Reza-Madani/FaithDial_HardHal

---
dataset_info:
features:
- name: dialog_idx
dtype: int32
- name: faithdial_response
dtype: string
- name: response
dtype: string
- name: history
sequence: string
- name: knowledge
dtype: string
- name: BEGIN
sequence: string
- name: VRM
sequence: string
splits:
- name: test
num_bytes: 648189.9384006781
num_examples: 828
- name: test_random_split
num_bytes: 301224.9055944056
num_examples: 381
- name: test_topic_split
num_bytes: 347281.72956664837
num_examples: 447
- name: train
num_bytes: 2247686.7684262134
num_examples: 2867
- name: validation
num_bytes: 624910.3453321627
num_examples: 790
- name: valid_random_split
num_bytes: 255306.94597839136
num_examples: 316
- name: valid_topic_split
num_bytes: 367558.17018846376
num_examples: 474
download_size: 2426029
dataset_size: 4792158.803486964
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: test_random_split
path: data/test_random_split-*
- split: test_topic_split
path: data/test_topic_split-*
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: valid_random_split
path: data/valid_random_split-*
- split: valid_topic_split
path: data/valid_topic_split-*
---

The dataset includes multiple features such as dialog index, faithdial response, response, history, knowledge, etc., and different splits of the dataset like train, validation, and test sets. The size and configuration of the dataset are also provided in detail.

原始信息汇总

数据集概述

数据特征

• dialog_idx: 数据类型为 int32
• faithdial_response: 数据类型为 string
• response: 数据类型为 string
• history: 数据类型为 string 的序列
• knowledge: 数据类型为 string
• BEGIN: 数据类型为 string 的序列
• VRM: 数据类型为 string 的序列

数据分割

• test: 字节数为 648189.9384006781，样本数为 828
• test_random_split: 字节数为 301224.9055944056，样本数为 381
• test_topic_split: 字节数为 347281.72956664837，样本数为 447
• train: 字节数为 2247686.7684262134，样本数为 2867
• validation: 字节数为 624910.3453321627，样本数为 790
• valid_random_split: 字节数为 255306.94597839136，样本数为 316
• valid_topic_split: 字节数为 367558.17018846376，样本数为 474

数据集大小

• 下载大小: 2426029 字节
• 数据集大小: 4792158.803486964 字节

配置信息

• 配置名称: default
  • 数据文件:
    • test: 路径为 data/test-*
    • test_random_split: 路径为 data/test_random_split-*
    • test_topic_split: 路径为 data/test_topic_split-*
    • train: 路径为 data/train-*
    • validation: 路径为 data/validation-*
    • valid_random_split: 路径为 data/valid_random_split-*
    • valid_topic_split: 路径为 data/valid_topic_split-*