Papers › BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning

BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning

27 Sep 2023CVPR 2024 1arXiv:2309.15785archive 2025-07-28

Ruyang Liu, Chen Li, Yixiao Ge, Ying Shan, Thomas H. Li, Ge Li

The recent progress in Large Language Models (LLM) has spurred various advancements in image-language conversation agents, while how to build a proficient video-based dialogue system is still under exploration. Considering the extensive scale of LLM and visual backbone, minimal GPU memory is left for facilitating effective temporal modeling, which is crucial for comprehending and providing feedback on videos. To this end, we propose Branching Temporal Adapter (BT-Adapter), a novel method for extending image-language pretrained models into the video domain. Specifically, BT-Adapter serves as a plug-and-use temporal modeling branch alongside the pretrained visual encoder, which is tuned while keeping the backbone frozen. Just pretrained once, BT-Adapter can be seamlessly integrated into all image conversation models using this version of CLIP, enabling video conversations without the need for video instructions. Besides, we develop a unique asymmetric token masking strategy inside the branch with tailor-made training tasks for BT-Adapter, facilitating faster convergence and better results. Thanks to BT-Adapter, we are able to empower existing multimodal dialogue models with strong video understanding capabilities without incurring excessive GPU costs. Without bells and whistles, BT-Adapter achieves (1) state-of-the-art zero-shot results on various video tasks using thousands of fewer GPU hours. (2) better performance than current video chatbots without any video instruction tuning. (3) state-of-the-art results of video chatting using video instruction tuning, outperforming previous SOTAs by a large margin.

PaperPDFConference PDFCodeCode 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 4 code samples harvested from 1 repository linked to this paper: 3 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; 1 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 4 samples from 1 repository, 3 ran (2 with no instrument failure; 1 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 between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

farewellthree/BT-Adapter officialmentioned in papermentioned on GitHubpytorchApache-2.0Syntology: 4 samples, 3 ran (1 instrument-failed) · 1 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

4 samples harvested: 3 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; 1 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 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
1ran · violated contract
1ran; the contract check could not run
1unverified

Licence: 2 of the 4 samples are pointer only, meaning Syntology does not serve that copy's text. For 2 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from farewellthree/BT-Adapter. “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.

farewellthree/BT-Adapter official repository 4 samples, 3 ran (1 instrument-failed) · 1 unverified
get_seq_frames in 6 papers llava_base/bt_adapter/eval/model_utils.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · 5764189adab1c8ff · report fixture: signature total_num_frames:int/0/int, desired_num_frames:int/0/int · returned list · level recorded 2026-09-25
get_spatio_temporal_features_torch llava_base/bt_adapter/inference.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted Apache-2.0 (permissive) here; the stored copy is pointer only · 37fcf26fc4a105af · report fixture: signature features:Tensor/3/float · returned Tensor · level recorded 2026-09-22
violates_moderation in 59 papers llava_base/bt_adapter/utils.py ran · violated contract Apache-2.0 (permissive) here; the stored copy is pointer only · f9939a84b9a65279 · report fixture: signature text:str/0/str · returned bool · level recorded 2026-09-26
pretty_print_semaphore in 55 papers llava_base/bt_adapter/utils.py unverified Apache-2.0 (permissive) · 37899f22fb191b37 · report

Tasks

Video Question AnsweringVideo UnderstandingVideo-based Generative Performance BenchmarkingVideo-based Generative Performance Benchmarking (Consistency)Video-based Generative Performance Benchmarking (Contextual Understanding)Video-based Generative Performance Benchmarking (Correctness of Information)Video-based Generative Performance Benchmarking (Detail Orientation))Video-based Generative Performance Benchmarking (Temporal Understanding)Zero-Shot Video Question AnswerZero-Shot Video Retrieval

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
VCGBench-Diverse VideoInstruct BT-Adapter Consistency 2.27 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Contextual Understanding 2.59 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Correctness of Information 2.20 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Dense Captioning 1.03 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Detail Orientation 2.62 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Reasoning 3.62 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Spatial Understanding 2.35 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter Temporal Understanding 1.29 #4 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct BT-Adapter mean 2.19 #4 of 6 Archive leaderboard report
Video Question Answering ActivityNet-QA BT-Adapter (zero-shot) Accuracy 46.1 #14 of 36 Archive leaderboard report
Video Question Answering ActivityNet-QA BT-Adapter (zero-shot) Confidence score 3.6 #14 of 36 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter Consistency 2.46 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter Contextual Understanding 3.27 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter Correctness of Information 2.68 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter Detail Orientation 2.69 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter Temporal Understanding 2.34 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter mean 2.69 #18 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) Consistency 2.2 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) Contextual Understanding 2.89 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) Correctness of Information 2.16 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) Detail Orientation 2.46 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) Temporal Understanding 2.13 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct BT-Adapter (zero-shot) mean 2.46 #19 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking (Consistency) VideoInstruct BT-Adapter gpt-score 2.46 #12 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Consistency) VideoInstruct BT-Adapter (zero-shot) gpt-score 2.2 #16 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Contextual Understanding) VideoInstruct BT-Adapter gpt-score 3.27 #12 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Contextual Understanding) VideoInstruct BT-Adapter (zero-shot) gpt-score 2.89 #14 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Correctness of Information) VideoInstruct BT-Adapter gpt-score 2.68 #13 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Correctness of Information) VideoInstruct BT-Adapter (zero-shot) gpt-score 2.16 #16 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Detail Orientation)) VideoInstruct BT-Adapter gpt-score 2.69 #13 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Detail Orientation)) VideoInstruct BT-Adapter (zero-shot) gpt-score 2.46 #16 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Temporal Understanding) VideoInstruct BT-Adapter gpt-score 2.34 #12 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Temporal Understanding) VideoInstruct BT-Adapter (zero-shot) gpt-score 2.13 #14 of 18 Archive leaderboard report
Zero-Shot Video Question Answer ActivityNet-QA BT-Adapter (zero-shot) Accuracy 46.1 #20 of 28 Archive leaderboard report
Zero-Shot Video Question Answer ActivityNet-QA BT-Adapter (zero-shot) Confidence Score 3.2 #20 of 28 Archive leaderboard report
Zero-Shot Video Question Answer MSRVTT-QA BT-Adapter (zero-shot) Accuracy 51.2 #25 of 30 Archive leaderboard report
Zero-Shot Video Question Answer MSRVTT-QA BT-Adapter (zero-shot) Confidence Score 2.9 #25 of 30 Archive leaderboard report
Zero-Shot Video Question Answer MSRVTT-QA BT-Adapter (zero-shot) Accuracy 51.2 #26 of 30 Archive leaderboard report
Zero-Shot Video Question Answer MSRVTT-QA BT-Adapter (zero-shot) Confidence Score 2.9 #26 of 30 Archive leaderboard report
Zero-Shot Video Question Answer MSVD-QA BT-Adapter (zero-shot) Accuracy 67.0 #22 of 28 Archive leaderboard report
Zero-Shot Video Question Answer MSVD-QA BT-Adapter (zero-shot) Confidence Score 3.6 #22 of 28 Archive leaderboard report
Zero-Shot Video Question Answer MSVD-QA BT-Adapter (zero-shot) Accuracy 67.0 #23 of 28 Archive leaderboard report
Zero-Shot Video Question Answer MSVD-QA BT-Adapter (zero-shot) Confidence Score 3.6 #23 of 28 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet BT-Adapter text-to-video R@1 37.0 #8 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet BT-Adapter text-to-video R@10 78.9 #8 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet BT-Adapter text-to-video R@5 66.7 #8 of 12 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo BT-Adapter text-to-video R@1 35.6 #14 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo BT-Adapter text-to-video R@10 72.6 #14 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo BT-Adapter text-to-video R@5 61.9 #14 of 26 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC BT-Adapter text-to-video R@1 19.5 #6 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC BT-Adapter text-to-video R@10 45.0 #6 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC BT-Adapter text-to-video R@5 35.9 #6 of 16 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT BT-Adapter text-to-video R@1 40.9 #13 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT BT-Adapter text-to-video R@10 73.5 #13 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT BT-Adapter text-to-video R@5 64.7 #13 of 41 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

AdapterCLIP

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