Tanqizheng/auto-hl
Auto-HL Dataset 轻量视界（LIGHTVISION）训练数据 · 端侧视频高光检测 由华东师范大学大创项目团队发布。本仓库只包含标注结果与视频索引，不包含原始视频文件（各源视频受其原有许可约束，见文末）。 教师：InternVideo2.5-8B 自动标注 · 学生：MobileViCLIP S0（17.3M 参数）蒸馏 概览 Auto-HL 是一个面向端侧（on-device）视频高光检测的大规模自动标注基准：由 8B 教师模型 InternVideo2.5 自动标注，再蒸馏训练 17.3M 参数的轻量学生模型 MobileViCLIP-S0。配套论文题为 Auto-HL: A 21K-Video Auto-Annotated Benchmark for On-Device Video Highlight Detection。 核心贡献： 首个面向端侧高光检测的大规模自动标注基准：21,375 视频 / 70,282 窗口，比 TVSum（50 视频）、YouTube Highlights（433… See the full description on the dataset page: https://huggingface.co/datasets/Tanqizheng/auto-hl.
| Author / organization | Tanqizheng |
|---|---|
| License (as declared) | cc-by-nc-4.0 |
| Downloads | 152 (as reported by Hugging Face) |
| Likes | 1 |
| Created | 1 October 2026 |
| Last updated at source | 7 October 2026 |
Popularity, last 3 days
152 now (+9 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.