andersonbcdefg/nli_pairs_v2

---
dataset_info:
features:
- name: query
dtype: string
- name: pos
dtype: string
splits:
- name: train
num_bytes: 729187184.8374316
num_examples: 903218
download_size: 204019551
dataset_size: 729187184.8374316
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---
# nli pairs
j = datasets.load_dataset("andersonbcdefg/jina_negation_v2", split="train").select_columns(["query", "pos"])
syn = datasets.load_dataset("andersonbcdefg/synthetic_nli_combined_mnli_filtered", split="train").select_columns(["query", "pos"])
a = datasets.load_dataset("andersonbcdefg/anli_triples", split="train").select_columns(["query", "pos"])
sim = datasets.load_dataset("andersonbcdefg/simcse_nli", split="train").select_columns(["query", "pos"])
doc = datasets.load_dataset("andersonbcdefg/doc_nli_pos_pairs", split="train")
fever = datasets.load_dataset("pietrolesci/nli_fever", split="train").filter(lambda x: x["label"] == 0).map(
lambda x: {"query": x["premise"], "pos": x["hypothesis"]}
).select_columns(["query", "pos"])
ling = datasets.load_dataset("metaeval/lingnli", split="train").filter(lambda x: x["label"] == "entailment").map(
lambda x: {"query": x["premise"], "pos": x["hypothesis"]}
).select_columns(["query", "pos"])

原始信息汇总

数据集概述

数据集特征

• 名称: query
  • 数据类型: string
• 名称: pos
  • 数据类型: string

数据分割

• 名称: train
  • 字节数: 729187184.8374316
  • 样本数: 903218

数据集大小

• 下载大小: 204019551
• 数据集大小: 729187184.8374316

配置

• 配置名称: default
  • 数据文件:
    • 分割: train
    • 路径: data/train-*