AI Paper Atlas

Research map / Video / Video-language understanding

LongVLM: Efficient Long Video Understanding via Large Language Models

Yuetian Weng, Mingfei Han, Haoyu He, Xiaojun Chang, Bohan Zhuang

ECCV 2024 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 109 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace 3D average pooling with hierarchical token merging for local features (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.64 -> 2.72Improved
replace 3D average pooling with token merging (ablation)LongVLM (local only) · Egoschemaaccuracy: 48.0 -> 52.4Improved
add global [CLS] semantic features to local features (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.72 -> 2.89Improved
add global features on Egoschema (ablation)LongVLM · Egoschemaaccuracy: 52.4 -> 57.6Improved
prepend global feature before local features instead of appending (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.82 -> 2.89Improved
increase tokens per segment M from 10 to 30 (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.72 -> 2.89Improved
increase M from 30 to 40 (ablation)LongVLM · ANET-QAaccuracy: 47.6 -> 46.0Hurt
increase M from 30 to 40 (ablation)LongVLM · ANET-QAGPU memory (G): 14.86 -> 14.96Hurt
use CLS tokens from last 5 layers rather than 1 (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.83 -> 2.89Improved
increase E from 5 to 24 CLS layers (ablation)LongVLM · Video-ChatGPT benchmarkMean score: 2.89 -> 2.70Hurt
raise LLM temperature from 0.2 to 2.0 (ablation)LongVLM · Video-ChatGPT benchmarkConsistency: 3.11 -> 1.01Hurt
lower temperature from 0.2 to 0.01 (ablation)LongVLM · Video-ChatGPT benchmarkTemporal Understanding: 2.39 -> 2.22Hurt

Later papers that compared with this method

1 later accepted paper reports LongVLM as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy