hdhacker/connan_30k
收藏Hugging Face2026-04-28 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/hdhacker/connan_30k
下载链接
链接失效反馈官方服务：
资源简介：
---
license: other
task_categories:
- text-to-video
language:
- en
tags:
- video
- webdataset
- parquet
private_backup: true
---
# Connan 30K Video Caption Dataset
Private Hugging Face backup of 30,617 short anime video clips and English visual captions.
## Layout
```text
metadata/manifest.parquet # sample index, captions, original paths, shard/member mapping
wds/*.tar # WebDataset shards, about 10GB each
dataset_info.json # summary metadata
```
Each WebDataset sample uses a stable key such as `001_shot_00003` and contains:
```text
001_shot_00003.mp4
001_shot_00003.json
```
The JSON sidecar contains the caption and original source paths. The Parquet manifest is the canonical table for captions and shard lookup.
## Summary
- Samples: 30617
- Shards: 10
- Caption language: English
- Recommended LoRA trigger: `DCANIME`
Captions are generic visual descriptions and intentionally avoid character names, franchise names, and trigger words.
提供机构：
hdhacker
