SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.16005 · CVPR · 2024

Knowledge-Enhanced Dual-stream Zero-shot Composed Image Retrieval

Yi Yang, Linchao Zhu, Yucheng Suo, Fan Ma

arXiv · PDF · Open in the Atlas

Code that ran

10 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; 2 honoured the contract we drafted, 2 violated it and 2 have no contract verdict; 4 failures of our instrument; 7 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/2403.16005.json · badge: /api/ran/2403.16005.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
suoych/keds canonical (classifier label)
classifier basis: census
8 ran · 6 no recorded run
suoych/KEDs community
Papers with Code archive
2 ran · 1 no recorded run
FunctionStatusWhere it lives
Transformer Ran
ran on a synthesized input
suoych/KEDs/src/model/model.py
code served (permissive licence) · get_code("fd7c4feaec54d930")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (3, 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
from torch import nn
from collections import OrderedDict

class LayerNorm(nn.LayerNorm):
    """Subclass torch's LayerNorm to handle fp16."""

    def forward(self, x: torch.Tensor):
        orig_type = x.dtype
        ret = super().forward(x.type(torch.float32))
        return ret.type(orig_type)

class QuickGELU(nn.Module):

    def forward(self, x: torch.Tensor):
        return x * torch.sigmoid(1.702 * x)

class ResidualAttentionBlock(nn.Module):

    def __init__(self, d_model: int, n_head: int, attn_mask: torch.Tensor=None):
        super().__init__()
        self.attn = nn.MultiheadAttention(d_model, n_head)
        self.ln_1 = LayerNorm(d_model)
        self.mlp = nn.Sequential(OrderedDict([('c_fc', nn.Linear(d_model, d_model * 4)), ('gelu', QuickGELU()), ('c_proj', nn.Linear(d_model * 4, d_model))]))
        self.ln_2 = LayerNorm(d_model)
        self.attn_mask = attn_mask

    def attention(self, x: torch.Tensor):
        self.attn_mask = self.attn_mask.to(dtype=x.dtype, device=x.device) if self.attn_mask is not None else None
        return self.attn(x, x, x, need_weights=False, attn_mask=self.attn_mask)[0]

    def forward(self, x: torch.Tensor):
        x = x + self.attention(self.ln_1(x))
        x = x + self.mlp(self.ln_2(x))
        return x

class Transformer(nn.Module):

    def __init__(self, width: int, layers: int, heads: int, attn_mask: torch.Tensor=None):
        super().__init__()
        self.width = width
        self.layers = layers
        self.resblocks = nn.Sequential(*[ResidualAttentionBlock(width, heads, attn_mask) for _ in range(layers)])

    def forward(self, x: torch.Tensor, cross_layers=None, kv_features=None, collect_ind=None, mid_feature=False, text_feature=None, img2text=None):
        if mid_feature:
            result_list = []
            for block in self.resblocks:
                x = block(x)
                result_list.append(x.permute(1, 0, 2))
            return (x, result_list)
        elif cross_layers is not None and kv_features is not None:
            for i in range(len(self.resblocks)):
                if i != len(self.resblocks) - 1 and i <= 5:
                    eos = x[collect_ind]
                    eos = cross_layers(eos.unsqueeze(1), kv_features)
                    x[collect_ind] = x[collect_ind] + eos.squeeze(1)
                    x = self.resblocks[i](x)
                else:
                    x = self.resblocks[i](x)
            return x
        elif text_feature is not None:
            for i in range(len(self.resblocks)):
                if i == len(self.resblocks) - 6:
                    x = x.permute(1, 0, 2)
                    x = torch.cat([x[:, 0].unsqueeze(1) + img2text(x, text_feature).squeeze(1), x[:, 1:]], dim=1)
                    x = x.permute(1, 0, 2)
                    x = self.resblocks[i](x)
                else:
                    x = self.resblocks[i](x)
            return x
        else:
            return self.resblocks(x)

INIT_ARGS = {
    'width': 4,
    'layers': 2,
    'heads': 2,
    'attn_mask': None,
}

ARGS = {
    'x': torch.randn(3, 2, 4),
    'cross_layers': None,
    'kv_features': None,
    'collect_ind': None,
    'mid_feature': False,
    'text_feature': None,
    'img2text': None,
}

if __name__ == '__main__':
    model = Transformer(**INIT_ARGS)
    output = model(**ARGS)
    print("Output shape:", output.shape)
    print("Test passed!")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 (3, 2, 4) — shape (3, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
VisualTransformer Ran
ran on a synthesized input
suoych/KEDs/src/model/model.py
code served (permissive licence) · get_code("6eeb0850e867ee04")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# INIT_ARGS for VisualTransformer.__init__
INIT_ARGS = {
    'input_resolution': 32,  # small image size
    'patch_size': 8,         # divides input_resolution evenly: 32/8 = 4 patches per side
    'width': 16,             # embedding dimension
    'layers': 2,             # number of transformer layers
    'heads': 4,              # number of attention heads (must divide width)
    'output_dim': 12,        # output projection dimension
    'attn_mask': None        # no attention mask
}

# ARGS for VisualTransformer.forward
# Input image: batch_size=2, channels=3, height=32, width=32
ARGS = {
    'x': torch.randn(2, 3, 32, 32),
    'mid_feature': False
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
collate_fn Ran
ran, honoured the contract we drafted
suoych/keds/src/data.py
pointer only (licence: MIT); the stored copy get_code serves, from Chen-Junyang-cn/PLI, is pointer only (licence: NONE) · get_code("b6c1f837e6faaeba")
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 Chen-Junyang-cn/PLI, 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.

find_first_nonzero_indices Ran
ran; our contract draft was wrong, not the code
suoych/keds/src/model/model.py
code served (permissive licence) · get_code("bd7ba008537a41f8")
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 list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes a tensor and finds the first nonzero index in each row (dim=1)
# It needs a 2D tensor with some zeros and some nonzeros
tensor = torch.tensor([
    [0, 0, 5, 0],
    [0, 3, 0, 2]
], dtype=torch.float32)

ARGS = {
    "tensor": tensor
}
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 list, ndim 1, dtype family int — a list 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
get_default_params Ran
ran; our contract draft was wrong, not the code
suoych/keds/src/params.py
code served (permissive licence) · get_code("0ad517bf07a3dc06")
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 dict. 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)

# The function only takes model_name as a parameter
# It checks the model_name against predefined lists and returns appropriate parameter dicts
# No complex objects needed - just a string

ARGS = {
    'model_name': 'RN50'
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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_retrieved_features Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
suoych/keds/src/eval_utils.py
code served (permissive licence) · get_code("7d6d8742930e7202")
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 4, 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)

# Mock faiss GPU index class
class MockFaissGPUIndex:
    def __init__(self, features, metric='L2'):
        self.features = features
        self.metric = metric
    
    def search(self, query, k):
        # query: (batch, dim), k: int
        # returns (distances, indices)
        batch_size = query.shape[0]
        num_features = self.features.shape[0]
        # Simple mock: return random indices
        indices = np.random.randint(0, num_features, size=(batch_size, k))
        distances = np.random.rand(batch_size, k).astype(np.float32)
        return distances, indices

# Create tiny feature tensors
batch_size = 2
feature_dim = 4
topk = 16
database_size = 32

# Feature query tensor (batch_size, feature_dim)
feature = torch.randn(batch_size, feature_dim, dtype=torch.float32)

# Image and text base features
image_base = torch.randn(database_size, feature_dim, dtype=torch.float32)
text_base = torch.randn(database_size, feature_dim, dtype=torch.float32)

# Basenames list
basenames = [f"item_{i}" for i in range(database_size)]

# Create mock faiss indices
image_gpu_index = MockFaissGPUIndex(image_base.numpy())
text_gpu_index = MockFaissGPUIndex(text_base.numpy())

# Build database tuple as expected by function
database = (image_base, text_base, basenames, image_gpu_index, text_gpu_index)

# Mock args object
class MockArgs:
    pass

args = MockArgs()

# Define all function arguments
ARGS = {
    'feature': feature,
    'database': database,
    'args': args,
    'topk': topk,
    'use_faiss': True
}
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 4, 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
is_master Ran
ran, violated the contract we drafted
suoych/keds/src/utils.py
code served (permissive licence) · get_code("c97b5d8516c10994")
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 bool, 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

class Args:
    def __init__(self, distributed=False, gpu=0, dp=False):
        self.distributed = distributed
        self.gpu = gpu
        self.dp = dp

# Test case 1: non-distributed setup (should return True)
args_1 = Args(distributed=False, gpu=0, dp=False)

# Test case 2: distributed with gpu=0 (should return True)
args_2 = Args(distributed=True, gpu=0, dp=False)

# Test case 3: distributed with gpu=1 and no dp (should return False)
args_3 = Args(distributed=True, gpu=1, dp=False)

# Test case 4: distributed with gpu=1 but dp=True (should return True)
args_4 = Args(distributed=True, gpu=1, dp=True)

# Primary ARGS dict for execution
ARGS = {
    'args': Args(distributed=False, gpu=0, dp=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 bool, 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
log_and_continue Ran
ran, violated the contract we drafted
suoych/keds/src/data.py
pointer only (licence: MIT); the stored copy get_code serves, from CSIPlab/SLUG, is pointer only (licence: NONE) · get_code("be4e00be50540042")
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 CSIPlab/SLUG, 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.

prepare_img Ran
ran, honoured the contract we drafted
suoych/keds/src/eval_utils.py
code served (permissive licence) · get_code("c97029a509c1d772")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (3, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torchvision.transforms as transforms
from PIL import Image
import io

torch.manual_seed(0)

# Create a small dummy image in memory
img_data = Image.new('RGB', (4, 3), color='red')
img_bytes = io.BytesIO()
img_data.save(img_bytes, format='PNG')
img_bytes.seek(0)

# Create a temporary file-like object
class FakeFile:
    def __init__(self, img_bytes):
        self.img_bytes = img_bytes
    
    def __enter__(self):
        return self
    
    def __exit__(self, *args):
        pass
    
    def read(self):
        return self.img_bytes.getvalue()

# Create transform that converts PIL image to tensor
transform = transforms.Compose([
    transforms.ToTensor(),
])

# Create a file-like object that Image.open can read from
img_file = io.BytesIO()
img_data.save(img_file, format='PNG')
img_file.seek(0)

ARGS = {
    'img_file': img_file,
    'transform': transform,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, torch, torchvision — 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 (3, 3, 4) — shape (3, 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
setup_primary_logging Ran
ran; our contract draft was wrong, not the code
suoych/keds/src/logger.py
code served (permissive licence) · get_code("050a8682c03e5b7a")
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 Queue. 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 logging
import tempfile
import os
from torch.multiprocessing import Queue

# Create a temporary log file for testing
temp_dir = tempfile.gettempdir()
log_file = os.path.join(temp_dir, "test_log.txt")

# Define arguments for setup_primary_logging
ARGS = {
    "log_file": log_file,
    "level": logging.INFO
}
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 Queue — a Queue output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Queue 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
CLIP Not yet run
unverified (harvested, no recorded run)
suoych/KEDs/src/model/model.py
code served (permissive licence) · get_code("3949802de58e42db")
How it ran

no invocation recorded.

build_model Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/model/model.py
code served (permissive licence) · get_code("340afd63600e948b")
How it ran

no invocation recorded.

concat_all_gather Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/model/model.py
code served (permissive licence) · get_code("73cecca9f3575f09")
How it ran

no invocation recorded.

download_pretrained_from_hf Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/model/clip.py
code served (permissive licence) · get_code("16a61f88e426e60e")
How it ran

no invocation recorded.

has_hf_hub Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/model/clip.py
code served (permissive licence) · get_code("576e20dbc1f9c66e")
How it ran

no invocation recorded.

load_model_without_definition Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/eval_retrieval.py
code served (permissive licence) · get_code("65adb49ebd943e40")
How it ran

no invocation recorded.

load_model_without_definition Not yet run
unverified (harvested, no recorded run)
suoych/keds/src/eval_utils.py
code served (permissive licence) · get_code("cf91dcd7b37a526e")
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 16 distinct functions with an identity, 10 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 2 the same month, 4 later) and 6 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
concat_all_gather 138 other papers
104 dated earlier, 2 the same month, 29 later, 3 same year, order not determinable · 33 only as an assembly brick
1710.09412 (2017-10)
mixup: BEYOND EMPIRICAL RISK MINIMIZATION
1905.04899 (2019-05)
CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
and 135 more not named here
build_model 119 other papers
58 dated earlier, 6 the same month, 48 later, 5 same year, order not determinable, 2 undated
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1901.07441 (2019-01)
PadChest: A large chest x-ray image dataset with multi-label annotated reports
2104.14294 (2021-04)
Emerging Properties in Self-Supervised Vision Transformers
and 116 more not named here
collate_fn 15 other papers
9 dated earlier, 6 later
2003.04618 (2020-03)
Convolutional Occupancy Networks
2209.09552 (2022-09)
Cross-modal Learning for Image-Guided Point Cloud Shape Completion
2301.13823 (2023-01)
Grounding Language Models to Images for Multimodal Inputs and Outputs
and 12 more not named here
log_and_continue 3 other papers
3 later
2405.21070 (2024-05)
What Makes CLIP More Robust to Long-Tailed Pre-Training Data? A Controlled Study for Transferable Insights
2407.11867 (2024-07)
Targeted Unlearning with Single Layer Unlearning Gradient
2502.07972 (2025-02)
Training Sparse Mixture Of Experts Text Embedding Models
download_pretrained_from_hf 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval
get_default_params 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval
has_hf_hub 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval
is_master 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval
prepare_img 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval
setup_primary_logging 1 other paper
1 same year, order not determinable
aaai_28324 (2024 (year only))
Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval

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

We study the zero-shot Composed Image Retrieval (ZS-CIR) task, which is to retrieve the target image given a reference image and a description without training on the triplet datasets. Previous works generate pseudo-word tokens by projecting the reference image features to the text embedding space. However, they focus on the global visual representation, ignoring the representation of detailed attributes, e.g., color, object number and layout. To address this challenge, we propose a Knowledge-Enhanced Dual-stream zero-shot composed image retrieval framework (KEDs). KEDs implicitly models the attributes of the reference images by incorporating a database. The database enriches the pseudo-word tokens by providing relevant images and captions, emphasizing shared attribute information in various aspects. In this way, KEDs recognizes the reference image from diverse perspectives. Moreover, KEDs adopts an extra stream that aligns pseudo-word tokens with textual concepts, leveraging pseudo-triplets mined from image-text pairs. The pseudo-word tokens generated in this stream are explicitly aligned with fine-grained semantics in the text embedding space. Extensive experiments on widely used benchmarks, i.e. ImageNet-R, COCO object, Fashion-IQ and CIRR, show that KEDs outperforms previous zero-shot composed image retrieval methods. Code is available at https://github.com/suoych/KEDs.

For agents

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

get_harvested_code_for_paper("2403.16005")
get_code_for_paper("2403.16005")
have("2403.16005")

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

curl https://syntology.ai/api/ran/2403.16005.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.16005.svg)](https://syntology.ai/paper/2403.16005)

Connect an agent — have() is free.