SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.08541 · ICCV · 2021

TransVG: End-to-End Visual Grounding with Transformers

Wengang Zhou, Houqiang Li, Tianlang Chen, Zhengyuan Yang, Jiajun Deng

arXiv · PDF · Open in the Atlas

Code that ran

18 of 20 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; 0 honoured the contract we drafted, 0 violated it and 10 have no contract verdict; 8 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/2104.08541.json · badge: /api/ran/2104.08541.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
nku-shengzheliu/Pytorch-TransVG community (classifier label)
classifier basis not recorded on the link
18 ran · 2 no recorded run
FunctionStatusWhere it lives
GIoU_Loss Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/models/loss.py
pointer only (licence: MIT); the stored copy get_code serves, from csir1996/vln-gela, is pointer only (licence: NONE) · get_code("f1ea5386ce472284")
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 csir1996/vln-gela, 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.

Reg_Loss Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/models/loss.py
pointer only (licence: MIT); the stored copy get_code serves, from csir1996/vln-gela, is pointer only (licence: NONE) · get_code("33a4a4477537ad1c")
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 csir1996/vln-gela, 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.

all_gather Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/utils/misc.py
pointer only (licence: MIT); the stored copy get_code serves, from kellyiss/situformer, is pointer only (licence: NONE) · get_code("aa6da43d05cfaf65")
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 kellyiss/situformer, 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.

bbox_iou_numpy Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/utils.py
code served (permissive licence) · get_code("ec48fe4bebce9e11")
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, 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 numpy as np

# Create sample bounding boxes in format (x_min, y_min, x_max, y_max)
# box1: 3 bboxes
box1 = np.array([
    [10, 20, 50, 80],
    [15, 25, 60, 90],
    [5, 10, 40, 70]
], dtype=np.float32)

# box2: 2 bboxes
box2 = np.array([
    [20, 30, 70, 100],
    [8, 15, 45, 75]
], dtype=np.float32)

ARGS = {
    "box1": box1,
    "box2": box2
}
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, 2) — shape (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
build_vis_transformer Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/models/transformer.py
code served (permissive licence) · get_code("5ed44f7213d4ad07")
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 Transformer_vis. 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)

# Create a minimal args object with required attributes
class Args:
    def __init__(self):
        self.hidden_dim = 256
        self.dropout = 0.1
        self.nheads = 8
        self.dim_feedforward = 2048
        self.enc_layers = 6
        self.pre_norm = False

ARGS = {
    'args': 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 Transformer_vis — a Transformer_vis output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Transformer_vis 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
collate_fn Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/utils/misc.py
pointer only (licence: MIT); the stored copy get_code serves, from pinxueguo/x-prompt, is pointer only (licence: NONE) · get_code("3f82133185c8bcb4")
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 pinxueguo/x-prompt, 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.

convert_examples_to_features Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/dataset/data_loader.py
pointer only (licence: MIT); the stored copy get_code serves, from Marloweeee/BARE, is pointer only (licence: NONE) · get_code("694e413f977737ba")
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 Marloweeee/BARE, 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_coord Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/models/model.py
code served (permissive licence) · get_code("0336cb59d505cda3")
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, 8, 4, 3) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 = {
    "batch": 2,
    "height": 4,
    "width": 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 Tensor, ndim 4, dtype family float, shape (2, 8, 4, 3) — shape (2, 8, 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
letterbox Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/transforms.py
code served (permissive licence) · get_code("b202ce3cdac3a1ce")
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 cv2
import numpy as np

# Create a small test image (height=100, width=150, channels=3)
img = np.random.randint(0, 256, (100, 150, 3), dtype=np.uint8)

# Create a small test mask (height=100, width=150, single channel)
mask = np.random.randint(0, 256, (100, 150), dtype=np.uint8)

# Define height for letterbox operation
height = 224

# Define color tuple
color = (123.7, 116.3, 103.5)

ARGS = {
    "img": img,
    "mask": mask,
    "height": height,
    "color": color
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — 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
load_pretrain Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/checkpoint.py
code served (permissive licence) · get_code("8dc1056f0f6f5443")
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 MinimalModel. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import logging
import tempfile
import os

torch.manual_seed(0)

# Create a minimal model
class MinimalModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create model instance
model = MinimalModel()

# Create a temporary pretrain checkpoint file
temp_dir = tempfile.gettempdir()
pretrain_path = os.path.join(temp_dir, 'pretrain_checkpoint.pt')

# Save a checkpoint with matching state_dict keys
checkpoint = {
    'state_dict': model.state_dict().copy()
}
torch.save(checkpoint, pretrain_path)

# Create args object with pretrain attribute
class Args:
    def __init__(self, pretrain_path):
        self.pretrain = pretrain_path

args = Args(pretrain_path)

# Create a logger
logger = logging.getLogger('test_logger')
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
logger.addHandler(handler)

ARGS = {
    'model': model,
    'args': args,
    'logging': logger
}
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 MinimalModel — a MinimalModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MinimalModel 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
load_resume Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/checkpoint.py
code served (permissive licence) · get_code("44ac945d7427e3cc")
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 MinimalModel. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 os
import torch
import torch.nn as nn
import logging
import tempfile

torch.manual_seed(0)

# Create a minimal model
class MinimalModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create a minimal args object
class Args:
    def __init__(self, resume_path):
        self.resume = resume_path
        self.start_epoch = 0

# Create a temporary checkpoint file
model = MinimalModel()
state_dict = model.state_dict()

# Create temporary directory and checkpoint file
temp_dir = tempfile.gettempdir()
checkpoint_path = os.path.join(temp_dir, 'test_checkpoint.pth')

checkpoint = {
    'epoch': 5,
    'best_loss': 0.42,
    'state_dict': state_dict
}

torch.save(checkpoint, checkpoint_path)

# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Create args
args = Args(checkpoint_path)

# Create model instance for loading
model_for_loading = MinimalModel()

ARGS = {
    'model': model_for_loading,
    'args': args,
    'logging': logger
}
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 MinimalModel — a MinimalModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MinimalModel 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
load_weights Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/models/model.py
code served (permissive licence) · get_code("395306b13bd68afc")
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 SimpleModel. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import tempfile
import os

torch.manual_seed(0)

# Create a simple model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

# Create model instance
model = SimpleModel()

# Create a temporary file to save weights
temp_dir = tempfile.gettempdir()
load_path = os.path.join(temp_dir, "test_weights.pt")

# Create a checkpoint with trained weights (simulating a saved model)
trained_model = SimpleModel()
checkpoint = {
    'model': trained_model.state_dict()
}
torch.save(checkpoint, load_path)

ARGS = {
    'model': model,
    'load_path': load_path
}
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 SimpleModel — a SimpleModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModel 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
random_affine Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/transforms.py
code served (permissive licence) · get_code("408bfee1417b817e")
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 cv2
import numpy as np
import random
import math

# Set seed for reproducibility
random.seed(0)
np.random.seed(0)

# Create a small test image (height=100, width=100, 3 channels)
img = np.random.randint(0, 255, (100, 100, 3), dtype=np.uint8)

# Create a small test mask (same size as image)
mask = np.random.randint(0, 2, (100, 100), dtype=np.uint8) * 255

# Create a bounding box target as a 1D array [x1, y1, x2, y2]
targets = np.array([10.0, 10.0, 50.0, 60.0], dtype=np.float32)

# Alternative: create a list of bounding boxes for testing the list case
targets_list = [
    np.array([10.0, 10.0, 50.0, 60.0], dtype=np.float32),
    np.array([20.0, 20.0, 70.0, 80.0], dtype=np.float32)
]

# Create all_bbox array (multiple bounding boxes)
all_bbox = np.array([
    [15.0, 15.0, 55.0, 65.0],
    [25.0, 25.0, 75.0, 85.0]
], dtype=np.float32)

ARGS = {
    'img': img,
    'mask': mask,
    'targets': targets,
    'degrees': (-10, 10),
    'translate': (0.1, 0.1),
    'scale': (0.9, 1.1),
    'shear': (-2, 2),
    'borderValue': (123.7, 116.3, 103.5),
    'all_bbox': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — 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
read_examples Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/dataset/data_loader.py
pointer only (licence: MIT); the stored copy get_code serves, from Marloweeee/BARE, is pointer only (licence: NONE) · get_code("e4f50d4eeb6144e9")
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 Marloweeee/BARE, 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.

reduce_dict Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/utils/misc.py
pointer only (licence: MIT); the stored copy get_code serves, from kellyiss/situformer, is pointer only (licence: NONE) · get_code("05992505df9ef0cd")
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 kellyiss/situformer, 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.

wrap_points Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/transforms.py
code served (permissive licence) · get_code("cb5d2317201b6913")
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 1, dtype family float, shape (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
import math

# Create valid arguments for wrap_points function

# targets: array of 4 bounding box coordinates [x1, y1, x2, y2]
targets = np.array([10.0, 20.0, 100.0, 150.0])

# M: 3x3 transformation matrix (for affine transformation)
# The function uses M.T to transform points, so it needs to be a proper matrix
M = np.array([
    [1.0, 0.0, 0.0],
    [0.0, 1.0, 0.0],
    [5.0, 10.0, 1.0]
], dtype=np.float32)

# height: scalar value representing image height (used for clipping)
height = 200

# a: angle in degrees (used for rotation and area reduction calculation)
a = 30

ARGS = {
    'targets': targets,
    'M': M,
    'height': height,
    'a': a
}
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 1, dtype family float, shape (4,) — shape (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
xywh2xyxy Ran
ran on a synthesized input
nku-shengzheliu/Pytorch-TransVG/utils/utils.py
code served (permissive licence) · get_code("7950108b58fb67bc")
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)

# The function xywh2xyxy takes a tensor x with shape (batch_size, 4)
# where each row contains [x_center, y_center, width, height]
# and converts it to [x_min, y_min, x_max, y_max] format

x = torch.tensor([
    [10.0, 20.0, 4.0, 6.0],
    [30.0, 40.0, 8.0, 10.0]
], dtype=torch.float32)

ARGS = {
    "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 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
xyxy2xywh Ran
ran; our contract draft was wrong, not the code
nku-shengzheliu/Pytorch-TransVG/utils/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from csir1996/vln-gela, is pointer only (licence: NONE) · get_code("3d629255e4cf4ea2")
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 csir1996/vln-gela, 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.

build_VLFusion Not yet run
unverified (harvested, no recorded run)
nku-shengzheliu/Pytorch-TransVG/models/detr.py
code served (permissive licence) · get_code("75b3520299159feb")
How it ran

no invocation recorded.

build_transformer Not yet run
unverified (harvested, no recorded run)
nku-shengzheliu/Pytorch-TransVG/models/transformer.py
code served (permissive licence) · get_code("87992961bac26fc7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 15 are also present in other papers whose code we hold (12 with at least one paper dated earlier, 4 the same month, 9 later) and 5 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
reduce_dict 33 other papers
3 dated earlier, 1 the same month, 29 later
1811.07628 (2018-11)
ATOM: Accurate Tracking by Overlap Maximization
1811.10200 (2018-11)
IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments
2007.02419 (2020-07)
Attention-based Joint Detection of Object and Semantic Part
and 30 more not named here
build_transformer 32 other papers
2 dated earlier, 2 the same month, 28 later
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2103.12115 (2021-03)
End-to-End Trainable Multi-Instance Pose Estimation with Transformers
2104.04369 (2021-04)
Video-aided Unsupervised Grammar Induction
and 29 more not named here
all_gather 30 other papers
3 dated earlier, 1 the same month, 26 later
1811.07628 (2018-11)
ATOM: Accurate Tracking by Overlap Maximization
1811.10200 (2018-11)
IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments
2007.02419 (2020-07)
Attention-based Joint Detection of Object and Semantic Part
and 27 more not named here
collate_fn 16 other papers
1 dated earlier, 1 the same month, 14 later
1811.07628 (2018-11)
ATOM: Accurate Tracking by Overlap Maximization
2104.00969 (2021-04)
TubeR: Tubelet Transformer for Video Action Detection
2111.10135 (2021-11)
Grounded Situation Recognition with Transformers
and 13 more not named here
read_examples 7 other papers
2 dated earlier, 5 later
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
2106.03089 (2021-06)
Referring Transformer: A One-step Approach to Multi-task Visual Grounding
and 4 more not named here
convert_examples_to_features 6 other papers
2 dated earlier, 4 later
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
2106.03089 (2021-06)
Referring Transformer: A One-step Approach to Multi-task Visual Grounding
and 3 more not named here
bbox_iou_numpy 2 other papers
2 dated earlier
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
generate_coord 2 other papers
2 dated earlier
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
random_affine 2 other papers
2 dated earlier
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
wrap_points 2 other papers
2 dated earlier
1908.06354 (2019-08)
A Fast and Accurate One-Stage Approach to Visual Grounding
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
GIoU_Loss 1 other paper
1 later
2308.12587 (2023-08)
Grounded Entity-Landmark Adaptive Pre-training for Vision-and-Language Navigation
Reg_Loss 1 other paper
1 later
2308.12587 (2023-08)
Grounded Entity-Landmark Adaptive Pre-training for Vision-and-Language Navigation
load_pretrain 1 other paper
1 dated earlier
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
load_resume 1 other paper
1 dated earlier
2008.01059 (2020-08)
Improving One-stage Visual Grounding by Recursive Sub-query Construction
xyxy2xywh 1 other paper
1 later
2308.12587 (2023-08)
Grounded Entity-Landmark Adaptive Pre-training for Vision-and-Language Navigation

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

In this paper, we present a neat yet effective transformerbased framework for visual grounding, namely TransVG, to address the task of grounding a language query to the corresponding region onto an image. The state-of-the-art methods, including two-stage or one-stage ones, rely on a complex module with manually-designed mechanisms to perform the query reasoning and multi-modal fusion. However, the involvement of certain mechanisms in fusion module design, such as query decomposition and image scene graph, makes the models easily overfit to datasets with specific scenarios, and limits the plenitudinous interaction between the visual-linguistic context. To avoid this caveat, we propose to establish the multi-modal correspondence by leveraging transformers, and empirically show that the complex fusion modules (e.g., modular attention network, dynamic graph, and multi-modal tree) can be replaced by a simple stack of transformer encoder layers with higher performance. Moreover, we re-formulate the visual grounding as a direct coordinates regression problem and avoid making predictions out of a set of candidates (i.e., region proposals or anchor boxes). Extensive experiments are conducted on five widely used datasets, and a series of stateof-the-art records are set by our TransVG. We build the benchmark of transformer-based visual grounding framework and make the code available at https://github. com/djiajunustc/TransVG.

For agents

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

get_harvested_code_for_paper("2104.08541")
get_code_for_paper("2104.08541")
have("2104.08541")

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

curl https://syntology.ai/api/ran/2104.08541.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.08541.svg)](https://syntology.ai/paper/2104.08541)

Connect an agent — have() is free.