Neural Language Models for Nineteenth-Century English: bert_1760_1900

Introduction

BERT model trained on a large historical dataset of books in English, published between 1760-1900 and comprised of ~5.1 billion tokens.

• Data paper: http://doi.org/10.5334/johd.48
• Github repository: https://github.com/Living-with-machines/histLM

License

The models are released under open license CC BY 4.0, available at https://creativecommons.org/licenses/by/4.0/legalcode.

Funding Statement

This work was supported by Living with Machines (AHRC grant AH/S01179X/1) and The Alan Turing Institute (EPSRC grant EP/N510129/1).

Dataset creators

Kasra Hosseini, Kaspar Beelen and Mariona Coll Ardanuy (The Alan Turing Institute) preprocessed the text, created a database, trained and fine-tuned language models as described in the accompanying paper. Giovanni Colavizza (University of Amsterdam), David Beavan (The Alan Turing Institute) and James Hetherington (University College London) helped with planning, accessing the datasets and designing the experiments.