This contains the trained checkpoint for LENS, as introduced in LENS: A Learnable Evaluation Metric for Text Simplification (ACL, 2023). For more information, please refer to the LENS repository.
pip install lens-metric
from lens import download_model, LENS
lens_path = download_model("davidheineman/lens")
lens = LENS(lens_path, rescale=True)
complex = [
"They are culturally akin to the coastal peoples of Papua New Guinea."
]
simple = [
"They are culturally similar to the people of Papua New Guinea."
]
references = [[
"They are culturally similar to the coastal peoples of Papua New Guinea.",
"They are similar to the Papua New Guinea people living on the coast."
]]
scores = lens.score(complex, simple, references, batch_size=8, devices=[0])
print(scores) # [78.6344531130125]
For an example, please see the quick demo Google Collab notebook.
Intended uses
This model is for reference-based text simplification evaluation, for a model requiring no references, please see LENS-SALSA.
Cite LENS
If you find our paper, code or data helpful, please consider citing our work:
@inproceedings{maddela-etal-2023-lens,
title = "{LENS}: A Learnable Evaluation Metric for Text Simplification",
author = "Maddela, Mounica and
Dou, Yao and
Heineman, David and
Xu, Wei",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.905",
doi = "10.18653/v1/2023.acl-long.905",
pages = "16383--16408",
}