chan030609/MUSE-News-2
收藏Hugging Face2024-06-15 更新2024-06-29 收录
下载链接：
https://hf-mirror.com/datasets/chan030609/MUSE-News-2
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
- config_name: knowmem
features:
- name: answer
dtype: string
- name: question
dtype: string
splits:
- name: retain2_qa_icl
num_bytes: 1265
num_examples: 10
- name: forget_qa_icl
num_bytes: 1154
num_examples: 10
- name: retain2_qa
num_bytes: 11425
num_examples: 100
- name: forget_qa
num_bytes: 11019
num_examples: 100
download_size: 26647
dataset_size: 24863
- config_name: main
features:
- name: text
dtype: string
splits:
- name: forget
num_bytes: 3281546
num_examples: 889
- name: holdout
num_bytes: 9095347
num_examples: 3043
- name: retain2
num_bytes: 6437312
num_examples: 1778
- name: retain1
num_bytes: 6456895
num_examples: 1777
download_size: 14881285
dataset_size: 25271100
- config_name: scal
features:
- name: text
dtype: string
splits:
- name: forget_4
num_bytes: 13219611
num_examples: 3554
- name: forget_1
num_bytes: 3281546
num_examples: 889
- name: forget_2
num_bytes: 6551494
num_examples: 1778
- name: forget_3
num_bytes: 9874079
num_examples: 2667
download_size: 19240874
dataset_size: 32926730
- config_name: sust
features:
- name: text
dtype: string
splits:
- name: forget_4
num_bytes: 3345532
num_examples: 887
- name: forget_1
num_bytes: 3281546
num_examples: 889
- name: forget_2
num_bytes: 3269948
num_examples: 889
- name: forget_3
num_bytes: 3322585
num_examples: 889
download_size: 7721364
dataset_size: 13219611
- config_name: train
features:
- name: text
dtype: string
splits:
- name: forget
num_bytes: 13219611
num_examples: 3554
- name: retain2
num_bytes: 6437312
num_examples: 1778
- name: retain1
num_bytes: 6456895
num_examples: 1777
download_size: 15207155
dataset_size: 26113818
configs:
- config_name: knowmem
data_files:
- split: retain2_qa_icl
path: knowmem/retain2_qa_icl-*
- split: forget_qa_icl
path: knowmem/forget_qa_icl-*
- split: retain2_qa
path: knowmem/retain2_qa-*
- split: forget_qa
path: knowmem/forget_qa-*
- config_name: main
data_files:
- split: forget
path: main/forget-*
- split: holdout
path: main/holdout-*
- split: retain2
path: main/retain2-*
- split: retain1
path: main/retain1-*
- config_name: scal
data_files:
- split: forget_4
path: scal/forget_4-*
- split: forget_1
path: scal/forget_1-*
- split: forget_2
path: scal/forget_2-*
- split: forget_3
path: scal/forget_3-*
- config_name: sust
data_files:
- split: forget_4
path: sust/forget_4-*
- split: forget_1
path: sust/forget_1-*
- split: forget_2
path: sust/forget_2-*
- split: forget_3
path: sust/forget_3-*
- config_name: train
data_files:
- split: forget
path: train/forget-*
- split: retain2
path: train/retain2-*
- split: retain1
path: train/retain1-*
---
提供机构：
chan030609
原始信息汇总
数据集概述
数据集配置
配置：knowmem
- 特征：
answer：字符串类型question：字符串类型
- 分割：
retain2_qa_icl：1265字节，10个样本forget_qa_icl：1154字节，10个样本retain2_qa：11425字节，100个样本forget_qa：11019字节，100个样本
- 下载大小：26647字节
- 数据集大小：24863字节
配置：main
- 特征：
text：字符串类型
- 分割：
forget：3281546字节，889个样本holdout：9095347字节，3043个样本retain2：6437312字节，1778个样本retain1：6456895字节，1777个样本
- 下载大小：14881285字节
- 数据集大小：25271100字节
配置：scal
- 特征：
text：字符串类型
- 分割：
forget_4：13219611字节，3554个样本forget_1：3281546字节，889个样本forget_2：6551494字节，1778个样本forget_3：9874079字节，2667个样本
- 下载大小：19240874字节
- 数据集大小：32926730字节
配置：sust
- 特征：
text：字符串类型
- 分割：
forget_4：3345532字节，887个样本forget_1：3281546字节，889个样本forget_2：3269948字节，889个样本forget_3：3322585字节，889个样本
- 下载大小：7721364字节
- 数据集大小：13219611字节
配置：train
- 特征：
text：字符串类型
- 分割：
forget：13219611字节，3554个样本retain2：6437312字节，1778个样本retain1：6456895字节，1777个样本
- 下载大小：15207155字节
- 数据集大小：26113818字节
数据文件路径
配置：knowmem
retain2_qa_icl：knowmem/retain2_qa_icl-*forget_qa_icl：knowmem/forget_qa_icl-*retain2_qa：knowmem/retain2_qa-*forget_qa：knowmem/forget_qa-*
配置：main
forget：main/forget-*holdout：main/holdout-*retain2：main/retain2-*retain1：main/retain1-*
配置：scal
forget_4：scal/forget_4-*forget_1：scal/forget_1-*forget_2：scal/forget_2-*forget_3：scal/forget_3-*
配置：sust
forget_4：sust/forget_4-*forget_1：sust/forget_1-*forget_2：sust/forget_2-*forget_3：sust/forget_3-*
配置：train
forget：train/forget-*retain2：train/retain2-*retain1：train/retain1-*
