{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/pllava-parameter-free-llava-extension-from-1","title":"PLLaVA : Parameter-free LLaVA Extension from Images to Videos for Video Dense Captioning","arxiv_id":"2404.16994","date":"2024-04-25","proceeding":"arXiv 2024 4","authors":["Lin Xu","Yilin Zhao","Daquan Zhou","Zhijie Lin","See Kiong Ng","Jiashi Feng"],"abstract":"Vision-language pre-training has significantly elevated performance across a wide range of image-language applications. Yet, the pre-training process for video-related tasks demands exceptionally large computational and data resources, which hinders the progress of video-language models. This paper investigates a straight-forward, highly efficient, and resource-light approach to adapting an existing image-language pre-trained model for dense video understanding. Our preliminary experiments reveal that directly fine-tuning pre-trained image-language models with multiple frames as inputs on video datasets leads to performance saturation or even a drop. Our further investigation reveals that it is largely attributed to the bias of learned high-norm visual features. Motivated by this finding, we propose a simple but effective pooling strategy to smooth the feature distribution along the temporal dimension and thus reduce the dominant impacts from the extreme features. The new model is termed Pooling LLaVA, or PLLaVA in short. PLLaVA achieves new state-of-the-art performance on modern benchmark datasets for both video question-answer and captioning tasks. Notably, on the recent popular VideoChatGPT benchmark, PLLaVA achieves a score of 3.48 out of 5 on average of five evaluated dimensions, exceeding the previous SOTA results from GPT4V (IG-VLM) by 9%. On the latest multi-choice benchmark MVBench, PLLaVA achieves 58.1% accuracy on average across 20 sub-tasks, 14.5% higher than GPT4V (IG-VLM). Code is available at https://pllava.github.io/","url_abs":"https://arxiv.org/abs/2404.16994v2","url_pdf":"https://arxiv.org/pdf/2404.16994v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"pllava-parameter-free-llava-extension-from-1","repo_url":"https://github.com/magic-research/PLLaVA","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"dense-captioning","task_name":"Dense Captioning"},{"task_slug":null,"task_name":"MVBench"},{"task_slug":"video-question-answering","task_name":"Video Question Answering"},{"task_slug":"video-understanding","task_name":"Video Understanding"},{"task_slug":"video-based-generative-performance","task_name":"Video-based Generative Performance Benchmarking"},{"task_slug":"video-based-generative-performance-5","task_name":"Video-based Generative Performance Benchmarking (Consistency)"},{"task_slug":"video-based-generative-performance-3","task_name":"Video-based Generative Performance Benchmarking (Contextual Understanding)"},{"task_slug":"video-based-generative-performance-1","task_name":"Video-based Generative Performance Benchmarking (Correctness of Information)"},{"task_slug":"video-based-generative-performance-2","task_name":"Video-based Generative Performance Benchmarking (Detail Orientation))"},{"task_slug":"video-based-generative-performance-4","task_name":"Video-based Generative Performance Benchmarking (Temporal Understanding)"},{"task_slug":"zeroshot-video-question-answer","task_name":"Zero-Shot Video Question Answer"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/video-question-answering-on-mvbench","task":"Video Question Answering","dataset":"MVBench","model":"PLLaVA","rank_in_archive_order":11,"of":22,"metrics":{"Avg.":"58.1"},"uses_additional_data":false},{"leaderboard":"/sota/video-question-answering-on-tvbench","task":"Video Question Answering","dataset":"TVBench","model":"PLLaVA-34B","rank_in_archive_order":20,"of":28,"metrics":{"Average Accuracy":"42.3"},"uses_additional_data":false},{"leaderboard":"/sota/video-question-answering-on-tvbench","task":"Video Question Answering","dataset":"TVBench","model":"PLLaVA-13B","rank_in_archive_order":25,"of":28,"metrics":{"Average Accuracy":"36.4"},"uses_additional_data":false},{"leaderboard":"/sota/video-question-answering-on-tvbench","task":"Video Question Answering","dataset":"TVBench","model":"PLLaVA-7B","rank_in_archive_order":28,"of":28,"metrics":{"Average Accuracy":"34.9"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance","task":"Video-based Generative Performance Benchmarking","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":4,"of":23,"metrics":{"Consistency":"3.25","Contextual Understanding":"3.90","Correctness of Information":"3.60","Detail Orientation":"3.20","Temporal Understanding":"2.67","mean":"3.32"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance-2","task":"Video-based Generative Performance Benchmarking (Consistency)","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":5,"of":18,"metrics":{"gpt-score":"3.25"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance-3","task":"Video-based Generative Performance Benchmarking (Contextual Understanding)","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":2,"of":18,"metrics":{"gpt-score":"3.9"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance-1","task":"Video-based Generative Performance Benchmarking (Correctness of Information)","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":2,"of":18,"metrics":{"gpt-score":"3.60"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance-4","task":"Video-based Generative Performance Benchmarking (Detail Orientation))","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":2,"of":18,"metrics":{"gpt-score":"3.20"},"uses_additional_data":false},{"leaderboard":"/sota/video-based-generative-performance-5","task":"Video-based Generative Performance Benchmarking (Temporal Understanding)","dataset":"VideoInstruct","model":"PLLaVA-34B","rank_in_archive_order":6,"of":18,"metrics":{"gpt-score":"2.67"},"uses_additional_data":false},{"leaderboard":"/sota/zeroshot-video-question-answer-on-activitynet","task":"Zero-Shot Video Question Answer","dataset":"ActivityNet-QA","model":"PLLaVA (34B)","rank_in_archive_order":2,"of":28,"metrics":{"Accuracy":"60.9","Confidence Score":"3.7"},"uses_additional_data":false},{"leaderboard":"/sota/zeroshot-video-question-answer-on-msrvtt-qa","task":"Zero-Shot Video Question Answer","dataset":"MSRVTT-QA","model":"PLLaVA (34B)","rank_in_archive_order":2,"of":30,"metrics":{"Accuracy":"68.7","Confidence Score":"3.6"},"uses_additional_data":false},{"leaderboard":"/sota/zeroshot-video-question-answer-on-msvd-qa","task":"Zero-Shot Video Question Answer","dataset":"MSVD-QA","model":"PLLaVA (34B)","rank_in_archive_order":5,"of":28,"metrics":{"Accuracy":"79.9","Confidence Score":"4.2"},"uses_additional_data":false},{"leaderboard":"/sota/zeroshot-video-question-answer-on-tgif-qa","task":"Zero-Shot Video Question Answer","dataset":"TGIF-QA","model":"PLLaVA","rank_in_archive_order":4,"of":14,"metrics":{"Accuracy":"80.6","Confidence Score":"4.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2404.16994","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2404.16994"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/magic-research/PLLaVA","reach":{"status":"ok"}}],"summary":{"unverified":2},"by_repo_kind":{"official":{"samples":2,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"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"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"770e3b5004c211d6","entry":"get_state_maybe_zero_3","repo":"magic-research/PLLaVA","repo_kind":"official","path":"tasks/train/train_pllava_nframe_accel.py","file_url":"https://github.com/magic-research/PLLaVA/blob/HEAD/tasks/train/train_pllava_nframe_accel.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"770e3b5004c211d6"}},{"code_sha256_prefix":"735025744c1ab0cf","entry":"maybe_zero_3","repo":"magic-research/PLLaVA","repo_kind":"official","path":"tasks/train/train_pllava_nframe_accel.py","file_url":"https://github.com/magic-research/PLLaVA/blob/HEAD/tasks/train/train_pllava_nframe_accel.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"735025744c1ab0cf"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}