Papers › BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large...

BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models

30 Jan 2023Conference 2023 2arXiv:2301.12597archive 2025-07-28

Junnan Li, Dongxu Li, Silvio Savarese, Steven Hoi

The cost of vision-and-language pre-training has become increasingly prohibitive due to end-to-end training of large-scale models. This paper proposes BLIP-2, a generic and efficient pre-training strategy that bootstraps vision-language pre-training from off-the-shelf frozen pre-trained image encoders and frozen large language models. BLIP-2 bridges the modality gap with a lightweight Querying Transformer, which is pre-trained in two stages. The first stage bootstraps vision-language representation learning from a frozen image encoder. The second stage bootstraps vision-to-language generative learning from a frozen language model. BLIP-2 achieves state-of-the-art performance on various vision-language tasks, despite having significantly fewer trainable parameters than existing methods. For example, our model outperforms Flamingo80B by 8.7% on zero-shot VQAv2 with 54x fewer trainable parameters. We also demonstrate the model's emerging capabilities of zero-shot image-to-text generation that can follow natural language instructions.

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

No sample from the archive's official repository (salesforce/lavis) is linked to this paper; the 8 samples below come from 4 community repositories. Syntology harvested code from salesforce/lavis for 2 other papers.

salesforce/lavis: Syntology's graph links these samples to Moonshot: Towards Controllable Video Generation and Editing with Multimodal Conditions (arXiv:2401.01827; that paper's own run record: 5 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; 5 where Syntology's instrument failed) · 2 unverified) and When and How Does CLIP Enable Domain and Compositional Generalization? (arXiv:2502.09507; that paper's own run record: 24 ran (of which 0 constructed an object rather than computing a result; 16 with no instrument failure: 0 honoured, 0 violated, 16 with no contract checked; 8 where Syntology's instrument failed) · 8 unverified); nothing checks that they implement this paper's method.

Syntology 8 code samples harvested from 4 repositories linked to this paper: 4 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified; every one of the 4 samples that ran constructed an object rather than computing a result. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: community (archive-listed): 8 samples from 4 repositories, 4 ran (4 with no instrument failure; 0 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 9 of its repositories that Syntology harvested for other papers:

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

17 repositories listed; official and paper-mentioned ones first.

salesforce/lavis officialmentioned in papermentioned on GitHubpytorchBSD-3-ClauseSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Moonshot: Towards Controllable Video Generation and Editing with Multimodal Conditions (arXiv:2401.01827; that paper's own run record: 5 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; 5 where Syntology's instrument failed) · 2 unverified) and When and How Does CLIP Enable Domain and Compositional Generalization? (arXiv:2502.09507; that paper's own run record: 24 ran (of which 0 constructed an object rather than computing a result; 16 with no instrument failure: 0 honoured, 0 violated, 16 with no contract checked; 8 where Syntology's instrument failed) · 8 unverified); nothing checks that they implement this paper's method. report
albertotestoni/ndq_visual_objects mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Grounding Language Models to Images for Multimodal Inputs and Outputs (arXiv:2301.13823; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
alibaba/graphtranslator mentioned on GitHubpytorchBSD-3-ClauseSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to GraphTranslator: Aligning Graph Model to Large Language Model for Open-ended Tasks (arXiv:2402.07197; that paper's own run record: 8 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 1 violated, 0 with no contract checked; 6 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
baaivision/eva mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to EVA: Exploring the Limits of Masked Visual Representation Learning at Scale (arXiv:2211.07636; that paper's own run record: 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) · 2 unverified) and EVA-CLIP-18B: Scaling CLIP to 18 Billion Parameters (arXiv:2402.04252; that paper's own run record: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report
facebookresearch/multimodal mentioned on GitHubpytorchSyntology: 3 samples, 3 ran (0 instrument-failed) · 0 unverified report
gregor-ge/mblip mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to mBLIP: Efficient Bootstrapping of Multilingual Vision-LLMs (arXiv:2307.06930; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
huggingface/transformers mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Attention Is All You Need (arXiv:1706.03762; that paper's own run record: 610 ran (of which 293 constructed an object rather than computing a result; 529 with no instrument failure: 45 honoured, 15 violated, 469 with no contract checked; 81 where Syntology's instrument failed) · 336 unverified), HuggingFace's Transformers: State-of-the-art Natural Language Processing (arXiv:1910.03771; that paper's own run record: 3 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; 3 where Syntology's instrument failed) · 0 unverified), DeBERTa: Decoding-enhanced BERT with Disentangled Attention (arXiv:2006.03654; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 2 honoured, 1 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 9 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
jiwanchung/vlis mentioned on GitHubpytorchSyntology: 1 sample, 0 ran · 1 unverified report
junshutang/Make-It-3D mentioned on GitHubpytorchSyntology: not harvested report
kdr/videorag-mrr2024 mentioned on GitHubNOASSERTIONSyntology: not harvested report
linzhiqiu/clip-flant5 mentioned on GitHubpytorchSyntology: 2 samples, 1 ran (0 instrument-failed) · 1 unverified report
rabiulcste/vqazero mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Investigating Prompting Techniques for Zero- and Few-Shot Visual Question Answering (arXiv:2306.09996; that paper's own run record: 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) · 0 unverified); nothing checks that they implement this paper's method. report
thudm/visualglm-6b mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to TheoremQA: A Theorem-driven Question Answering dataset (arXiv:2305.12524; that paper's own run record: 10 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
unispac/visual-adversarial-examples-jailbreak-large-language-models mentioned on GitHubpytorchSyntology: not harvested report
yukw777/videoblip mentioned on GitHubpytorchSyntology: 2 samples, 0 ran · 2 unverified report
2024-MindSpore-1/Code2 mindsporeSyntology: not harvested report
yangyucheng000/University mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution (arXiv:2409.12191; that paper's own run record: 12 ran (of which 1 constructed an object rather than computing a result; 10 with no instrument failure: 4 honoured, 1 violated, 5 with no contract checked; 2 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

8 samples harvested: 4 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified; every one of the 4 samples that ran constructed an object rather than computing a result. “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.

4ran
4unverified

Licence: 1 of the 8 samples is 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 4 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.

facebookresearch/multimodal community (archive-listed) 3 samples, 3 ran (0 instrument-failed) · 0 unverified
BLIP2 torchmultimodal/models/blip2/blip2.py ran BSD-3-Clause (permissive) · dbb9b4ef6b0f95dd · report fixture: returned BLIP2 · level recorded 2026-09-22
Blip2Output torchmultimodal/models/blip2/blip2.py ran · metamorphic tier: deterministic BSD-3-Clause (permissive) · b939479f3d2573b8 · report fixture: returned Blip2Output · level recorded 2026-09-22
TransformerOutput in 5 papers torchmultimodal/models/blip2/blip2.py ran · metamorphic tier: deterministic BSD-3-Clause (permissive) · e62bffe433f61ff1 · report fixture: returned TransformerOutput · level recorded 2026-09-22
jiwanchung/vlis community (archive-listed) 1 sample, 0 ran · 1 unverified
LynxBase code/lynx/lynx_llm/models/lynx.py unverified no licence file found · pointer only · 6e342d36cf541d23 · report
linzhiqiu/clip-flant5 community (archive-listed) 2 samples, 1 ran (0 instrument-failed) · 1 unverified
CLIPT5Config llava/model/language_model/clip_t5.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 196ad3b77304e23c · report fixture: returned CLIPT5Config · level recorded 2026-09-22
CLIPT5ForConditionalGeneration llava/model/language_model/clip_t5.py unverified Apache-2.0 (permissive) · 11e6ad12cf0667b0 · report
yukw777/videoblip community (archive-listed) 2 samples, 0 ran · 2 unverified
VideoBlipForConditionalGeneration video_blip/model.py unverified MIT (permissive) · b67db6a1705fb783 · report
VideoBlipVisionModel video_blip/model.py unverified MIT (permissive) · c4125cdecc4964aa · 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.

4 assemblies on this paper: 4 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).

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

harvested Assembly for BLIP-2 (graph method: BLIP-2). 3 bricks; individually on a fixture: 3 ran (of which 3 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; every one of the 3 samples that ran constructed an object rather than computing a result. The assembly has not been run end-to-end.

one of 4 siblings

  1. BLIP2 entry · level 1 · ran · dbb9b4ef6b0f95dd
  2. Blip2Output helper · level 1 · ran · b939479f3d2573b8
  3. TransformerOutput helper · level 1 · ran · e62bffe433f61ff1

harvested Assembly for BLIP-2 (graph method: BLIP-2). 2 bricks; individually on a fixture: 0 ran · 2 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. VideoBlipForConditionalGeneration entry · level 0 · unverified · b67db6a1705fb783
  2. VideoBlipVisionModel helper · level 0 · unverified · c4125cdecc4964aa

harvested Assembly for BLIP-2 (graph method: BLIP-2). 6 bricks; individually on a fixture: 3 ran (of which 2 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. CLIPT5ForConditionalGeneration entry · level 0 · unverified · 11e6ad12cf0667b0
  2. CLIPT5Config helper · level 1 · ran · 196ad3b77304e23c
  3. build_vision_projector helper · level 1 · ran · our draft was wrong · ef4db0da176c674c
  4. build_vision_tower helper · level 0 · unverified · 014050859277825c
  5. IdentityMap helper · level 1 · ran · 9f0f5d62445adabe
  6. CLIPVisionTower model · level 0 · unverified · 6422a6162a5e6169

harvested Assembly for BLIP-2 (graph method: BLIP-2). 2 bricks; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. LynxBase entry · level 0 · unverified · 6e342d36cf541d23
  2. disabled_train helper · level 1 · ran · violated contract · 4cb732f513d69dfd

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="2301.12597") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Generative Visual Question AnsweringImage CaptioningImage RetrievalImage to textImage-to-Text RetrievalLanguage ModelingLanguage ModellingMedical Visual Question AnsweringMultiple-choiceOpen Vocabulary Attribute DetectionRepresentation LearningText GenerationVisual Question AnsweringVisual Question Answering (VQA)Visual ReasoningZero-shot Text-to-Image Retrievalvisual instruction following

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Generative Visual Question Answering PMC-VQA BLIP-2 BLEU-1 7.6 #2 of 3 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G OPT 2.7B (zero-shot) BLEU-4 43.7 #4 of 41 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G OPT 2.7B (zero-shot) CIDER 145.8 #4 of 41 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G OPT 6.7B (zero-shot) BLEU-4 43.5 #5 of 41 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G OPT 6.7B (zero-shot) CIDER 145.2 #5 of 41 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G FlanT5 XL (zero-shot) BLEU-4 42.4 #8 of 41 Archive leaderboard report
Image Captioning COCO Captions BLIP-2 ViT-G FlanT5 XL (zero-shot) CIDER 144.5 #8 of 41 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) CIDEr 123.7 #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) Pre-train (#images) 1.1B #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) SPICE 16.3 #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) CIDEr 123.7 #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) Pre-train (#images) 1.1B #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) SPICE 15.8 #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) CIDEr 123 #3 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) Pre-train (#images) 1.1B #3 of 11 Archive leaderboard report
Image Captioning nocaps-val-in-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) SPICE 15.8 #3 of 11 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) CIDEr 120.2 #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) Pre-train (#images) 1.1B #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) SPICE 15.9 #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) CIDEr 119.2 #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) Pre-train (#images) 1.1B #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) SPICE 15.3 #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) CIDEr 117.8 #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) Pre-train (#images) 1.1B #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-near-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) SPICE 15.4 #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) CIDEr 124.8 #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) Pretrain (#images) 1.1B #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G FlanT5 XL (zero-shot) SPICE 15.1 #1 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) CIDEr 124.4 #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) Pretrain (#images) 1.1B #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 6.7B (zero-shot) SPICE 14.8 #2 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) CIDEr 123.4 #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) Pretrain (#images) 1.1B #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-out-domain BLIP-2 ViT-G OPT 2.7B (zero-shot) SPICE 15.1 #3 of 10 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G FlanT5 XL (zero-shot) CIDEr 121.6 #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G FlanT5 XL (zero-shot) Pretrain (#images) 1.1B #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G FlanT5 XL (zero-shot) SPICE 15.8 #1 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 6.7B (zero-shot) CIDEr 121.0 #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 6.7B (zero-shot) Pretrain (#images) 1.1B #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 6.7B (zero-shot) SPICE 15.3 #2 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 2.7B (zero-shot) CIDEr 119.7 #3 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 2.7B (zero-shot) Pretrain (#images) 1.1B #3 of 11 Archive leaderboard report
Image Captioning nocaps-val-overall BLIP-2 ViT-G OPT 2.7B (zero-shot) SPICE 15.4 #3 of 11 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-G (fine-tuned) Recall@10 92.6 #1 of 6 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-G (fine-tuned) recall@1 68.3 #1 of 6 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-G (fine-tuned) recall@5 87.7 #1 of 6 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-L (fine-tuned) Recall@10 91.8 #3 of 6 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-L (fine-tuned) recall@1 66.3 #3 of 6 Archive leaderboard report
Image Retrieval COCO (Common Objects in Context) BLIP-2 ViT-L (fine-tuned) recall@5 86.5 #3 of 6 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@1 89.7 #1 of 9 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@10 98.9 #1 of 9 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@5 98.1 #1 of 9 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@1 88.6 #2 of 9 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@10 98.9 #2 of 9 Archive leaderboard report
Image Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@5 97.6 #2 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-G, fine-tuned) Recall@1 85.4 #1 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-G, fine-tuned) Recall@10 98.5 #1 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-G, fine-tuned) Recall@5 97.0 #1 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-L, fine-tuned) Recall@1 83.5 #3 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-L, fine-tuned) Recall@10 98.0 #3 of 9 Archive leaderboard report
Image-to-Text Retrieval COCO (Common Objects in Context) BLIP-2 (ViT-L, fine-tuned) Recall@5 96.0 #3 of 9 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@1 97.6 #2 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@10 100 #2 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-G (zero-shot, 1K test set) Recall@5 100 #2 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@1 96.9 #5 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@10 100 #5 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k BLIP-2 ViT-L (zero-shot, 1K test set) Recall@5 100 #5 of 11 Archive leaderboard report
Open Vocabulary Attribute Detection OVAD-Box benchmark BLIP 2 (pretrained) mean average precision 25.5 #2 of 7 Archive leaderboard report
Visual Question Answering MM-Vet BLIP-2-12B GPT-4 score 22.4±0.2 #226 of 231 Archive leaderboard report
Visual Question Answering MM-Vet BLIP-2-12B Params 12B #226 of 231 Archive leaderboard report
Visual Question Answering VQA v2 test-dev BLIP-2 ViT-G OPT 6.7B (fine-tuned) Accuracy 82.30 #1 of 11 Archive leaderboard report
Visual Question Answering VQA v2 test-dev BLIP-2 ViT-G OPT 2.7B (fine-tuned) Accuracy 81.74 #4 of 11 Archive leaderboard report
Visual Question Answering VQA v2 test-dev BLIP-2 ViT-G FlanT5 XL (fine-tuned) Accuracy 81.66 #5 of 11 Archive leaderboard report
Visual Question Answering VQA v2 val BLIP-2 ViT-G OPT 6.7B (fine-tuned) Accuracy 82.19 #1 of 4 Archive leaderboard report
Visual Question Answering VQA v2 val BLIP-2 ViT-G OPT 2.7B (fine-tuned) Accuracy 81.59 #2 of 4 Archive leaderboard report
Visual Question Answering VQA v2 val BLIP-2 ViT-G FlanT5 XL (fine-tuned) Accuracy 81.55 #3 of 4 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-G FlanT5 XXL (zero-shot) Accuracy 44.7 #10 of 17 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-L FlanT5 XL (zero-shot) Accuracy 44.4 #11 of 17 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-G FlanT5 XL (zero-shot) Accuracy 44.2 #12 of 17 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-G OPT 6.7B (zero-shot) Accuracy 36.4 #14 of 17 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-G OPT 2.7B (zero-shot) Accuracy 34.6 #15 of 17 Archive leaderboard report
Visual Question Answering (VQA) GQA test-dev BLIP-2 ViT-L OPT 2.7B (zero-shot) Accuracy 33.9 #16 of 17 Archive leaderboard report
Visual Question Answering (VQA) InfiMM-Eval BLIP-2-OPT2.7B Abductive 18.96 #12 of 14 Archive leaderboard report
Visual Question Answering (VQA) InfiMM-Eval BLIP-2-OPT2.7B Analogical 7.5 #12 of 14 Archive leaderboard report
Visual Question Answering (VQA) InfiMM-Eval BLIP-2-OPT2.7B Deductive 2.76 #12 of 14 Archive leaderboard report
Visual Question Answering (VQA) InfiMM-Eval BLIP-2-OPT2.7B Overall score 19.31 #12 of 14 Archive leaderboard report
Visual Question Answering (VQA) InfiMM-Eval BLIP-2-OPT2.7B Params 3B #12 of 14 Archive leaderboard report
Visual Question Answering (VQA) InfoSeek BLIP2 Accuracy 14.6 #6 of 7 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-G FlanT5 XXL (zero-shot) Accuracy 45.9 #23 of 37 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-G FlanT5 XL (zero-shot) Accuracy 40.7 #28 of 37 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-L FlanT5 XL (zero-shot) Accuracy 39.4 #29 of 37 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-G OPT 6.7B (zero-shot) Accuracy 36.4 #30 of 37 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-G OPT 2.7B (zero-shot) Accuracy 31.7 #32 of 37 Archive leaderboard report
Visual Question Answering (VQA) OK-VQA BLIP-2 ViT-L OPT 2.7B (zero-shot) Accuracy 30.2 #33 of 37 Archive leaderboard report
Visual Question Answering (VQA) PMC-VQA BLIP-2 Accuracy 24.3 #4 of 4 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-G FlanT5 XXL (zero-shot) Accuracy 65 #42 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-G FlanT5 XL (zero-shot) Accuracy 63 #47 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-L FlanT5 XL (zero-shot) Accuracy 62.3 #48 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-G OPT 6.7B (zero-shot) Accuracy 52.6 #50 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-G OPT 2.7B (zero-shot) Accuracy 52.3 #51 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-dev BLIP-2 ViT-L OPT 2.7B (zero-shot) Accuracy 49.7 #54 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-G FlanT5 XXL (zero-shot) Accuracy 65.2 #1 of 11 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-G FlanT5 XL (zero-shot) Accuracy 63.1 #3 of 11 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-L FlanT5 XL (zero-shot) Accuracy 62.6 #4 of 11 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-G OPT 6.7B (zero-shot) Accuracy 54.3 #5 of 11 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-G OPT 2.7B (zero-shot) Accuracy 53.5 #6 of 11 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 val BLIP-2 ViT-L OPT 2.7B (zero-shot) Accuracy 50.1 #7 of 11 Archive leaderboard report
visual instruction following LLaVA-Bench BLIP-2 avg score 38.1 #8 of 8 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

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