Datasets:
metadata
language:
- en
license: apache-2.0
dataset_info:
features:
- name: id
dtype: int64
- name: anchor
dtype: string
- name: positive
dtype: string
splits:
- name: train
num_bytes: 10865560
num_examples: 10136
- name: test
num_bytes: 2673216
num_examples: 2534
download_size: 6056861
dataset_size: 13538776
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
Dataset Card for Dataset Name
Dataset for finetuning embedding models for patent language.
Dataset Details
Anchor/Positive pairs extracted from text
and text_b
columns with label=1
in the ultrabalanced version of the PatentMatch dataset
- Repository: https://hpi.de/naumann/projects/web-science/paar-patent-analysis-and-retrieval/patentmatch.html
- Paper [optional]: https://arxiv.org/abs/2012.13919
Uses
For fine tuning embedding models.