|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: condition |
|
|
dtype: string |
|
|
- name: group |
|
|
sequence: string |
|
|
- name: __index_level_0__ |
|
|
dtype: int64 |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 736350 |
|
|
num_examples: 2759 |
|
|
- name: validation |
|
|
num_bytes: 84745 |
|
|
num_examples: 345 |
|
|
- name: test |
|
|
num_bytes: 92647 |
|
|
num_examples: 345 |
|
|
download_size: 503833 |
|
|
dataset_size: 913742 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
- split: validation |
|
|
path: data/validation-* |
|
|
- split: test |
|
|
path: data/test-* |
|
|
--- |
|
|
|
|
|
## 介绍 |
|
|
quora的这个检索数据集并不怎么好, query和corpus都时问题,而且没有label为0的qerl |
|
|
然后才一共出产了3449条数据, 有点少了. |
|
|
这不是我理想的数据集. |
|
|
|
|
|
一共囊括了11750句语料,3449条query,平均1个query对应3.4个corpus,组合的话,能有14000条数据. |
