effective 3D DETR in Point Clouds
2408.07931 (2024-08)
Surgical SAM 2: Real-time Segment Anything in Surgical Video by Efficient Frame Pruning
and 7 more not named here
get_1d_sine_pe 9 other papers
3 the same month, 6 later · 1 only as an assembly brick
2408.00714 (2024-08)
SAM 2: Segment Anything in Images and Videos
2408.07931 (2024-08)
Surgical SAM 2: Real-time Segment Anything in Surgical Video by Efficient Frame Pruning
2408.08870 (2024-08)
SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation
and 6 more not named here
select_closest_cond_frames 9 other papers
3 the same month, 6 later · 1 only as an assembly brick
2408.00714 (2024-08)
SAM 2: Segment Anything in Images and Videos
2408.07931 (2024-08)
Surgical SAM 2: Real-time Segment Anything in Surgical Video by Efficient Frame Pruning
2408.08870 (2024-08)
SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation
and 6 more not named here
window_partition 7 other papers
4 dated earlier, 1 the same month, 1 later, 1 same year, order not determinable
2212.00280 (2022-12)
GRiT: A Generative Region-to-text Transformer for Object Understanding
2306.04121 (2023-06)
Matte Anything: Interactive Natural Image Matting with Segment Anything Models
aaai_28465 (2024 (year only))
Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation
and 4 more not named here
window_unpartition 7 other papers
4 dated earlier, 1 the same month, 1 later, 1 same year, order not determinable
2212.00280 (2022-12)
GRiT: A Generative Region-to-text Transformer for Object Understanding
2306.04121 (2023-06)
Matte Anything: Interactive Natural Image Matting with Segment Anything Models
aaai_28465 (2024 (year only))
Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation
and 4 more not named here
build_sam2 4 other papers
4 the same month
2408.02635 (2024-08)
Interactive 3D Medical Image Segmentation with SAM 2
2408.03286 (2024-08)
Biomedical SAM 2: Segment Anything in Biomedical Images and Videos
2408.08870 (2024-08)
SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation
and 1 more not named here
do_pool 2 other papers
1 the same month, 1 later
2408.08870 (2024-08)
SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation
2410.16268 (2024-10)
SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree

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

Medical image segmentation plays a pivotal role in clinical diagnostics and treatment planning, yet existing models often face challenges in generalization and in handling both 2D and 3D data uniformly. In this paper, we introduce Medical SAM 2 (MedSAM-2), a generalized auto-tracking model for universal 2D and 3D medical image segmentation. The core concept is to leverage the Segment Anything Model 2 (SAM2) pipeline to treat all 2D and 3D medical segmentation tasks as a video object tracking problem. To put it into practice, we propose a novel \emph{self-sorting memory bank} mechanism that dynamically selects informative embeddings based on confidence and dissimilarity, regardless of temporal order. This mechanism not only significantly improves performance in 3D medical image segmentation but also unlocks a \emph{One-Prompt Segmentation} capability for 2D images, allowing segmentation across multiple images from a single prompt without temporal relationships. We evaluated MedSAM-2 on five 2D tasks and nine 3D tasks, including white blood cells, optic cups, retinal vessels, mandibles, coronary arteries, kidney tumors, liver tumors, breast cancer, nasopharynx cancer, vestibular schwannoma, mediastinal lymph nodules, cerebral artery, inferior alveolar nerve, and abdominal organs, comparing it against state-of-the-art (SOTA) models in task-tailored, general and interactive segmentation settings. Our findings demonstrate that MedSAM-2 surpasses a wide range of existing models and updates new SOTA on several benchmarks. The code is released on the project page: https://supermedintel.github.io/Medical-SAM2/.

For agents

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

get_harvested_code_for_paper("2408.00874")
get_code_for_paper("2408.00874")
have("2408.00874")

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

curl https://syntology.ai/api/ran/2408.00874.json

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

[![Syntology run record](https://syntology.ai/api/ran/2408.00874.svg)](https://syntology.ai/paper/2408.00874)

Connect an agent — have() is free.