Ftest/VTdataset
收藏Hugging Face2024-04-21 更新2024-06-12 收录
下载链接：
https://hf-mirror.com/datasets/Ftest/VTdataset
下载链接
链接失效反馈官方服务：
资源简介：
---
license: apache-2.0
configs:
- config_name: default
data_files:
- split: labels
path: "vtllama3_cleaned.json"
---
# Dataset Card for Dataset Name
Youtube clips video data processed for conversational llava model.
This dataset card aims to be a base template for new datasets. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1).
### Dataset Description
Video data are segmented into intervals of 30 seconds. Each interval is converted into a collage of 3 x 3 frames uniformaly selected.
Dataset is generated in two-folds:
1) Basic Llava model tasked with describing the 3 x 3 collage.
2) Llama 3 prompted with image description + video transcription + Character card "Maple" to generate a conversational chain.
提供机构：
Ftest
原始信息汇总
数据集卡片 for Dataset Name
数据集描述
视频数据被分割成30秒的间隔。每个间隔被转换成一个3x3帧的拼贴画，均匀选择。数据集以两种方式生成：
- 基本的Llava模型被任务描述3x3的拼贴画。
- Llama 3被提示使用图像描述 + 视频转录 + 角色卡片“Maple”来生成对话链。
