Papers › Hulk: A Universal Knowledge Translator for Human-Centric Tasks

Hulk: A Universal Knowledge Translator for Human-Centric Tasks

4 Dec 2023arXiv:2312.01697archive 2025-07-28

Yizhou Wang, Yixuan Wu, Shixiang Tang, Weizhen He, Xun Guo, Feng Zhu, Lei Bai, Rui Zhao, Jian Wu, Tong He, Wanli Ouyang

Human-centric perception tasks, e.g., pedestrian detection, skeleton-based action recognition, and pose estimation, have wide industrial applications, such as metaverse and sports analysis. There is a recent surge to develop human-centric foundation models that can benefit a broad range of human-centric perception tasks. While many human-centric foundation models have achieved success, they did not explore 3D and vision-language tasks for human-centric and required task-specific finetuning. These limitations restrict their application to more downstream tasks and situations. To tackle these problems, we present Hulk, the first multimodal human-centric generalist model, capable of addressing 2D vision, 3D vision, skeleton-based, and vision-language tasks without task-specific finetuning. The key to achieving this is condensing various task-specific heads into two general heads, one for discrete representations, e.g., languages, and the other for continuous representations, e.g., location coordinates. The outputs of two heads can be further stacked into four distinct input and output modalities. This uniform representation enables Hulk to treat diverse human-centric tasks as modality translation, integrating knowledge across a wide range of tasks. Comprehensive evaluations of Hulk on 12 benchmarks covering 8 human-centric tasks demonstrate the superiority of our proposed method, achieving state-of-the-art performance in 11 benchmarks. The code is available on https://github.com/OpenGVLab/Hulk.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 24 code samples harvested from 2 repositories linked to this paper: 13 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 9 where Syntology's instrument failed) · 11 unverified. Where Syntology's instrument failed: 7 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 24 samples from 2 repositories, 13 ran (4 with no instrument failure; 9 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-24 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

opengvlab/hulk officialmentioned in papermentioned on GitHubpytorchMITSyntology: 19 samples, 12 ran (8 instrument-failed) · 7 unverified report
opengvlab/humanbench officialmentioned in paperpytorchMITSyntology: 5 samples, 1 ran (1 instrument-failed) · 4 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

24 samples harvested: 13 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 9 where Syntology's instrument failed) · 11 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
7ran · our draft was wrong
2ran; the contract check could not run
3ran
11unverified

Licence: 11 of the 24 samples are pointer only, meaning Syntology does not serve that copy's text. For 11 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

opengvlab/hulk official repository 19 samples, 12 ran (8 instrument-failed) · 7 unverified
accuracy in 872 papers
census: 8 implementations from 7 papers, 7 distinct outputs across 6 buckets, one shared input per bucket
core/utils.py ran; the contract check could not run (RAISES) MIT (permissive) here; the stored copy is pointer only · 78d1f9f766be74b7 · report fixture: signature output:Tensor/2/float, target:Tensor/1/int, topk:tuple/1/int · returned list · level recorded 2026-09-25
accuracy_multi core/utils.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · 8a63c5da44344ba6 · report fixture: signature output:Tensor/2/float, target:Tensor/2/float · returned Tensor · level recorded 2026-09-24
all_gather in 119 papers core/comm_.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 41c46b0d5713eb14 · report fixture: signature data:dict/0/?, group:NoneType/0/? · returned list · level recorded 2026-09-24
allreduce in 12 papers core/distributed_utils.py ran MIT (permissive) · ae5f2eed69658ac0 · report fixture: signature x:Tensor/2/float, group_idx:int/0/int · returned Tensor · level recorded 2026-09-24
broadcast_object core/comm_.py ran MIT (permissive) · c632684a9726a77a · report fixture: returned str · level recorded 2026-09-24
flat in 11 papers core/config.py ran · honoured contract fingerprinted MIT (permissive) here; the stored copy is pointer only · 4570d6d70bccdfb6 · report fixture: signature nums:list/1/int · returned list · level recorded 2026-09-24
gather in 75 papers core/comm_.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 00a9f7918ac39329 · report fixture: signature data:dict/0/?, dst:int/0/int, group:NoneType/0/? · returned list · level recorded 2026-09-24
init in 55 papers core/fp16/amp.py ran MIT (permissive) here; the stored copy is pointer only · eb8a2600afc1de95 · report fixture: signature enable_caching:bool/0/bool, verbose:bool/0/bool · returned AmpHandle · level recorded 2026-09-24
param_group_fc core/make_param_group.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 70165e0988ce9571 · report fixture: signature model:callable/0/? · returned list · level recorded 2026-09-24
param_group_multitask core/make_param_group.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · e4a7663d686ce91a · report fixture: signature model:callable/0/? · returned list · level recorded 2026-09-24
param_group_no_wd core/make_param_group.py ran · our draft was wrong MIT (permissive) · e4f5737bd720668d · report fixture: signature model:callable/0/? · returned tuple · level recorded 2026-09-24
retry_if_cuda_oom core/memory.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · ff1253be17efe1ae · report fixture: signature func:callable/0/? · returned function · level recorded 2026-09-24
allreduce_async core/distributed_utils.py unverified MIT (permissive) · 02640c5911ca6b56 · report
broadcast in 21 papers core/distributed_utils.py unverified MIT (permissive) · 305673e9acaebcc6 · report
float_function core/fp16/amp.py unverified MIT (permissive) · f27c7ab614e20039 · report
half_function core/fp16/amp.py unverified MIT (permissive) · 6b4eee4b4bb1f238 · report
load_last_iter core/utils.py unverified MIT (permissive) · 7c26a4ba6aca54f0 · report
specific_group_split core/config.py unverified MIT (permissive) · 831beea1b4db57e3 · report
specific_group_split_modality_groups core/config.py unverified MIT (permissive) · 6f5d8cf806cbe77c · report
opengvlab/humanbench official repository 5 samples, 1 ran (1 instrument-failed) · 4 unverified
dist_init in 7 papers PATH/core/distributed_utils.py ran; the contract check could not run (RAISES) MIT (permissive) · 5d01dded5497144b · report fixture: signature method:str/0/str, port:str/0/str · returned tuple · level recorded 2026-09-24
param_group_no_wd PATH/core/make_param_group.py unverified MIT (permissive) · b14254d606cb6719 · report
simple_group_split in 6 papers PATH/core/distributed_utils.py unverified MIT (permissive) here; the stored copy is pointer only · b1471eccec180d87 · report
specific_group_split PATH/core/distributed_utils.py unverified MIT (permissive) here; the stored copy is pointer only · 4e36f197c125d4d3 · report
specific_group_split PATH/core/config.py unverified MIT (permissive) · 62ad1cdb3bc504fe · report

Tasks

3D Human Pose EstimationAction RecognitionHuman Mesh RecoveryHuman Part SegmentationObject DetectionPedestrian Attribute RecognitionPedestrian DetectionPedestrian Image CaptionPose EstimationSemantic SegmentationSkeleton Based Action Recognition

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Human Pose Estimation 3DPW Hulk(ViT-L) MPJPE 66.3 #8 of 119 Archive leaderboard report
3D Human Pose Estimation 3DPW Hulk(ViT-L) MPVPE 77.4 #8 of 119 Archive leaderboard report
3D Human Pose Estimation 3DPW Hulk(ViT-L) PA-MPJPE 38.5 #8 of 119 Archive leaderboard report
3D Human Pose Estimation 3DPW Hulk(ViT-B) MPJPE 67 #17 of 119 Archive leaderboard report
3D Human Pose Estimation 3DPW Hulk(ViT-B) MPVPE 79.8 #17 of 119 Archive leaderboard report
3D Human Pose Estimation 3DPW Hulk(ViT-B) PA-MPJPE 39.9 #17 of 119 Archive leaderboard report
Human Part Segmentation CIHP Hulk(Finetune, ViT-L) Mean IoU 72.68 #1 of 6 Archive leaderboard report
Human Part Segmentation CIHP Hulk(Finetune, ViT-B) Mean IoU 71.26 #2 of 6 Archive leaderboard report
Human Part Segmentation Human3.6M Hulk(Finetune, ViT-L) mIoU 69.89 #1 of 3 Archive leaderboard report
Human Part Segmentation Human3.6M Hulk(Finetune, ViT-B) mIoU 68.56 #2 of 3 Archive leaderboard report
Object Detection CrowdHuman (full body) Hulk(Finetune, ViT-L) AP 93 #6 of 19 Archive leaderboard report
Object Detection CrowdHuman (full body) Hulk(Finetune, ViT-L) mMR 36.5 #6 of 19 Archive leaderboard report
Object Detection CrowdHuman (full body) Hulk(Finetune, ViT-B) AP 92.4 #9 of 19 Archive leaderboard report
Object Detection CrowdHuman (full body) Hulk(Finetune, ViT-B) mMR 40.7 #9 of 19 Archive leaderboard report
Pedestrian Attribute Recognition PA-100K Hulk(Finetune, ViT-L) Accuracy 88.97 #2 of 13 Archive leaderboard report
Pedestrian Attribute Recognition PA-100K Hulk(Finetune, ViT-B) Accuracy 87.85 #3 of 13 Archive leaderboard report
Pedestrian Attribute Recognition RAPv2 Hulk(Finetune, ViT-L) Accuracy 85.86 #1 of 4 Archive leaderboard report
Pedestrian Attribute Recognition RAPv2 Hulk(Finetune, ViT-B) Accuracy 85.26 #2 of 4 Archive leaderboard report
Pose Estimation AIC Hulk(Finetune, ViT-L) AP 37.1 #1 of 10 Archive leaderboard report
Pose Estimation AIC Hulk(Finetune, ViT-B) AP 35.6 #2 of 10 Archive leaderboard report
Pose Estimation COCO (Common Objects in Context) Hulk(Finetune, ViT-L) AP 78.7 #3 of 10 Archive leaderboard report
Pose Estimation COCO (Common Objects in Context) Hulk(Finetune, ViT-B) AP 77.5 #5 of 10 Archive leaderboard report
Semantic Segmentation LIP val Hulk(Finetune, ViT-L) mIoU 66.02% #1 of 13 Archive leaderboard report
Semantic Segmentation LIP val Hulk(Finetune, ViT-B) mIoU 63.98% #2 of 13 Archive leaderboard report
Skeleton Based Action Recognition NTU RGB+D Hulk(Finetune, ViT-L) Accuracy (CS) 94.3 #1 of 135 Archive leaderboard report
Skeleton Based Action Recognition NTU RGB+D Hulk(Finetune, ViT-B) Accuracy (CS) 94 #3 of 135 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections