SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.12407 · 2020

Visual Compositional Learning for Human-Object Interaction Detection

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 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 0 have no contract verdict; 1 failure of our instrument; 0 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/2007.12407.json · badge: /api/ran/2007.12407.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
zhihou7/HOI-CL-OneStage reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
get_convert_matrix Ran
ran; our contract draft was wrong, not the code
zhihou7/HOI-CL-OneStage/lib/modeling/roi_heads.py
code served (permissive licence) · get_code("ec81a538da82399f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) 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 numpy as np
import pickle
import os
import tempfile

# Create a temporary directory to mock the data directory
temp_dir = tempfile.mkdtemp()
DATA_DIR = temp_dir

# Create mock pickle files
# hoi_to_vb.pkl: maps HOI interaction indices to verb class indices
hoi_to_vb = {
    0: 0,
    1: 1,
    2: 2,
    3: 0,
    4: 1,
    5: 3,
    10: 5,
    20: 10,
    50: 15,
    100: 20,
    200: 30,
    300: 40,
    400: 50,
    500: 60,
    599: 116
}

# hoi_to_obj.pkl: maps HOI interaction indices to HICO object class indices
hoi_to_obj = {
    0: 0,
    1: 1,
    2: 2,
    3: 3,
    4: 4,
    5: 5,
    10: 10,
    20: 20,
    50: 30,
    100: 40,
    200: 50,
    300: 60,
    400: 70,
    500: 75,
    599: 79
}

# Save the pickle files
with open(os.path.join(temp_dir, 'hoi_to_vb.pkl'), 'wb') as f:
    pickle.dump(hoi_to_vb, f)

with open(os.path.join(temp_dir, 'hoi_to_obj.pkl'), 'wb') as f:
    pickle.dump(hoi_to_obj, f)

# Monkey patch the DATA_DIR in the function's scope
import sys
import types

# We need to modify the function to use our temp directory
# Since we can't modify the function directly, we'll set up the environment
# so that when the function runs, it finds our mock data

# Create a wrapper that will temporarily modify the DATA_DIR
original_function_globals = {}

def setup_data_dir():
    global temp_dir
    return temp_dir

# The function uses a hardcoded DATA_DIR, so we need to make sure our pickle files
# are accessible. We'll rely on the function's internal DATA_DIR assignment.
# However, since the function has its own DATA_DIR variable, we need to ensure
# our pickle files exist at that path or modify the approach.

# Since we can't change the hardcoded path in the function, let's create
# the directory structure it expects (if possible) or work around it.
# For the setup, we'll just provide the arguments.

ARGS = {
    'verb_class_num': 117,
    'obj_class_num': 80
}


import numpy as np
import pickle
import os
import tempfile

# Since the function has a hardcoded DATA_DIR, we need to mock the file system
# We'll create the necessary directory structure and files

DATA_DIR = '/public/data0/users/houzhi28/data0/Data/'

# Try to create the directory if it doesn't exist (will likely fail in sandbox)
# Instead, we'll monkey-patch the pickle.load and open functions

# Create mock data
hoi_to_vb_mock = {i: i % 117 for i in range(600)}
hoi_to_obj_mock = {i: i % 80 for i in range(600)}

# Store original functions
original_open = open
original_pickle_load = pickle.load

# Create a mock open function
def mock_open(filename, mode='r'):
    if 'hoi_to_vb.pkl' in filename:
        import io
        return io.BytesIO(pickle.dumps(hoi_to_vb_mock))
    elif 'hoi_to_obj.pkl' in filename:
        import io
        return io.BytesIO(pickle.dumps(hoi_to_obj_mock))
    else:
        return original_open(filename, mode)

# Monkey patch
import builtins
builtins.open = mock_open

ARGS = {
    'verb_class_num': 117,
    'obj_class_num': 80
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
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

Also present in other papers' code

Of 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 1 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

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

Human-Object interaction (HOI) detection aims to localize and infer relationships between human and objects in an image. It is challenging because an enormous number of possible combinations of objects and verbs types forms a long-tail distribution. We devise a deep Visual Compositional Learning (VCL) framework, which is a simple yet efficient framework to effectively address this problem. VCL first decomposes an HOI representation into object and verb specific features, and then composes new interaction samples in the feature space via stitching the decomposed features. The integration of decomposition and composition enables VCL to share object and verb features among different HOI samples and images, and to generate new interaction samples and new types of HOI, and thus largely alleviates the long-tail distribution problem and benefits low-shot or zero-shot HOI detection. Extensive experiments demonstrate that the proposed VCL can effectively improve the generalization of HOI detection on HICO-DET and V-COCO and outperforms the recent state-of-the-art methods on HICO-DET. Code is available at https://github.com/zhihou7/VCL.

For agents

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

get_harvested_code_for_paper("2007.12407")
get_code_for_paper("2007.12407")
have("2007.12407")

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

curl https://syntology.ai/api/ran/2007.12407.json

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

[![Syntology run record](https://syntology.ai/api/ran/2007.12407.svg)](https://syntology.ai/paper/2007.12407)

Connect an agent — have() is free.