seungheondoh/if-music-encodec

--- dataset_info: features: - name: ytid dtype: string - name: task dtype: string - name: instruction dtype: string - name: input dtype: string - name: output dtype: string - name: audio_token dtype: string - name: is_audioset_eval dtype: bool splits: - name: qa num_bytes: 1053372614 num_examples: 59691 - name: tag_to_music num_bytes: 95601878 num_examples: 5521 - name: caption_to_music num_bytes: 96425125 num_examples: 5521 - name: music_to_tag num_bytes: 95441769 num_examples: 5521 - name: music_to_caption num_bytes: 96226369 num_examples: 5521 download_size: 464841106 dataset_size: 1437067755 configs: - config_name: default data_files: - split: qa path: data/qa-* - split: tag_to_music path: data/tag_to_music-* - split: caption_to_music path: data/caption_to_music-* - split: music_to_tag path: data/music_to_tag-* - split: music_to_caption path: data/music_to_caption-* ---The dataset includes multiple features such as video ID (ytid), task type (task), instruction (instruction), input (input), output (output), audio token (audio_token), and whether it is for audio set evaluation (is_audioset_eval). The dataset is divided into several subsets including question-answering (qa), tag-to-music, caption-to-music, music-to-tag, and music-to-caption, each with specific data files and example counts. The download size of the dataset is 464,841,106 bytes, and the total size is 1,437,067,755 bytes.

原始信息汇总

数据集概述

数据特征

• ytid: 数据类型为字符串。
• task: 数据类型为字符串。
• instruction: 数据类型为字符串。
• input: 数据类型为字符串。
• output: 数据类型为字符串。
• audio_token: 数据类型为字符串。
• is_audioset_eval: 数据类型为布尔值。

数据分割

• qa: 字节数为1053372614，样本数为59691。
• tag_to_music: 字节数为95601878，样本数为5521。
• caption_to_music: 字节数为96425125，样本数为5521。
• music_to_tag: 字节数为95441769，样本数为5521。
• music_to_caption: 字节数为96226369，样本数为5521。

数据大小

• 下载大小: 464841106字节。
• 数据集大小: 1437067755字节。

配置

• 默认配置: 包含以下数据文件路径：
  • qa: data/qa-*
  • tag_to_music: data/tag_to_music-*
  • caption_to_music: data/caption_to_music-*
  • music_to_tag: data/music_to_tag-*
  • music_to_caption: data/music_to_caption-*