Papers › MELTR: Meta Loss Transformer for Learning to Fine-tune Video Foundation Models

MELTR: Meta Loss Transformer for Learning to Fine-tune Video Foundation Models

23 Mar 2023CVPR 2023 1arXiv:2303.13009archive 2025-07-28

Dohwan Ko, Joonmyung Choi, Hyeong Kyu Choi, Kyoung-Woon On, Byungseok Roh, Hyunwoo J. Kim

Foundation models have shown outstanding performance and generalization capabilities across domains. Since most studies on foundation models mainly focus on the pretraining phase, a naive strategy to minimize a single task-specific loss is adopted for fine-tuning. However, such fine-tuning methods do not fully leverage other losses that are potentially beneficial for the target task. Therefore, we propose MEta Loss TRansformer (MELTR), a plug-in module that automatically and non-linearly combines various loss functions to aid learning the target task via auxiliary learning. We formulate the auxiliary learning as a bi-level optimization problem and present an efficient optimization algorithm based on Approximate Implicit Differentiation (AID). For evaluation, we apply our framework to various video foundation models (UniVL, Violet and All-in-one), and show significant performance gain on all four downstream tasks: text-to-video retrieval, video question answering, video captioning, and multi-modal sentiment analysis. Our qualitative analyses demonstrate that MELTR adequately `transforms' individual loss functions and `melts' them into an effective unified loss. Code is available at https://github.com/mlvlab/MELTR.

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 1 code sample harvested from 1 repository linked to this paper: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample 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: official repository: 1 sample from 1 repository, 1 ran (1 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.

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

mlvlab/MELTR officialmentioned in papermentioned on GitHubpytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 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

1 sample harvested: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample 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.

1ran

Licence: 0 of the 1 sample 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 mlvlab/MELTR. “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.

mlvlab/MELTR official repository 1 sample, 1 ran (0 instrument-failed) · 0 unverified
MELTR univl/modules/meltr.py ran · metamorphic tier: deterministic MIT (permissive) · a3964b84fb27d6e4 · report fixture: returned MELTR · level recorded 2026-09-22

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

  1. MELTR entry · level 1 · ran · a3964b84fb27d6e4

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="2303.13009") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Auxiliary LearningMultimodal Sentiment AnalysisQuestion AnsweringRetrievalSentiment AnalysisTGIF-ActionTGIF-FrameTGIF-TransitionText to Video RetrievalVideo CaptioningVideo Question AnsweringVideo RetrievalVisual Question Answering (VQA)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Multimodal Sentiment Analysis CMU-MOSI UniVL + MELTR Acc-2 85.3 #4 of 12 Archive leaderboard report
Multimodal Sentiment Analysis CMU-MOSI UniVL + MELTR Corr 0.789 #4 of 12 Archive leaderboard report
Multimodal Sentiment Analysis CMU-MOSI UniVL + MELTR F1 85.4 #4 of 12 Archive leaderboard report
Multimodal Sentiment Analysis CMU-MOSI UniVL + MELTR MAE 0.759 #4 of 12 Archive leaderboard report
Video Captioning MSR-VTT UniVL + MELTR BLEU-4 44.17 #24 of 24 Archive leaderboard report
Video Captioning MSR-VTT UniVL + MELTR CIDEr 52.77 #24 of 24 Archive leaderboard report
Video Captioning MSR-VTT UniVL + MELTR METEOR 29.26 #24 of 24 Archive leaderboard report
Video Captioning MSR-VTT UniVL + MELTR ROUGE-L 62.35 #24 of 24 Archive leaderboard report
Video Captioning YouCook2 UniVL + MELTR BLEU-3 24.12 #2 of 14 Archive leaderboard report
Video Captioning YouCook2 UniVL + MELTR BLEU-4 17.92 #2 of 14 Archive leaderboard report
Video Captioning YouCook2 UniVL + MELTR CIDEr 1.90 #2 of 14 Archive leaderboard report
Video Captioning YouCook2 UniVL + MELTR METEOR 22.56 #2 of 14 Archive leaderboard report
Video Captioning YouCook2 UniVL + MELTR ROUGE-L 47.04 #2 of 14 Archive leaderboard report
Video Retrieval MSR-VTT All-in-one + MELTR text-to-video R@1 38.6 #15 of 40 Archive leaderboard report
Video Retrieval MSR-VTT All-in-one + MELTR text-to-video R@10 84.7 #15 of 40 Archive leaderboard report
Video Retrieval MSR-VTT All-in-one + MELTR text-to-video R@5 74.4 #15 of 40 Archive leaderboard report
Video Retrieval MSR-VTT VIOLET + MELTR text-to-video Median Rank 3 #20 of 40 Archive leaderboard report
Video Retrieval MSR-VTT VIOLET + MELTR text-to-video R@1 33.6 #20 of 40 Archive leaderboard report
Video Retrieval MSR-VTT VIOLET + MELTR text-to-video R@10 77.8 #20 of 40 Archive leaderboard report
Video Retrieval MSR-VTT VIOLET + MELTR text-to-video R@5 63.7 #20 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UniVL + MELTR text-to-video Median Rank 4 #28 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UniVL + MELTR text-to-video R@1 28.5 #28 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UniVL + MELTR text-to-video R@10 67.6 #28 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UniVL + MELTR text-to-video R@5 55.5 #28 of 40 Archive leaderboard report
Video Retrieval MSR-VTT-1kA All-in-one + MELTR text-to-video R@1 41.3 #38 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA All-in-one + MELTR text-to-video R@10 82.5 #38 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA All-in-one + MELTR text-to-video R@5 73.5 #38 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VIOLET + MELTR text-to-video Median Rank 3 #46 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VIOLET + MELTR text-to-video R@1 35.5 #46 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VIOLET + MELTR text-to-video R@10 78.4 #46 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VIOLET + MELTR text-to-video R@5 67.2 #46 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA UniVL + MELTR text-to-video Median Rank 4 #48 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA UniVL + MELTR text-to-video R@1 31.1 #48 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA UniVL + MELTR text-to-video R@10 68.3 #48 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA UniVL + MELTR text-to-video R@5 55.7 #48 of 63 Archive leaderboard report
Video Retrieval YouCook2 UniVL + MELTR text-to-video Median Rank 3 #2 of 16 Archive leaderboard report
Video Retrieval YouCook2 UniVL + MELTR text-to-video R@1 33.7 #2 of 16 Archive leaderboard report
Video Retrieval YouCook2 UniVL + MELTR text-to-video R@10 74.8 #2 of 16 Archive leaderboard report
Video Retrieval YouCook2 UniVL + MELTR text-to-video R@5 63.1 #2 of 16 Archive leaderboard report
Visual Question Answering (VQA) MSVD-QA VIOLET + MELTR Accuracy 0.517 #20 of 36 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