pa-shk/re_dial_processed

该数据集主要包含与电影相关的对话和互动信息。数据集的特征包括电影提及、受访者问题、消息、对话ID、受访者ID、发起者ID、发起者问题、推荐电影、喜欢的电影、对话、格式化对话和嵌入向量等。数据集分为训练集和测试集，分别包含9864和1310个样本。

原始信息汇总

数据集概述

数据集特征

• movieMentions
  • movieId: string
  • movieName: string
• respondentQuestions
  • movieId: string
  • suggested: int32
  • seen: int32
  • liked: int32
• messages
  • timeOffset: int32
  • text: string
  • senderWorkerId: int32
  • messageId: int32
• conversationId: int32
• respondentWorkerId: int32
• initiatorWorkerId: int32
• initiatorQuestions
  • movieId: string
  • suggested: int32
  • seen: int32
  • liked: int32
• recommended_movies: sequence of string
• liked_movies: sequence of string
• dialogs
  • messageId: int64
  • senderWorkerId: int64
  • text: string
  • timeOffset: int64
• formatted_dialogs: sequence of string
• embed: sequence of float32

数据集分割

• train
  • num_bytes: 111451027
  • num_examples: 9864
• test
  • num_bytes: 13820141
  • num_examples: 1310

数据集大小

• download_size: 65951601
• dataset_size: 125271168

配置

• default
  • data_files
    • split: train path: data/train-*
    • split: test path: data/test-*