Papers › ProVision: Programmatically Scaling Vision-centric Instruction Data for Multimodal...

ProVision: Programmatically Scaling Vision-centric Instruction Data for Multimodal Language Models

9 Dec 2024arXiv:2412.07012archive 2025-07-28

Jieyu Zhang, Le Xue, Linxin Song, Jun Wang, Weikai Huang, Manli Shu, An Yan, Zixian Ma, Juan Carlos Niebles, Silvio Savarese, Caiming Xiong, Zeyuan Chen, Ranjay Krishna, ran Xu

With the rise of multimodal applications, instruction data has become critical for training multimodal language models capable of understanding complex image-based queries. Existing practices rely on powerful but costly large language models (LLMs) or multimodal language models (MLMs) to produce instruction data. These are often prone to hallucinations, licensing issues and the generation process is often hard to scale and interpret. In this work, we present a programmatic approach that employs scene graphs as symbolic representations of images and human-written programs to systematically synthesize vision-centric instruction data. Our approach ensures the interpretability and controllability of the data generation process and scales efficiently while maintaining factual accuracy. By implementing a suite of 24 single-image, 14 multi-image instruction generators, and a scene graph generation pipeline, we build a scalable, cost-effective system: ProVision which produces diverse question-answer pairs concerning objects, attributes, relations, depth, etc., for any given image. Applied to Visual Genome and DataComp datasets, we generate over 10 million instruction data points, ProVision-10M, and leverage them in both pretraining and instruction tuning stages of MLMs. When adopted in the instruction tuning stage, our single-image instruction data yields up to a 7% improvement on the 2D split and 8% on the 3D split of CVBench, along with a 3% increase in performance on QBench2, RealWorldQA, and MMMU. Our multi-image instruction data leads to an 8% improvement on Mantis-Eval. Incorporation of our data in both pre-training and fine-tuning stages of xGen-MM-4B leads to an averaged improvement of 1.6% across 11 benchmarks.

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 15 code samples harvested from 1 repository linked to this paper: 7 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 0 violated, 4 with no contract checked; 2 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 15 samples from 1 repository, 7 ran (5 with no instrument failure; 2 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 on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

jieyuz2/provision officialmentioned in paperpytorchApache-2.0Syntology: 15 samples, 7 ran (2 instrument-failed) · 8 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

15 samples harvested: 7 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 0 violated, 4 with no contract checked; 2 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 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
2ran · our draft was wrong
4ran
8unverified

Licence: 0 of the 15 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 jieyuz2/provision. “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.

jieyuz2/provision official repository 15 samples, 7 ran (2 instrument-failed) · 8 unverified
boxOverlap provision/dataset.py ran Apache-2.0 (permissive) · 1b9b3ba31d9f08f9 · report fixture: signature box1:list/1/int, box2:list/1/int, vertical:bool/0/bool, horizontal:bool/0/bool · returned bool · level recorded 2026-09-25
load_image in 273 papers provision/annotation/attribute_llava.py ran · honoured contract Apache-2.0 (permissive) · 5f90221712cb63f4 · report fixture: signature image_file:str/0/str · returned Image · level recorded 2026-09-25
normalize_relations provision/dataset.py ran · our draft was wrong Apache-2.0 (permissive) · 3cfade61c1d29398 · report fixture: signature relations:list/1/tuple · returned list · level recorded 2026-09-25
rel_parsing provision/annotation/relation_grounding.py ran Apache-2.0 (permissive) · a1e26f63b595bca0 · report fixture: signature rel_output:str/0/str, detection_mode:str/0/str · returned list · level recorded 2026-09-25
save_depth_res provision/annotation/depth_estimation.py ran Apache-2.0 (permissive) · 9008ca38b586fa68 · report fixture: signature image_size:tuple/1/int, depth:ndarray/2/float, save_path:str/0/str, grayscale:bool/0/bool · returned str · level recorded 2026-09-25
save_mask_as_png provision/annotation/utils.py ran Apache-2.0 (permissive) · b67cc1d3454644b8 · report fixture: signature masks:list/1/ndarray, file_path:str/0/str, grey_scale:bool/0/bool · returned str · level recorded 2026-09-25
save_sep_mask provision/annotation/utils.py ran · our draft was wrong Apache-2.0 (permissive) · 2ac58005d7f320e4 · report fixture: signature masks:list/1/dict, save_path:str/0/str · returned tuple · level recorded 2026-09-25
cate_grounding provision/annotation/relation_grounding.py unverified Apache-2.0 (permissive) · 7f28431ff5823dfe · report
depth_estimation provision/annotation/depth_estimation.py unverified Apache-2.0 (permissive) · 7a269cd987117ade · report
get_attributes in 10 papers provision/annotation/attribute_llava.py unverified Apache-2.0 (permissive) · 3010b213fcb650e6 · report
get_image in 52 papers provision/annotation/utils.py unverified Apache-2.0 (permissive) · 4354809003526c37 · report
normalize_attributes provision/dataset.py unverified Apache-2.0 (permissive) · a72c327b29ab95e1 · report
object_detection provision/annotation/object_detection.py unverified Apache-2.0 (permissive) · a13bc543209b74ab · report
sam_mask_generation provision/annotation/segmentation.py unverified Apache-2.0 (permissive) · f69b840efc02bd76 · report
sam_mask_prediction provision/annotation/segmentation.py unverified Apache-2.0 (permissive) · 8b3900eb71dae6ba · report

Tasks

Graph GenerationScene Graph GenerationVisual Question Answering

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Visual Question Answering MM-Vet LLaVA-1.5-7B (VG-S) GPT-4 score 40.4 #107 of 231 Archive leaderboard report
Visual Question Answering MM-Vet LLaVA-1.5-7B (DC-S) GPT-4 score 38.5 #123 of 231 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