SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.08021 · NeurIPS · 2024

OneRef: Unified One-tower Expression Grounding and Segmentation with Mask Referring Modeling

Yaowei Wang, Changsheng Xu, Linhui Xiao, Xiaoshan Yang, Fang Peng

arXiv · PDF · Open in the Atlas

Code that ran

11 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (7 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 4 failures of our instrument; 5 no recorded run), measured 2026-11-21; 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/2410.08021.json · badge: /api/ran/2410.08021.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
linhuixiao/OneRef canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
linhuixiao/hivg community
Papers with Code archive
8 ran · 5 no recorded run
FunctionStatusWhere it lives
CLIPMLP Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("2fffd13c6fa5cd19")
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 an instance of its own class, CLIPMLP: the run constructed the object and did not call it, so no result was computed. 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)

class Config:
    def __init__(self, hidden_size, intermediate_size):
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size

config = Config(hidden_size=4, intermediate_size=8)

ARGS = {
    "config": config
}
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 CLIPMLP — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CLIPMLP 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
Modified_CLIPVisionEmbeddings Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("86707e290b068897")
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 an instance of its own class, Modified_CLIPVisionEmbeddings: the run constructed the object and did not call it, so no result was computed. 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
import math

torch.manual_seed(0)

# Create a minimal stand-in config
class MinimalConfig:
    pass

# Create a minimal CLIP embedding module
class MinimalCLIPEmbedding(nn.Module):
    def __init__(self, embed_dim=4, image_size=8, patch_size=4):
        super().__init__()
        self.config = MinimalConfig()
        self.embed_dim = embed_dim
        self.image_size = image_size
        self.patch_size = patch_size
        self.num_patches = (image_size // patch_size) ** 2
        self.num_positions = self.num_patches + 1
        
        # Class embedding: (1, embed_dim)
        self.class_embedding = nn.Parameter(torch.randn(1, embed_dim))
        
        # Patch embedding: Conv2d that outputs patch embeddings
        self.patch_embedding = nn.Conv2d(
            in_channels=3,
            out_channels=embed_dim,
            kernel_size=patch_size,
            stride=patch_size
        )
        
        # Position embedding: (num_positions, embed_dim)
        self.position_embedding = nn.Embedding(self.num_positions, embed_dim)

# Create instance of minimal CLIP embedding
clip_embed = MinimalCLIPEmbedding(embed_dim=4, image_size=8, patch_size=4)

# Create pixel values: (batch_size=2, channels=3, height=8, width=8)
pixel_values = torch.randn(2, 3, 8, 8)

ARGS = {
    "clip_embed": clip_embed,
}
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 Modified_CLIPVisionEmbeddings — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Modified_CLIPVisionEmbeddings 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
NestedTensor Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("5a5b1be86e3ab9a4")
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 an instance of its own class, NestedTensor: the run constructed the object and did not call it, so no result was computed. 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
from typing import Optional

torch.manual_seed(0)

# Construct a small tensor (batch=2, dim=4)
tensors = torch.randn(2, 4)

# Construct a mask tensor (same shape or compatible shape)
mask = torch.randint(0, 2, (2, 4)).bool()

ARGS = {
    "tensors": tensors,
    "mask": mask,
}
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 NestedTensor — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a NestedTensor 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
TOKEN_MLP Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("8c79c33fa972f11f")
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 an instance of its own class, TOKEN_MLP: the run constructed the object and did not call it, so no result was computed. 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)

class QuickGELU(nn.Module):
    def forward(self, x: torch.Tensor):
        return x * torch.sigmoid(1.702 * x)

class TOKEN_MLP(nn.Module):
    def __init__(self, hidden_size, intermediate_size):
        super().__init__()
        self.activation_fn = QuickGELU()
        self.fc1 = nn.Linear(hidden_size, intermediate_size)
        self.fc2 = nn.Linear(intermediate_size, hidden_size)

    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
        hidden_states = self.fc1(hidden_states)
        hidden_states = self.activation_fn(hidden_states)
        hidden_states = self.fc2(hidden_states)
        return hidden_states

ARGS = {
    "hidden_size": 4,
    "intermediate_size": 8,
}
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 TOKEN_MLP — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a TOKEN_MLP 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
TransformerEncoder Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("06fdb423fe1f1fe7")
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 an instance of its own class, TransformerEncoder: the run constructed the object and did not call it, so no result was computed. 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 encoder layer that satisfies the interface expected by TransformerEncoder
class MinimalEncoderLayer(nn.Module):
    def __init__(self, d_model=4):
        super().__init__()
        self.linear = nn.Linear(d_model, d_model)
    
    def forward(self, src, src_mask=None, src_key_padding_mask=None, pos=None):
        # The layer receives src and optional mask/pos arguments
        # It should return transformed output of same shape
        return self.linear(src)

# Create a minimal norm layer
norm_layer = nn.LayerNorm(4)

# Create encoder layer instance
encoder_layer = MinimalEncoderLayer(d_model=4)

# Number of layers
num_layers = 2

ARGS = {
    "encoder_layer": encoder_layer,
    "num_layers": num_layers,
    "norm": norm_layer
}
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 TransformerEncoder — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a TransformerEncoder 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
TransformerEncoderLayer Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("f003917559e3b08b")
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 an instance of its own class, TransformerEncoderLayer: the run constructed the object and did not call it, so no result was computed. 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
import torch.nn.functional as F

torch.manual_seed(0)

# Define parameters for TransformerEncoderLayer
d_model = 8
nhead = 2
dim_feedforward = 16
dropout = 0.1
activation = 'relu'
normalize_before = False

# Create small tensors for forward pass
batch_size = 2
seq_len = 3

# src: (seq_len, batch_size, d_model)
src = torch.randn(seq_len, batch_size, d_model)

# pos: optional positional embeddings, same shape as src
pos = torch.randn(seq_len, batch_size, d_model)

# src_mask: optional attention mask (seq_len, seq_len)
src_mask = None

# src_key_padding_mask: optional padding mask (batch_size, seq_len)
src_key_padding_mask = None

ARGS = {
    'd_model': d_model,
    'nhead': nhead,
    'dim_feedforward': dim_feedforward,
    'dropout': dropout,
    'activation': activation,
    'normalize_before': normalize_before,
}
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 TransformerEncoderLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a TransformerEncoderLayer 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
VisionLanguageEncoder Ran
ran on a synthesized input · constructed an object; no result was computed
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("8d9679b20f8d8cba")
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 an instance of its own class, VisionLanguageEncoder: the run constructed the object and did not call it, so no result was computed. 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)

ARGS = {
    "d_model": 512,
    "nhead": 8,
    "num_encoder_layers": 6,
    "num_decoder_layers": 6,
    "dim_feedforward": 2048,
    "dropout": 0.1,
    "activation": "relu",
    "normalize_before": False,
}
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 VisionLanguageEncoder — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a VisionLanguageEncoder 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
build_vl_transformer Ran
ran; our contract draft was wrong, not the code
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("dd59acd7b7246b31")
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 VisionLanguageEncoder. 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)

class Args:
    def __init__(self):
        self.vl_hidden_dim = 64
        self.vl_dropout = 0.1
        self.vl_nheads = 4
        self.vl_dim_feedforward = 128
        self.vl_enc_layers = 2
        self.normalize_before = 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 VisionLanguageEncoder — a VisionLanguageEncoder output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a VisionLanguageEncoder 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
get_object_phrase Ran
ran; our contract draft was wrong, not the code
linhuixiao/OneRef/models/OneRef_model.py
code served (permissive licence) · get_code("4a9deb3a2060d602")
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 Token. 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

# Create a minimal spaCy-like doc object
class Token:
    def __init__(self, i, dep_):
        self.i = i
        self.dep_ = dep_

class Doc:
    def __init__(self, tokens):
        self.tokens = tokens
    
    def __iter__(self):
        return iter(self.tokens)
    
    def __getitem__(self, index):
        return self.tokens[index]

# Create a doc with tokens that have dep_ attributes
# One token should have a dep_ in OBJECTS to test the function
doc = Doc([
    Token(0, 'nsubj'),
    Token(1, 'verb'),
    Token(2, 'dobj'),  # This matches OBJECTS
    Token(3, 'punct')
])

ARGS = {
    'doc': doc
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Token — a Token output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Token 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
get_predicate_phrase Ran
ran; our contract draft was wrong, not the code
linhuixiao/OneRef/models/OneRef_model.py
code served (permissive licence) · get_code("596ccf0324e358f5")
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 Token. 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

class Token:
    def __init__(self, i, dep_, pos_, text="word"):
        self.i = i
        self.dep_ = dep_
        self.pos_ = pos_
        self.text = text

class Doc:
    def __init__(self, tokens):
        self.tokens = tokens
    
    def __iter__(self):
        return iter(self.tokens)
    
    def __getitem__(self, idx):
        return self.tokens[idx]

# Create a doc with tokens that have the required attributes
tokens = [
    Token(0, "nsubj", "NOUN", "subject"),
    Token(1, "ROOT", "VERB", "runs"),
    Token(2, "det", "DET", "the"),
    Token(3, "nmod", "NOUN", "park"),
]

doc = Doc(tokens)

ARGS = {
    "doc": doc
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Token — a Token output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Token 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
get_subject_phrase Ran
ran; our contract draft was wrong, not the code
linhuixiao/OneRef/models/OneRef_model.py
code served (permissive licence) · get_code("c366b440880ba949")
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 MockToken. 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
from unittest.mock import Mock

# Create a mock spaCy doc object with tokens that have dep_ attributes
class MockToken:
    def __init__(self, i, dep_, text="word"):
        self.i = i
        self.dep_ = dep_
        self.text = text

class MockDoc:
    def __init__(self, tokens):
        self.tokens = tokens
    
    def __iter__(self):
        return iter(self.tokens)
    
    def __getitem__(self, idx):
        return self.tokens[idx]

# Create a doc with a subject token
doc = MockDoc([
    MockToken(0, "det", "The"),
    MockToken(1, "nsubj", "dog"),
    MockToken(2, "verb", "runs"),
])

ARGS = {
    "doc": doc
}
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 MockToken — a MockToken output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockToken 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
CLIPEncoderLayer_with_Crossmodal_Bridge Not yet run
unverified (harvested, no recorded run)
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("d854f136869b0a41")
How it ran

no invocation recorded.

CLIPEncoder_with_Crossmodal_Bridge Not yet run
unverified (harvested, no recorded run)
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("7d374d3f2f83c0b4")
How it ran

no invocation recorded.

CLIP_Cross_Attention Not yet run
unverified (harvested, no recorded run)
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("9424f702eada219d")
How it ran

no invocation recorded.

CLIP_Vision_Model_with_Crossmodal_Bridge Not yet run
unverified (harvested, no recorded run)
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("f1d58c8cce8cd3ac")
How it ran

no invocation recorded.

HiVG Not yet run
unverified (harvested, no recorded run)
linhuixiao/hivg/models/HiVG.py
code served (permissive licence) · get_code("6465c6905e74b07c")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 7 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 5 later) and 9 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
NestedTensor 32 other papers
32 dated earlier · 28 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2101.01909 (2021-01)
Line Segment Detection Using Transformers without Edges
2103.04503 (2021-03)
End-to-End Human Object Interaction Detection with HOI Transformer
and 29 more not named here
TransformerEncoder 16 other papers
16 dated earlier · 7 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2103.11468 (2021-03)
Learning Multi-Scene Absolute Pose Regression with Transformers
2103.12236 (2021-03)
Instance-level Image Retrieval using Reranking Transformers
and 13 more not named here
TransformerEncoderLayer 16 other papers
14 dated earlier, 2 later · 8 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2103.11468 (2021-03)
Learning Multi-Scene Absolute Pose Regression with Transformers
2104.02439 (2021-04)
Few-Shot Transformation of Common Actions into Time and Space
and 13 more not named here
build_vl_transformer 1 other paper
1 later
2601.01526 (2026-01)
BARE: Towards Bias-Aware and Reasoning-Enhanced One-Tower Visual Grounding
get_object_phrase 1 other paper
1 later
2412.20206 (2024-12)
Toward Visual Grounding: A Survey
get_predicate_phrase 1 other paper
1 later
2412.20206 (2024-12)
Toward Visual Grounding: A Survey
get_subject_phrase 1 other paper
1 later
2412.20206 (2024-12)
Toward Visual Grounding: A Survey

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Constrained by the separate encoding of vision and language, existing grounding and referring segmentation works heavily rely on bulky Transformer-based fusion en-/decoders and a variety of early-stage interaction technologies. Simultaneously, the current mask visual language modeling (MVLM) fails to capture the nuanced referential relationship between image-text in referring tasks. In this paper, we propose OneRef, a minimalist referring framework built on the modality-shared one-tower transformer that unifies the visual and linguistic feature spaces. To modeling the referential relationship, we introduce a novel MVLM paradigm called Mask Referring Modeling (MRefM), which encompasses both referring-aware mask image modeling and referring-aware mask language modeling. Both modules not only reconstruct modality-related content but also cross-modal referring content. Within MRefM, we propose a referring-aware dynamic image masking strategy that is aware of the referred region rather than relying on fixed ratios or generic random masking schemes. By leveraging the unified visual language feature space and incorporating MRefM's ability to model the referential relations, our approach enables direct regression of the referring results without resorting to various complex techniques. Our method consistently surpasses existing approaches and achieves SoTA performance on both grounding and segmentation tasks, providing valuable insights for future research. Our code and models are available at https://github.com/linhuixiao/OneRef.

For agents

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

get_harvested_code_for_paper("2410.08021")
get_code_for_paper("2410.08021")
have("2410.08021")

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

curl https://syntology.ai/api/ran/2410.08021.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.08021.svg)](https://syntology.ai/paper/2410.08021)

Connect an agent — have() is free.