Papers › RLIPv2: Fast Scaling of Relational Language-Image Pre-training

RLIPv2: Fast Scaling of Relational Language-Image Pre-training

18 Aug 2023ICCV 2023 1arXiv:2308.09351archive 2025-07-28

Hangjie Yuan, Shiwei Zhang, Xiang Wang, Samuel Albanie, Yining Pan, Tao Feng, Jianwen Jiang, Dong Ni, Yingya Zhang, Deli Zhao

Relational Language-Image Pre-training (RLIP) aims to align vision representations with relational texts, thereby advancing the capability of relational reasoning in computer vision tasks. However, hindered by the slow convergence of RLIPv1 architecture and the limited availability of existing scene graph data, scaling RLIPv1 is challenging. In this paper, we propose RLIPv2, a fast converging model that enables the scaling of relational pre-training to large-scale pseudo-labelled scene graph data. To enable fast scaling, RLIPv2 introduces Asymmetric Language-Image Fusion (ALIF), a mechanism that facilitates earlier and deeper gated cross-modal fusion with sparsified language encoding layers. ALIF leads to comparable or better performance than RLIPv1 in a fraction of the time for pre-training and fine-tuning. To obtain scene graph data at scale, we extend object detection datasets with free-form relation labels by introducing a captioner (e.g., BLIP) and a designed Relation Tagger. The Relation Tagger assigns BLIP-generated relation texts to region pairs, thus enabling larger-scale relational pre-training. Through extensive experiments conducted on Human-Object Interaction Detection and Scene Graph Generation, RLIPv2 shows state-of-the-art performance on three benchmarks under fully-finetuning, few-shot and zero-shot settings. Notably, the largest RLIPv2 achieves 23.29mAP on HICO-DET without any fine-tuning, yields 32.22mAP with just 1% data and yields 45.09mAP with 100% data. Code and models are publicly available at https://github.com/JacobYuan7/RLIPv2.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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 30 code samples harvested from 3 repositories linked to this paper: 22 ran (of which 5 constructed an object rather than computing a result; 18 with no instrument failure: 0 honoured, 1 violated, 17 with no contract checked; 4 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 10 samples from 1 repository, 8 ran (8 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 20 samples from 2 repositories, 14 ran (10 with no instrument failure; 4 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 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 22 samples that ran has no recorded date.

jacobyuan7/rlipv2 officialmentioned in papermentioned on GitHubpytorchApache-2.0Syntology: 10 samples, 8 ran (0 instrument-failed) · 2 unverified report
jacobyuan7/ocn-hoi-benchmark mentioned on GitHubpytorchApache-2.0Syntology: 10 samples, 7 ran (3 instrument-failed) · 3 unverified report
jacobyuan7/rlip mentioned on GitHubpytorchSyntology: 10 samples, 7 ran (1 instrument-failed) · 3 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

30 samples harvested: 22 ran (of which 5 constructed an object rather than computing a result; 18 with no instrument failure: 0 honoured, 1 violated, 17 with no contract checked; 4 where Syntology's instrument failed) · 8 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · violated contract
4ran · our draft was wrong
17ran
8unverified

Licence: 0 of the 30 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 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”. For 2 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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.

jacobyuan7/rlipv2 official repository 10 samples, 8 ran (0 instrument-failed) · 2 unverified
RLIPv2_BiMultiHeadAttention models/fuse_helper.py ran Apache-2.0 (permissive) · 2303ca2dfb7c411f · report fixture: signature v:Tensor/3/float, l:Tensor/3/float, v_pos:Tensor/3/float, attention_mask_l:Tensor/2/float, attention_mask_v:Tensor/2/float · returned tuple · level recorded 2026-09-24
check_list_out_of_bounds generate_relations_using_verb_tagger.py ran Apache-2.0 (permissive) · 5425cd58fc9f60ff · report fixture: signature relation_candidates:list/1/tuple, boxes:list/1/dict · returned bool · level recorded 2026-09-24
compute_relation_distribution_from_scene_graphs datasets/rlipv2_helper/coco_caption_helper.py ran Apache-2.0 (permissive) · de2f09364131391c · report fixture: signature sng:dict/0/? · returned dict · level recorded 2026-09-24
grammartical_tranform datasets/rlipv2_helper/coco_caption_helper.py ran Apache-2.0 (permissive) · 46b10090febf5d85 · report fixture: signature sng:dict/0/? · returned dict · level recorded 2026-09-24
load_hico_object_txt datasets/rlipv2_helper/OD_rel_cand_select.py ran Apache-2.0 (permissive) · 45e4eb399a99d475 · report fixture: signature file_path:str/0/str · returned dict · level recorded 2026-09-24
load_hico_object_txt inference_on_custom_imgs_hico.py ran Apache-2.0 (permissive) · be161ebd7c1a3d07 · report fixture: signature file_path:str/0/str · returned list · level recorded 2026-09-24
load_hico_verb_txt inference_on_custom_imgs_hico.py ran Apache-2.0 (permissive) · 5e1b3b1dd052111f · report fixture: signature file_path:str/0/str · returned list · level recorded 2026-09-24
transform_coco_bbox_to_VG_format inference_on_custom_imgs_pseudo_coco.py ran Apache-2.0 (permissive) · 34ca1776c2c6b50f · report fixture: signature coco_bbox:list/1/dict · returned list · level recorded 2026-09-24
filter_by_gt_object_annotations inference_on_custom_imgs_pseudo_coco.py unverified Apache-2.0 (permissive) · 9b06d4ed21a99088 · report
transform_o365_official_to_VG_format datasets/rlipv2_helper/BLIP_o365_caption_helper.py unverified Apache-2.0 (permissive) · e040382fdc48b50a · report
jacobyuan7/ocn-hoi-benchmark community (archive-listed) 10 samples, 7 ran (3 instrument-failed) · 3 unverified
count_fusion models/hoi.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · b66569473e9784e1 · report fixture: signature x:Tensor/2/float, y:Tensor/2/float · returned Tensor · level recorded 2026-09-22
dice_loss in 193 papers
census: 10 implementations from 16 papers, 7 distinct outputs across 4 buckets, one shared input per bucket this sample: bucket 1 of 4, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
models/segmentation.py ran · violated contract fingerprinted Apache-2.0 (permissive) · ac8fe530cdad4d8c · report fixture: signature inputs:Tensor/2/float, targets:Tensor/2/float, num_boxes:int/0/int · returned Tensor · level recorded 2026-09-22
exponential_inc_iterative_loss engine.py ran Apache-2.0 (permissive) · 5dad3494aed8377b · report fixture: signature loss_dict_reduced:dict/0/?, weight_dict:dict/0/?, model:callable/0/? · returned dict · level recorded 2026-09-24
flip_lr models/decode.py ran Apache-2.0 (permissive) · 1d2943763d1f2ef8 · report fixture: signature x:Tensor/4/float, flip_idx:list/1/tuple · returned Tensor · level recorded 2026-09-24
flip_tensor models/decode.py ran fingerprinted Apache-2.0 (permissive) · 6e3e0dbceafbf01c · report fixture: signature x:Tensor/4/float · returned Tensor · level recorded 2026-09-24
norm_tensor in 5 papers models/hoi.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · d6d0aa8ce6bedcbb · report fixture: signature tensor:Tensor/2/float · returned Tensor · level recorded 2026-09-22
sigmoid_focal_loss in 122 papers
census: 3 implementations from 9 papers, 2 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
models/segmentation.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 5c0711aada67957e · report fixture: signature inputs:Tensor/2/float, targets:Tensor/2/float, num_boxes:int/0/int, alpha:float/0/float, gamma:float/0/float · returned Tensor · level recorded, date not recorded
build_transformer in 62 papers models/transformer.py unverified Apache-2.0 (permissive) · 9a524c433c47ed10 · report
flip_lr_off models/decode.py unverified Apache-2.0 (permissive) · 5f199ad5ac482895 · report
set_hoi convert_vcoco_annotations.py unverified Apache-2.0 (permissive) · 3755b0f2896c9d66 · report
jacobyuan7/rlip community (archive-listed) 10 samples, 7 ran (1 instrument-failed) · 3 unverified
CrossModelTransformerEncoder models/ParSetransformer.py ran fingerprinted Apache-2.0 (permissive) · 4a630df05e78ca0c · report fixture: returned CrossModelTransformerEncoder · level recorded 2026-09-22
DeformableTransformerEncoder in 7 papers models/ParSetransformer.py ran Apache-2.0 (permissive) · 4621c1ee7b2ff5e6 · report fixture: returned DeformableTransformerEncoder · level recorded 2026-09-22
FeatureResizer models/ParSetransformer.py ran fingerprinted Apache-2.0 (permissive) · b3b492f41ac84c02 · report fixture: returned FeatureResizer · level recorded 2026-09-22
MSDeformAttn in 13 papers models/ParSetransformer.py ran Apache-2.0 (permissive) · 5d8915fe61383bce · report fixture: returned MSDeformAttn · level recorded 2026-09-22
TransformerDecoderHOI models/ParSetransformer.py ran Apache-2.0 (permissive) · cf0fda09312fb592 · report fixture: returned TransformerDecoderHOI · level recorded 2026-09-22
TransformerEncoderLayer in 90 papers models/ParSetransformer.py ran fingerprinted Apache-2.0 (permissive) · 107ffdb07dbd2f9d · report fixture: signature src:Tensor/3/float, src_mask:NoneType/0/?, src_key_padding_mask:NoneType/0/?, pos:NoneType/0/? · returned Tensor · level recorded 2026-09-24
_get_clones in 28 papers models/ParSetransformer.py ran · our draft was wrong Apache-2.0 (permissive) · c03e17b75656115f · report fixture: signature module:callable/0/?, N:int/0/int · returned ModuleList · level recorded 2026-09-22
DeformableTransformerDecoderLayer in 5 papers models/ParSetransformer.py unverified Apache-2.0 (permissive) · 66064a88a6ab62de · report
DeformableTransformerEncoderLayer in 9 papers models/ParSetransformer.py unverified Apache-2.0 (permissive) · 64cee3e52db36a85 · report
ParSeDeformableTransformer models/ParSetransformer.py unverified Apache-2.0 (permissive) · 4107165e358237d6 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another.

3 assemblies on this paper: 3 harvested, 0 generated, listed in 2 groups (siblings lifted in one batch form one group).

harvested Assembly for RLIPv2 (graph method: RLIPv2). 13 bricks; individually on a fixture: 10 ran (of which 5 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 1 violated, 6 with no contract checked; 3 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

  1. ParSeDeformableTransformer entry · level 0 · unverified · 4107165e358237d6
  2. CrossModelTransformerEncoder helper · level 1 · ran · 4a630df05e78ca0c
  3. DeformableTransformerDecoderLayer helper · level 0 · unverified · 66064a88a6ab62de
  4. DeformableTransformerEncoder helper · level 1 · ran · 4621c1ee7b2ff5e6
  5. DeformableTransformerEncoderLayer helper · level 0 · unverified · 64cee3e52db36a85
  6. FeatureResizer helper · level 1 · ran · b3b492f41ac84c02
  7. MSDeformAttn helper · level 1 · ran · 5d8915fe61383bce
  8. TransformerDecoderHOI helper · level 1 · ran · cf0fda09312fb592
  9. TransformerEncoderLayer model · level 1 · ran · 107ffdb07dbd2f9d
  10. _get_clones helper · level 1 · ran · our draft was wrong · c03e17b75656115f
  11. _get_activation_fn helper · level 1 · ran · our draft was wrong · 0bb264f52e707e08
  12. _is_power_of_2 helper · level 1 · ran · violated contract · 2b45e8897cb1e2ea
  13. inverse_sigmoid helper · level 1 · ran · our draft was wrong · 83c03ceb1307c6c4

2 sibling assemblies for Asymmetric Language-Image Fusion (ALIF) (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for Asymmetric Language-Image Fusion (ALIF) (graph method: Asymmetric Language-Image Fusion (ALIF)). 1 brick; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. RLIPv2_BiMultiHeadAttention entry · level 1 · ran · 2303ca2dfb7c411f

harvested Assembly for Asymmetric Language-Image Fusion (ALIF) (graph method: Asymmetric Language-Image Fusion (ALIF)). 4 bricks; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. CrossModalCalibration entry · level 0 · unverified · 0c5d441376f6d47a
  2. MHCrossAttLayer model · level 0 · unverified · 62ea1d7f68ecba93
  3. TransformerLayer helper · level 0 · unverified · dafcb62dbe894ef0
  4. _get_clones helper · level 1 · ran · our draft was wrong · 698f5d9bbd2a4581

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2308.09351") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Graph GenerationHuman-Object Interaction DetectionObject DetectionRelational ReasoningScene Graph GenerationZero-Shot Human-Object Interaction Detectionobject-detection

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Human-Object Interaction Detection HICO-DET RLIPv2 (Swin-L) mAP 45.09 #2 of 55 Archive leaderboard report
Human-Object Interaction Detection V-COCO RLIPv2 AP(S1) 72.1 #1 of 34 Archive leaderboard report
Human-Object Interaction Detection V-COCO RLIPv2 AP(S2) 74.1 #1 of 34 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.

Methods

ALIGN

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