Why Is Video Still So Expensive? A Survey of Inference-Efficiency Mechanisms in Video and Audiovisual LLMs
- 类型:arxiv
- 标识:2609.10355
- 链接:https://arxiv.org/abs/2609.10355
- 主分类:multimodal
- 形态:survey
- 被引:0
- 被引来源:Semantic Scholar + OpenAlex
- S2被引:0
- OpenAlex被引:0
- 影响力被引:0
- TLDR:This survey covers inference-efficiency mechanisms for visual and audiovisual VideoLLMs that report concrete reductions in parameter count, FLOPs per input, latency, memory, or visual and audio token count, and organize methods by the pipeline stage at which they act.
- OpenAlex ID:W7212148774
- OpenAlex DOI:10.48550/arxiv.2609.10355
- DOI:10.48550/arxiv.2609.10355
- DOI来源:OpenAlex
- 开放获取:green
- 开放获取链接:https://doi.org/10.48550/arxiv.2609.10355
- OpenAlex更新:2026-09-13
- 副分类:llm-infra
- 待LLM分类:否
- 标题中文:视频为何仍然如此昂贵?视频与音视频 LLM 推理效率机制综述
- TLDR中文:本综述涵盖视觉与音视频 VideoLLM 的推理效率机制,这些机制报告了参数数量、每输入 FLOPs、延迟、内存或视觉/音频 token 数量的具体削减量,并按方法所作用的 pipeline 阶段加以组织。
- 来源文件:
- /inbox/tom/_candidates/2026-09-10-agent-rag-longcontext-candidates.json
- [OpenAlex backfill]
- [S2 enrich]