Papers › Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks

Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks

13 Apr 2020ECCV 2020 8arXiv:2004.06165archive 2025-07-28

Xiujun Li, Xi Yin, Chunyuan Li, Pengchuan Zhang, Xiao-Wei Hu, Lei Zhang, Lijuan Wang, Houdong Hu, Li Dong, Furu Wei, Yejin Choi, Jianfeng Gao

Large-scale pre-training methods of learning cross-modal representations on image-text pairs are becoming popular for vision-language tasks. While existing methods simply concatenate image region features and text features as input to the model to be pre-trained and use self-attention to learn image-text semantic alignments in a brute force manner, in this paper, we propose a new learning method Oscar (Object-Semantics Aligned Pre-training), which uses object tags detected in images as anchor points to significantly ease the learning of alignments. Our method is motivated by the observation that the salient objects in an image can be accurately detected, and are often mentioned in the paired text. We pre-train an Oscar model on the public corpus of 6.5 million text-image pairs, and fine-tune it on downstream tasks, creating new state-of-the-arts on six well-established vision-language understanding and generation tasks.

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 23 code samples harvested from 3 repositories linked to this paper: 13 ran (of which 6 constructed an object rather than computing a result; 10 with no instrument failure: 1 honoured, 0 violated, 9 with no contract checked; 3 where Syntology's instrument failed) · 10 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 1 sample from 1 repository, 1 ran (1 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 22 samples from 2 repositories, 12 ran (9 with no instrument failure; 3 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.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

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.

microsoft/Oscar officialmentioned in papermentioned on GitHubpytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report
ThanThoai/Visual-Question-Answering_Vietnamese mentioned on GitHubpytorchApache-2.0Syntology: 6 samples, 3 ran (1 instrument-failed) · 3 unverified report
milvlg/rosita mentioned on GitHubpytorchSyntology: 16 samples, 9 ran (2 instrument-failed) · 7 unverified report
rmokady/clip_prefix_caption mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to ClipCap: CLIP Prefix for Image Captioning (arXiv:2111.09734; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

23 samples harvested: 13 ran (of which 6 constructed an object rather than computing a result; 10 with no instrument failure: 1 honoured, 0 violated, 9 with no contract checked; 3 where Syntology's instrument failed) · 10 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 · honoured contract
3ran · our draft was wrong
9ran
10unverified

Licence: 1 of the 23 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, 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 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 3 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.

microsoft/Oscar official repository 1 sample, 1 ran (0 instrument-failed) · 0 unverified
make_data_sampler in 11 papers oscar/run_captioning.py ran MIT (permissive) · e17f1049b8148f71 · report fixture: signature dataset:DummyDataset/0/?, shuffle:bool/0/bool, distributed:bool/0/bool · returned RandomSampler · level recorded 2026-09-22
ThanThoai/Visual-Question-Answering_Vietnamese community (archive-listed) 6 samples, 3 ran (1 instrument-failed) · 3 unverified
get_list_text merge_raw_data.py ran Apache-2.0 (permissive) · b9f7aa42c0afe2be · report fixture: signature path_en:str/0/str, path_vi:str/0/str, path_idx:str/0/str, type_:str/0/str · returned tuple · level recorded 2026-09-24
make_mask in 16 papers model.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) here; the stored copy is pointer only · e766a0a3063c15d0 · report fixture: signature feature:Tensor/3/float · returned Tensor · level recorded 2026-09-22
read_json in 250 papers merge_raw_data.py ran Apache-2.0 (permissive) · ecf115f3542eefd4 · report fixture: signature path_json:str/0/str · returned dict · level recorded 2026-09-24
get_data_en merge_raw_data.py unverified Apache-2.0 (permissive) · 5d4575f5eacfe1f9 · report
prepare_data in 87 papers extract_image.py unverified Apache-2.0 (permissive) · e6bf616333cbcc3b · report
train_model in 103 papers extract_image.py unverified Apache-2.0 (permissive) · bdc984dffbfed9a3 · report
milvlg/rosita community (archive-listed) 16 samples, 9 ran (2 instrument-failed) · 7 unverified
Backbone in 17 papers rosita/modeling/pretrain_tasks/rosita.py ran Apache-2.0 (permissive) · bb6ea058ef7c09e1 · report fixture: returned Backbone · level recorded 2026-09-22
FeedForward in 198 papers rosita/modeling/pretrain_tasks/rosita.py ran fingerprinted Apache-2.0 (permissive) · 1e4945df552f6045 · report fixture: returned FeedForward · level recorded 2026-09-22
LayerNorm in 121 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
rosita/modeling/pretrain_tasks/rosita.py ran Apache-2.0 (permissive) · fcab94bf5f3f29d5 · report fixture: returned LayerNorm · level recorded 2026-09-22
Pooler in 8 papers rosita/modeling/pretrain_tasks/rosita.py ran Apache-2.0 (permissive) · 14d39fe285bd537f · report fixture: returned Pooler · level recorded 2026-09-22
TextEmbeddings rosita/modeling/pretrain_tasks/rosita.py ran Apache-2.0 (permissive) · 1e2c2b0e7e10968a · report fixture: returned TextEmbeddings · level recorded 2026-09-22
VisualEmbeddings rosita/modeling/pretrain_tasks/rosita.py ran Apache-2.0 (permissive) · b0d904c1669cb4b0 · report fixture: returned VisualEmbeddings · level recorded 2026-09-22
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
rosita/modeling/pretrain_tasks/rosita.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · 8d619e54294fb0b7 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
optimal_transport_dist rosita/modeling/pretrain_tasks/rosita.py ran · our draft was wrong Apache-2.0 (permissive) · 06b045032beeb885 · report fixture: signature txt_emb:Tensor/3/float, img_emb:Tensor/3/float, txt_pad:Tensor/2/bool, img_pad:Tensor/2/bool, beta:float/0/float, iteration:int/0/int, k:int/0/int · returned Tensor · level recorded 2026-09-22
trace in 6 papers rosita/modeling/pretrain_tasks/rosita.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 412cdf6ed5178c2c · report fixture: signature x:Tensor/3/float · returned Tensor · level recorded 2026-09-22
ImgfeatHead rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · 01007aebf0b1b84a · report
MHAtt in 6 papers rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · b97a0ae0d5f606f5 · report
MmITMHead rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · 8bc6b6ae220fc251 · report
MmQAHead rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · b90f9bcd35cb01ad · report
Net in 54 papers rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · 4fde87aac2867c6f · report
SelfAtt rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · 00f7c9f9cc6a5223 · report
TextMLMHead rosita/modeling/pretrain_tasks/rosita.py unverified Apache-2.0 (permissive) · 5b65287416e63add · 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.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for Oscar (graph method: Oscar). 18 bricks; individually on a fixture: 10 ran (of which 6 constructed an object rather than computing a result; 7 with no instrument failure: 1 honoured, 0 violated, 6 with no contract checked; 3 where Syntology's instrument failed) · 8 unverified. The assembly has not been run end-to-end.

  1. Net entry · level 0 · unverified · 4fde87aac2867c6f
  2. Backbone helper · level 1 · ran · bb6ea058ef7c09e1
  3. ImgfeatHead loss · level 0 · unverified · 01007aebf0b1b84a
  4. LayerNorm helper · level 1 · ran · fcab94bf5f3f29d5
  5. MmITMHead loss · level 0 · unverified · 8bc6b6ae220fc251
  6. MmQAHead loss · level 0 · unverified · b90f9bcd35cb01ad
  7. Pooler helper · level 1 · ran · 14d39fe285bd537f
  8. TextEmbeddings model · level 1 · ran · 1e2c2b0e7e10968a
  9. TextMLMHead loss · level 0 · unverified · 5b65287416e63add
  10. VisualEmbeddings model · level 1 · ran · b0d904c1669cb4b0
  11. optimal_transport_dist helper · level 1 · ran · our draft was wrong · 06b045032beeb885
  12. SelfAtt model · level 0 · unverified · 00f7c9f9cc6a5223
  13. gelu helper · level 1 · ran · honoured contract · 8d619e54294fb0b7
  14. cost_matrix_cosine helper · level 1 · ran; the contract check could not run · 9c9d2823696faa43
  15. ipot helper · level 0 · unverified · cbf43203110456f0
  16. trace helper · level 1 · ran · our draft was wrong · 412cdf6ed5178c2c
  17. FeedForward model · level 1 · ran · 1e4945df552f6045
  18. MHAtt helper · level 0 · unverified · b97a0ae0d5f606f5

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="2004.06165") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Cross-Modal RetrievalImage CaptioningImage RetrievalImage-text matchingImage-to-Text RetrievalVisual Question Answering (VQA)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Cross-Modal Retrieval COCO 2014 Oscar Image-to-text R@1 73.5 #19 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 Oscar Image-to-text R@10 96.0 #19 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 Oscar Image-to-text R@5 92.2 #19 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 Oscar Text-to-image R@1 57.5 #19 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 Oscar Text-to-image R@10 89.8 #19 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 Oscar Text-to-image R@5 82.8 #19 of 36 Archive leaderboard report
Image Captioning COCO Captions Oscar BLEU-4 41.7 #11 of 41 Archive leaderboard report
Image Captioning COCO Captions Oscar CIDER 140 #11 of 41 Archive leaderboard report
Image Captioning COCO Captions Oscar METEOR 30.6 #11 of 41 Archive leaderboard report
Image Captioning COCO Captions Oscar SPICE 24.5 #11 of 41 Archive leaderboard report
Image Captioning nocaps-val-overall OSCAR CIDEr 80.9 #11 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall OSCAR Pretrain (#images) 345M #11 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall OSCAR SPICE 11.3 #11 of 11 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) Oscar Recall@10 98.3 #6 of 6 Archive leaderboard report
Image-text matching CommercialAdsDataset OSCAR ADD(S) AUC 87.45 #4 of 8 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) Oscar Recall@10 99.8 #7 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev Oscar Accuracy 73.82 #20 of 56 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

Introduced by this paper: OSCAR

OSCAR

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