mattymchen/synli
收藏Hugging Face2022-10-28 更新2024-03-04 收录
下载链接：
https://hf-mirror.com/datasets/mattymchen/synli
下载链接
链接失效反馈官方服务：
资源简介：
---
license: odc-by
dataset_info:
features:
- name: sent0
dtype: string
- name: sent1
dtype: string
- name: hard_neg
dtype: string
splits:
- name: train
num_bytes: 11441750654
num_examples: 60939492
download_size: 6904073153
dataset_size: 11441750654
---
# Dataset Card for SyNLI
A synthetic NLI datasets from open domain sentences using T5 as data synthesizer. The data can be used to train sentence embedding models.
## Data Fields
The data have several fields:
- `sent0`: premise as a string
- `sent1`: entailment hypothesis as a string
- `hard_neg`: contradiction hypothesis as a string
提供机构：
mattymchen
原始信息汇总
数据集卡片：SyNLI
数据集信息
-
特征字段：
sent0：前提，数据类型为字符串sent1：蕴含假设，数据类型为字符串hard_neg：矛盾假设，数据类型为字符串
-
数据分割：
train：训练集，包含60939492个样本，占用11441750654字节
-
数据大小：
- 下载大小：6904073153字节
- 数据集大小：11441750654字节
