metadata
dataset_info:
- config_name: corpus
features:
- name: _id
dtype: string
- name: title
dtype: string
- name: text
dtype: string
splits:
- name: test
num_bytes: 4876445
num_examples: 4307
download_size: 3171857
dataset_size: 4876445
- config_name: default
features:
- name: query-id
dtype: string
- name: corpus-id
dtype: string
- name: score
dtype: int64
splits:
- name: test
num_bytes: 218700
num_examples: 4374
download_size: 164426
dataset_size: 218700
- config_name: queries
features:
- name: _id
dtype: string
- name: text
dtype: string
splits:
- name: test
num_bytes: 797559
num_examples: 4374
download_size: 577217
dataset_size: 797559
configs:
- config_name: corpus
data_files:
- split: test
path: corpus/test-*
- config_name: default
data_files:
- split: test
path: data/test-*
- config_name: queries
data_files:
- split: test
path: queries/test-*
license: cc-by-4.0
IlPost (MTEB retrieval format)
Italian news summary-to-article retrieval task. Given an Italian summary, retrieve the relevant full news article from the corpus. Built from the test split.
Reformatted into MTEB retrieval format from ARTeLab/ilpost, an Italian abstractive summarization dataset of Il Post news articles.
License: CC BY 4.0.