Papers › A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs
A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs
Wangbo Zhao, Yizeng Han, Jiasheng Tang, Zhikai Li, Yibing Song, Kai Wang, Zhangyang Wang, Yang You
Vision-language models (VLMs) have shown remarkable success across various multi-modal tasks, yet large VLMs encounter significant efficiency challenges due to processing numerous visual tokens. A promising approach to accelerating large VLM inference is using partial information, such as attention maps from specific layers, to assess token importance and prune less essential tokens. However, our study reveals three key insights: (i) Partial attention information is insufficient for accurately identifying critical visual tokens, resulting in suboptimal performance, especially at low token retention ratios; (ii) Global attention information, such as the attention map aggregated across all layers, more effectively preserves essential tokens and maintains comparable performance under aggressive pruning. However, the attention maps from all layers requires a full inference pass, which increases computational load and is therefore impractical in existing methods; and (iii) The global attention map aggregated from a small VLM closely resembles that of a large VLM, suggesting an efficient alternative. Based on these findings, we introduce a training-free method, Small VLM Guidance for accelerating Large VLMs (SGL). Specifically, we employ the attention map aggregated from a small VLM to guide visual token pruning in a large VLM. Additionally, an early exiting mechanism is developed to fully use the small VLM's predictions, dynamically invoking the larger VLM only when necessary, yielding a superior trade-off between accuracy and computation. Extensive evaluations across 11 benchmarks demonstrate the effectiveness and generalizability of SGL, achieving up to 91% pruning ratio for visual tokens while retaining competitive performance.
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):
get_harvested_code_for_paper(arxiv_id="2412.03324")get_code(code_sha256="…")with a sample'scode_sha256from the list belowget_citation_path(paper_1="2412.03324", paper_2="…")with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="2412.03324")
Code
Syntology 5 code samples: 1 from 1 repository linked to this paper, 4 identical to code harvested from another paper's repository. Across all 5: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 2 honoured, 2 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.
By repository: official repository: 1 sample from 1 repository, 0 ran (0 with no instrument failure; 0 where Syntology's instrument failed); 4 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.
No date is recorded for any of the 4 samples that ran: their levels were written before Syntology's graph kept the date.
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
5 samples harvested: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 2 honoured, 2 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 1 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.
Licence: 5 of the 5 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 NUS-HPC-AI-Lab/SGL. 4 samples are listed last, in a group without a repository. “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.
88a4b26411ded833 · report
4 samples are identical to code Syntology harvested from another paper's repository; this paper's own copy was not located, so no repository is named and no licence is recorded.
5f3bdd5f3304a0e0 · report
fixture: signature v1:str/0/str, v2:str/0/str, op:str/0/str · returned bool · level recorded, date not recorded e159a001bd4cbfaa · report
fixture: signature entries:list/1/dict · returned float · level recorded, date not recorded 4ffcad52354b4806 · report
fixture: signature entries:list/1/dict · returned float · level recorded, date not recorded b3c8680ebd8739af · report
fixture: signature target:str/0/str, prediction:str/0/str, max_relative_change:float/0/float · returned bool · level recorded, date not recorded Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 64%) | GPT-4 score | 65.60 | #19 of 231 | Archive leaderboard | report |
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 64%) | Params | 26B | #19 of 231 | Archive leaderboard | report |
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 35%) | GPT-4 score | 63.20 | #28 of 231 | Archive leaderboard | report |
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 35%) | Params | 26B | #28 of 231 | Archive leaderboard | report |
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 9%) | GPT-4 score | 52.10 | #54 of 231 | Archive leaderboard | report |
| Visual Question Answering | MM-Vet | InternVL2-26B (SGP, token ratio 9%) | Params | 26B | #54 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.
Methods
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