<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lee-etal-2026-composable">
    <titleInfo>
        <title>A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Hodong</namePart>
        <namePart type="family">Lee</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sanghee</namePart>
        <namePart type="family">Park</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dohoon</namePart>
        <namePart type="family">Ryu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jungwhan</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Junyeob</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Soyoon</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Geewook</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Sweta</namePart>
            <namePart type="family">Agrawal</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alberto</namePart>
            <namePart type="family">Abad</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-461-3</identifier>
    </relatedItem>
    <abstract>Building an omni-modal foundation model means evaluating it across text, image, video, and audio. Excellent evaluation toolkits exist for each modality, but their inference engines, prompt conventions, and metric implementations are mutually incompatible, so practitioners end up maintaining separate environments for every toolchain and still struggle to compare results across them. OmniEvaluator grew out of this need in our own model development: rather than reimplementing benchmarks, it connects existing inference engines and curated evaluation libraries at a higher level, exposing four inference backends, four evaluation frameworks, and over a thousand benchmarks through a single interface. Every run is recorded as an artifact capturing the full configuration for exact reproduction, and results flow into a shared dashboard for cross-model comparison. A federated mode shares GPU inference servers across concurrent evaluations, and a built-in verifier, small enough to run on CPU, keeps its score stable across engines and prompts where rule-based scoring fluctuates under configuration mismatch, matching cost-efficient commercial LLM judges without their recurring API cost. The system, demo video, and dashboard are publicly available. (https://github.com/naver-ai/omni-evaluator, https://www.youtube.com/watch?v=4Z5VZZWyXqY, https://omni-evaluator.info)</abstract>
    <identifier type="citekey">lee-etal-2026-composable</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-demos.70/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>738</start>
            <end>749</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation
%A Lee, Hodong
%A Park, Sanghee
%A Ryu, Dohoon
%A Kim, Jungwhan
%A Kim, Junyeob
%A Kim, Soyoon
%A Kim, Geewook
%Y Agrawal, Sweta
%Y Abad, Alberto
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-461-3
%F lee-etal-2026-composable
%X Building an omni-modal foundation model means evaluating it across text, image, video, and audio. Excellent evaluation toolkits exist for each modality, but their inference engines, prompt conventions, and metric implementations are mutually incompatible, so practitioners end up maintaining separate environments for every toolchain and still struggle to compare results across them. OmniEvaluator grew out of this need in our own model development: rather than reimplementing benchmarks, it connects existing inference engines and curated evaluation libraries at a higher level, exposing four inference backends, four evaluation frameworks, and over a thousand benchmarks through a single interface. Every run is recorded as an artifact capturing the full configuration for exact reproduction, and results flow into a shared dashboard for cross-model comparison. A federated mode shares GPU inference servers across concurrent evaluations, and a built-in verifier, small enough to run on CPU, keeps its score stable across engines and prompts where rule-based scoring fluctuates under configuration mismatch, matching cost-efficient commercial LLM judges without their recurring API cost. The system, demo video, and dashboard are publicly available. (https://github.com/naver-ai/omni-evaluator, https://www.youtube.com/watch?v=4Z5VZZWyXqY, https://omni-evaluator.info)
%U https://aclanthology.org/2026.emnlp-demos.70/
%P 738-749

Markdown (Informal)

[A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation](https://aclanthology.org/2026.emnlp-demos.70/) (Lee et al., EMNLP 2026)

• A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation (Lee et al., EMNLP 2026)

ACL

• Hodong Lee, Sanghee Park, Dohoon Ryu, Jungwhan Kim, Junyeob Kim, Soyoon Kim, and Geewook Kim. 2026. A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 738–749, Budapest, Hungary. Association for Computational Linguistics.