five

Hemabhushan/bridge_network_mlm_hidden_states

收藏
Hugging Face2024-07-22 更新2024-07-22 收录
下载链接:
https://hf-mirror.com/datasets/Hemabhushan/bridge_network_mlm_hidden_states
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集包含多个部分的视频场景数据,每个部分都有详细的配置信息。数据集的特征字段包括视频的标识符、哈希标识符、URL链接、场景开始和结束时间、帧数、关键帧数、动画标签、用户标签、文本描述、美学评分、动态评分、评级、文本概率、分辨率、文件格式、帧率、时间分类、场地分类、媒体分类、拍摄分类、构图分类、角色分类等。此外,还包含与LLM(大语言模型)相关的输出隐藏状态和响应文本。数据集分为训练集,每个部分的训练集大小和样本数量也有所不同。

The dataset contains multiple parts of video scene data, each with detailed configuration information. The feature fields of the dataset include video identifiers, hash identifiers, URL links, scene start and end times, frame numbers, key frame numbers, anime tags, user tags, text descriptions, aesthetic scores, dynamic scores, ratings, text probabilities, resolutions, file formats, frame rates, time taxonomy, venue taxonomy, media taxonomy, filming taxonomy, composition taxonomy, character taxonomy, etc. Additionally, it includes LLM (Large Language Model) related output hidden states and response texts. The dataset is divided into training sets, with varying sizes and sample counts for each part.
提供机构:
Hemabhushan
原始信息汇总

数据集概述

数据集配置

第五部分 (fifth_part)

  • 特征:
    • identifier: string
    • hash_identifier: string
    • url_link: string
    • scene_start_time: string
    • scene_end_time: string
    • frame_number: float64
    • key_frame_number: float64
    • anime_tags: string
    • user_tags: string
    • text_description: string
    • aesthetic_score: float64
    • dynamic_score: float64
    • rating: string
    • text_prob: float64
    • width: int64
    • height: int64
    • file_ext: string
    • fps: float64
    • Taxonomy_Time: string
    • Taxonomy_Venue: string
    • Taxonomy_Media: string
    • Taxonomy_Filming: string
    • Taxonomy_Composition: string
    • Taxonomy_Character: string
    • index_level_0: int64
    • word_count: int64
    • video_id: string
    • video_segment_no: int64
    • llm_output_hidden_states: sequence of sequence of sequence of sequence of float32
    • llm_response_text: string
  • 分割:
    • train: 3004个样本, 2496075507字节
  • 下载大小: 1243687335字节
  • 数据集大小: 2496075507字节

第一部分 (first_part)

  • 特征:
    • identifier: string
    • hash_identifier: string
    • url_link: string
    • scene_start_time: string
    • scene_end_time: string
    • frame_number: float64
    • key_frame_number: float64
    • anime_tags: string
    • user_tags: string
    • text_description: string
    • aesthetic_score: float64
    • dynamic_score: float64
    • rating: string
    • text_prob: float64
    • width: int64
    • height: int64
    • file_ext: string
    • fps: float64
    • Taxonomy_Time: string
    • Taxonomy_Venue: string
    • Taxonomy_Media: string
    • Taxonomy_Filming: string
    • Taxonomy_Composition: string
    • Taxonomy_Character: string
    • index_level_0: int64
    • word_count: int64
    • video_id: string
    • video_segment_no: int64
    • llm_output_hidden_states: sequence of sequence of sequence of sequence of float32
    • llm_response_text: string
  • 分割:
    • train: 2998个样本, 2503715891字节
  • 下载大小: 1246647152字节
  • 数据集大小: 2503715891字节

第四部分 (fourth_part)

  • 特征:
    • identifier: string
    • hash_identifier: string
    • url_link: string
    • scene_start_time: string
    • scene_end_time: string
    • frame_number: float64
    • key_frame_number: float64
    • anime_tags: string
    • user_tags: string
    • text_description: string
    • aesthetic_score: float64
    • dynamic_score: float64
    • rating: string
    • text_prob: float64
    • width: int64
    • height: int64
    • file_ext: string
    • fps: float64
    • Taxonomy_Time: string
    • Taxonomy_Venue: string
    • Taxonomy_Media: string
    • Taxonomy_Filming: string
    • Taxonomy_Composition: string
    • Taxonomy_Character: string
    • index_level_0: int64
    • word_count: int64
    • video_id: string
    • video_segment_no: int64
    • llm_output_hidden_states: sequence of sequence of sequence of sequence of float32
    • llm_response_text: string
  • 分割:
    • train: 3005个样本, 2515835963字节
  • 下载大小: 1252505701字节
  • 数据集大小: 2515835963字节

第二部分 (second_part)

  • 特征:
    • identifier: string
    • hash_identifier: string
    • url_link: string
    • scene_start_time: string
    • scene_end_time: string
    • frame_number: float64
    • key_frame_number: float64
    • anime_tags: string
    • user_tags: string
    • text_description: string
    • aesthetic_score: float64
    • dynamic_score: float64
    • rating: string
    • text_prob: float64
    • width: int64
    • height: int64
    • file_ext: string
    • fps: float64
    • Taxonomy_Time: string
    • Taxonomy_Venue: string
    • Taxonomy_Media: string
    • Taxonomy_Filming: string
    • Taxonomy_Composition: string
    • Taxonomy_Character: string
    • index_level_0: int64
    • word_count: int64
    • video_id: string
    • video_segment_no: int64
    • llm_output_hidden_states: sequence of sequence of sequence of sequence of float32
    • llm_response_text: string
  • 分割:
    • train: 3004个样本, 2587034424字节
  • 下载大小: 1288190767字节
  • 数据集大小: 2587034424字节

第三部分 (third_part)

  • 特征:
    • identifier: string
    • hash_identifier: string
    • url_link: string
    • scene_start_time: string
    • scene_end_time: string
    • frame_number: float64
    • key_frame_number: float64
    • anime_tags: string
    • user_tags: string
    • text_description: string
    • aesthetic_score: float64
    • dynamic_score: float64
    • rating: string
    • text_prob: float64
    • width: int64
    • height: int64
    • file_ext: string
    • fps: float64
    • Taxonomy_Time: string
    • Taxonomy_Venue: string
    • Taxonomy_Media: string
    • Taxonomy_Filming: string
    • Taxonomy_Composition: string
    • Taxonomy_Character: string
    • index_level_0: int64
    • word_count: int64
    • video_id: string
    • video_segment_no: int64
    • llm_output_hidden_states: sequence of sequence of sequence of sequence of float32
    • llm_response_text: string
  • 分割:
    • train: 2999个样本, 2475896528字节
  • 下载大小: 1233461894字节
  • 数据集大小: 2475896528字节
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作