SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.10593 · CVPR · 2022

Open-Vocabulary One-Stage Detection with Hierarchical Visual-Language Knowledge Distillation

Yuxin Chen, Shaoru Wang, Jin Gao, Weiming Hu, Liang Li, Zongyang Ma, Congxuan Zhang, Guan Luo

arXiv · PDF · Open in the Atlas

Code that ran

2 of 3 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; 2 failures of our instrument; 1 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/2203.10593.json · badge: /api/ran/2203.10593.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
mengqidyangge/hierkd canonical (classifier label)
classifier basis not recorded on the link
2 ran · 1 no recorded run
FunctionStatusWhere it lives
process_checkpoint Ran
ran; our contract draft was wrong, not the code
mengqidyangge/hierkd/.dev_scripts/gather_models.py
code served (permissive licence) · get_code("138e112e22b0314e")
How it ran

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

import torch
import tempfile
import os

torch.manual_seed(0)

# Create a temporary directory for test files
tmpdir = tempfile.mkdtemp()

# Create a valid checkpoint file
in_file = os.path.join(tmpdir, 'checkpoint.pth')
out_file = os.path.join(tmpdir, 'processed.pth')

# Build a checkpoint with state_dict, optimizer, and ema_ keys
checkpoint = {
    'state_dict': {
        'layer1.weight': torch.randn(4, 3),
        'layer1.bias': torch.randn(4),
        'ema_layer1.weight': torch.randn(4, 3),
        'ema_layer1.bias': torch.randn(4),
    },
    'optimizer': {
        'state': {},
        'param_groups': []
    },
    'epoch': 10
}

# Save the checkpoint to disk
torch.save(checkpoint, in_file)

ARGS = {
    'in_file': in_file,
    'out_file': out_file
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
process_model_info Ran
ran; our contract draft was wrong, not the code
mengqidyangge/hierkd/.dev_scripts/convert_test_benchmark_script.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from opendatalab/MLS-BRN, is pointer only (licence: NONE) · get_code("893e052530d28946")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from opendatalab/MLS-BRN, 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.

ordered_yaml_dump Not yet run
unverified (harvested, no recorded run)
mengqidyangge/hierkd/.dev_scripts/gather_models.py
code served (permissive licence) · get_code("8ff99a6bd2ebf598")
How it ran

no invocation recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 3 the same month, 3 later) and 0 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
ordered_yaml_dump 31 other papers
8 dated earlier, 1 the same month, 22 later
1611.02644 (2016-11)
Multispectral Deep Neural Networks for Pedestrian Detection
1906.07155 (2019-06)
MMDetection: Open MMLab Detection Toolbox and Benchmark
2105.01928 (2021-05)
Instances as Queries
and 28 more not named here
process_checkpoint 22 other papers
6 dated earlier, 1 the same month, 15 later
1611.02644 (2016-11)
Multispectral Deep Neural Networks for Pedestrian Detection
1906.07155 (2019-06)
MMDetection: Open MMLab Detection Toolbox and Benchmark
2105.01928 (2021-05)
Instances as Queries
and 19 more not named here
process_model_info 10 other papers
3 dated earlier, 1 the same month, 6 later
1906.07155 (2019-06)
MMDetection: Open MMLab Detection Toolbox and Benchmark
2105.01928 (2021-05)
Instances as Queries
2107.00420 (2021-07)
CBNet: A Composite Backbone Network Architecture for Object Detection
and 7 more not named here

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

Open-vocabulary object detection aims to detect novel object categories beyond the training set. The advanced open-vocabulary two-stage detectors employ instance-level visual-to-visual knowledge distillation to align the visual space of the detector with the semantic space of the Pretrained Visual-Language Model (PVLM). However, in the more efficient one-stage detector, the absence of classagnostic object proposals hinders the knowledge distillation on unseen objects, leading to severe performance degradation. In this paper, we propose a hierarchical visual-language knowledge distillation method, i.e., Hi-erKD, for open-vocabulary one-stage detection. Specifically, a global-level knowledge distillation is explored to transfer the knowledge of unseen categories from the PVLM to the detector. Moreover, we combine the proposed globallevel knowledge distillation and the common instance-level knowledge distillation to learn the knowledge of seen and unseen categories simultaneously. Extensive experiments on MS-COCO show that our method significantly surpasses the previous best one-stage detector with 11.9% and 6.7% AP 50 gains under the zero-shot detection and generalized zero-shot detection settings, and reduces the AP 50 performance gap from 14% to 7.3% compared to the best twostage detector. Code will be released at this url 1 .

For agents

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

get_harvested_code_for_paper("2203.10593")
get_code_for_paper("2203.10593")
have("2203.10593")

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

curl https://syntology.ai/api/ran/2203.10593.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.10593.svg)](https://syntology.ai/paper/2203.10593)

Connect an agent — have() is free.