SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_28394 · AAAI · 2024

DTMFormer: Dynamic Token Merging for Boosting Transformer-Based Medical Image Segmentation

Yu Li, Kwang-Ting Cheng, Zhehao Wang, Xian Lin, Nannan Wu, Zengqiang Yan

Open in the Atlas

Code that ran

14 of 17 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 9 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_28394.json · badge: /api/ran/aaai_28394.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
iam-nacl/DTMFormer canonical (classifier label)
classifier basis: census
14 ran · 3 no recorded run
FunctionStatusWhere it lives
box_area Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/non_maximum_suppression.py
code served (permissive licence) · get_code("5a20e9de46fed2c3")
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 1, dtype family float, shape (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)

# boxes is a Tensor with shape (batch_size, 4)
# where each row represents [x1, y1, x2, y2] coordinates
# The function computes (x2 - x1) * (y2 - y1) for each box
boxes = torch.tensor([
    [0.0, 0.0, 4.0, 3.0],
    [1.0, 1.0, 5.0, 6.0]
], dtype=torch.float32)

ARGS = {
    "boxes": boxes
}
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 1, dtype family float, shape (2,) — shape (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
box_iou Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/non_maximum_suppression.py
code served (permissive licence) · get_code("c3725b35b6d0e285")
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, 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
from torch import Tensor

torch.manual_seed(0)

# Create two sets of bounding boxes in format [x1, y1, x2, y2]
# boxes1: 2 boxes
# boxes2: 3 boxes
boxes1 = torch.tensor([
    [0.0, 0.0, 4.0, 4.0],
    [1.0, 1.0, 5.0, 5.0]
], dtype=torch.float32)

boxes2 = torch.tensor([
    [0.0, 0.0, 3.0, 3.0],
    [2.0, 2.0, 6.0, 6.0],
    [5.0, 5.0, 7.0, 7.0]
], dtype=torch.float32)

ARGS = {
    "boxes1": boxes1,
    "boxes2": boxes2
}
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, 3) — shape (2, 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
dice_coefficient Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/metrics.py
code served (permissive licence) · get_code("456cc33b9d71d233")
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 0, dtype family float, shape () 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 numpy as np

torch.manual_seed(0)

# Create small binary tensors for predictions and ground truth
# Shape: (batch_size=2, spatial_dims=4)
pred = torch.randint(0, 2, (2, 4), dtype=torch.float32)
gt = torch.randint(0, 2, (2, 4), dtype=torch.float32)

ARGS = {
    "pred": pred,
    "gt": gt,
    "smooth": 1e-05
}
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 Tensor, ndim 0, dtype family float, shape () — shape () 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
flip Ran
ran; our contract draft was wrong, not the code
iam-nacl/DTMFormer/utils/tensor_utils.py
code served (permissive licence) · get_code("121824ff518da9ee")
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, 4, 5) 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 small tensor with distinct dimensions for probing
# Using shape (2, 3, 4, 5) to have distinct sizes across dimensions
x = torch.randn(2, 3, 4, 5)

# dim parameter - an integer specifying which dimension to flip
dim = 1

ARGS = {
    "x": x,
    "dim": dim,
}
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, 4, 5) — shape (2, 3, 4, 5) 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
flops_to_string Ran
ran; our contract draft was wrong, not the code
iam-nacl/DTMFormer/utils/flops_counter.py
pointer only (licence: MIT); the stored copy get_code serves, from rui-qian/SoTA-3D-Object-Detection, is pointer only (licence: NONE) · get_code("ad708b63b59159a6")
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 rui-qian/SoTA-3D-Object-Detection, 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.

index_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
iam-nacl/DTMFormer/models/model_utils.py
code served (permissive licence) · get_code("a1d91d4a1b5bbd6f")
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

torch.manual_seed(0)

# Create small tensors with distinct dimensions
# points: [B, N, C] where B=2 (batch), N=5 (num points), C=3 (channels)
points = torch.randn(2, 5, 3)

# idx: [B, S] where B=2 (batch), S=4 (num samples to index)
# Values should be valid indices into the N=5 dimension
idx = torch.tensor([[0, 2, 4, 1], [1, 3, 0, 2]], dtype=torch.long)

ARGS = {
    "points": points,
    "idx": idx,
}
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
mean_tensor Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/tensor_utils.py
code served (permissive licence) · get_code("89c15fb77f6b8f8f")
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
import numpy as np

torch.manual_seed(0)

# Create a small tensor to use as input
inp = torch.randn(2, 3, 4, 5)

# Define axes to reduce over
axes = np.array([1, 3])

# Define keepdim parameter
keepdim = False

ARGS = {
    "inp": inp,
    "axes": axes,
    "keepdim": keepdim
}
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 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
nms Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/non_maximum_suppression.py
code served (permissive licence) · get_code("c4f4dab1a6c80d83")
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 1, dtype family int, shape (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)

# boxes: Tensor[N, 4] with format (x1, y1, x2, y2)
# Create 2 boxes with distinct coordinates
boxes = torch.tensor([
    [10.0, 20.0, 50.0, 60.0],  # box 1
    [30.0, 40.0, 70.0, 80.0],  # box 2
], dtype=torch.float32)

# scores: Tensor[N] - confidence scores for each box
scores = torch.tensor([0.9, 0.7], dtype=torch.float32)

# iou_threshold: float - threshold for NMS filtering
iou_threshold = 0.5

ARGS = {
    "boxes": boxes,
    "scores": scores,
    "iou_threshold": iou_threshold,
}
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 1, dtype family int, shape (2,) — shape (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
norm_zscore Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/utils_gray.py
code served (permissive licence) · get_code("1a02dc2e586a5e97")
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 ndarray, ndim 2, dtype family float, shape (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 numpy as np

# Create a small test array with mixed zeros and positive values
# The function expects an array-like input that can be converted to np.array
test_data = np.array([
    [0, 1.5, 2.0, 0],
    [3.5, 0, 1.2, 4.1],
    [0, 0, 5.3, 2.8]
], dtype=np.float32)

ARGS = {
    "tx": test_data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 4) — shape (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
params_to_string Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/flops_counter.py
code served (permissive licence) · get_code("ddea91f7dddc66e6")
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 str. 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 numpy as np

# params_to_string takes params_num (required), units (optional), and precision (optional)
# The function does numerical operations on params_num and string formatting
# It checks if units is None, or equals 'M' or 'K'
# precision is used in round() calls when units is specified

ARGS = {
    'params_num': 2500000,  # A number large enough to trigger M conversion (2.5M)
    'units': None,          # Test default behavior
    'precision': 2          # Default precision value
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
sespiou_coefficient Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/metrics.py
code served (permissive licence) · get_code("b6e21886dcde716a")
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)

# Create small binary prediction and ground truth tensors
# Shape: (batch_size=2, spatial_dims=4)
pred = torch.randint(0, 2, (2, 4)).float()
gt = torch.randint(0, 2, (2, 4)).float()

ARGS = {
    "pred": pred,
    "gt": gt,
    "smooth": 1e-05
}
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
sespiou_coefficient2 Ran
ran on a synthesized input
iam-nacl/DTMFormer/utils/metrics.py
code served (permissive licence) · get_code("1cdc2a8ae6a38333")
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)

# The function expects:
# - pred: a tensor that will be reshaped to (N, -1) where N is batch size
# - gt: a tensor with same shape as pred, will be reshaped to (N, -1)
# - smooth: a float (default 1e-05)

# Create small binary prediction and ground truth tensors
# Shape: (batch_size=2, height=4, width=3) to have distinct dimensions
batch_size = 2
height = 4
width = 3

# Create binary predictions (values between 0 and 1)
pred = torch.rand(batch_size, height, width)
pred = (pred > 0.5).float()  # Convert to binary 0/1

# Create binary ground truth
gt = torch.rand(batch_size, height, width)
gt = (gt > 0.5).float()  # Convert to binary 0/1

# Smooth parameter
smooth = 1e-05

ARGS = {
    "pred": pred,
    "gt": gt,
    "smooth": smooth
}
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
sum_tensor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
iam-nacl/DTMFormer/utils/tensor_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from szc19990412/TransMIL, is pointer only (licence: NONE) · get_code("9e4e02bee8e90bea")
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 szc19990412/TransMIL, 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.

to_long_tensor Ran
ran, honoured the contract we drafted
iam-nacl/DTMFormer/utils/utils_gray.py
pointer only (licence: MIT); the stored copy get_code serves, from shashankhuang/reclmis, is pointer only (licence: NOASSERTION) · get_code("1f565d043b253a06")
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 shashankhuang/reclmis, 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.

cluster_dpc Not yet run
unverified (harvested, no recorded run)
iam-nacl/DTMFormer/models/model_utils.py
code served (permissive licence) · get_code("a30cf7667e54e598")
How it ran

no invocation recorded.

get_model_complexity_info Not yet run
unverified (harvested, no recorded run)
iam-nacl/DTMFormer/utils/flops_counter.py
code served (permissive licence) · get_code("03fe1448ec9d717b")
How it ran

no invocation recorded.

merge_tokens Not yet run
unverified (harvested, no recorded run)
iam-nacl/DTMFormer/models/model_utils.py
code served (permissive licence) · get_code("f5defd2df4c605bc")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 7 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 2 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.

FunctionAlso present inEarliest-datedOthers named
index_points 42 other papers
27 dated earlier, 4 later, 11 same year, order not determinable · 6 only as an assembly brick
1706.02413 (2017-06)
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
1811.07246 (2018-11)
PointConv: Deep Convolutional Networks on 3D Point Clouds
2003.13479 (2020-03)
RPM-Net: Robust Point Matching using Learned Features
and 39 more not named here
flops_to_string 5 other papers
5 dated earlier
2004.09576 (2020-04)
LSQ+: Improving low-bit quantization through learnable offsets and better initialization
2104.05707 (2021-04)
LocalViT: Analyzing Locality in Vision Transformers
2106.10823 (2021-06)
3D Object Detection for Autonomous Driving: A Survey
and 2 more not named here
flip 4 other papers
3 dated earlier, 1 same year, order not determinable
2002.00718 (2020-02)
Modeling the Background for Incremental Learning in Semantic Segmentation
2010.07092 (2020-10)
Data Augmentation for Meta-Learning
2210.07207 (2022-10)
Attribution-aware Weight Transfer: A Warm-Start Initialization for Class-Incremental Semantic Segmentation
and 1 more not named here
params_to_string 2 other papers
2 dated earlier
2104.05707 (2021-04)
LocalViT: Analyzing Locality in Vision Transformers
2212.00968 (2022-12)
UIU-Net: U-Net in U-Net for Infrared Small Object Detection
sum_tensor 2 other papers
1 later, 1 same year, order not determinable
2410.09299 (2024-10)
Hierarchical uncertainty estimation for learning-based registration in neuroimaging
2602.01951 (2026-02)
Enabling Progressive Whole-slide Image Analysis with Multi-scale Pyramidal Network
to_long_tensor 2 other papers
1 dated earlier, 1 same year, order not determinable
2203.04967 (2022-03)
UNeXt: MLP-based Rapid Medical Image Segmentation Network
2404.02845 (2024-04)
Cross-Modal Conditioned Reconstruction for Language-guided Medical Image Segmentation
get_model_complexity_info 1 other paper
1 dated earlier
2212.00968 (2022-12)
UIU-Net: U-Net in U-Net for Infrared Small Object Detection

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

Despite the great potential in capturing long-range dependency, one rarely-explored underlying issue of transformer in medical image segmentation is attention collapse, making it often degenerate into a bypass module in CNN-Transformer hybrid architectures. This is due to the high computational complexity of vision transformers requiring extensive training data while well-annotated medical image data is relatively limited, resulting in poor convergence. In this paper, we propose a plug-n-play transformer block with dynamic token merging, named DTMFormer, to avoid building long-range dependency on redundant and duplicated tokens and thus pursue better convergence. Specifically, DTMFormer consists of an attention-guided token merging (ATM) module to adaptively cluster tokens into fewer semantic tokens based on feature and dependency similarity and a light token reconstruction module to fuse ordinary and semantic tokens. In this way, as self-attention in ATM is calculated based on fewer tokens, DTMFormer is of lower complexity and more friendly to converge. Extensive experiments on publicly-available datasets demonstrate the effectiveness of DTMFormer working as a plug-n-play module for simultaneous complexity reduction and performance improvement. We believe it will inspire future work on rethinking transformers in medical image segmentation. Code: https://github.com/iam-nacl/DTMFormer.

For agents

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

get_harvested_code_for_paper("aaai_28394")
get_code_for_paper("aaai_28394")
have("aaai_28394")

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

curl https://syntology.ai/api/ran/aaai_28394.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_28394.svg)](https://syntology.ai/paper/aaai_28394)

Connect an agent — have() is free.