C. Sun, X. Qiu, Y. Xu, and X. Huang. Chinese Computational Linguistics, page 194--206. Cham, Springer International Publishing, (2019)
Abstract
Language model pre-training has proven to be useful in learning universal language representations. As a state-of-the-art language model pre-training model, BERT (Bidirectional Encoder Representations from Transformers) has achieved amazing results in many language understanding tasks. In this paper, we conduct exhaustive experiments to investigate different fine-tuning methods of BERT on text classification task and provide a general solution for BERT fine-tuning. Finally, the proposed solution obtains new state-of-the-art results on eight widely-studied text classification datasets.
%0 Conference Paper
%1 sun2019finetune
%A Sun, Chi
%A Qiu, Xipeng
%A Xu, Yige
%A Huang, Xuanjing
%B Chinese Computational Linguistics
%C Cham
%D 2019
%E Sun, Maosong
%E Huang, Xuanjing
%E Ji, Heng
%E Liu, Zhiyuan
%E Liu, Yang
%I Springer International Publishing
%K bert deep fine learning transfer tuning
%P 194--206
%T How to Fine-Tune BERT for Text Classification?
%X Language model pre-training has proven to be useful in learning universal language representations. As a state-of-the-art language model pre-training model, BERT (Bidirectional Encoder Representations from Transformers) has achieved amazing results in many language understanding tasks. In this paper, we conduct exhaustive experiments to investigate different fine-tuning methods of BERT on text classification task and provide a general solution for BERT fine-tuning. Finally, the proposed solution obtains new state-of-the-art results on eight widely-studied text classification datasets.
%@ 978-3-030-32381-3
@inproceedings{sun2019finetune,
abstract = {Language model pre-training has proven to be useful in learning universal language representations. As a state-of-the-art language model pre-training model, BERT (Bidirectional Encoder Representations from Transformers) has achieved amazing results in many language understanding tasks. In this paper, we conduct exhaustive experiments to investigate different fine-tuning methods of BERT on text classification task and provide a general solution for BERT fine-tuning. Finally, the proposed solution obtains new state-of-the-art results on eight widely-studied text classification datasets.},
added-at = {2020-03-02T10:18:58.000+0100},
address = {Cham},
author = {Sun, Chi and Qiu, Xipeng and Xu, Yige and Huang, Xuanjing},
biburl = {https://www.bibsonomy.org/bibtex/200a330c32642b180b2e4091ec42f88c6/nosebrain},
booktitle = {Chinese Computational Linguistics},
editor = {Sun, Maosong and Huang, Xuanjing and Ji, Heng and Liu, Zhiyuan and Liu, Yang},
interhash = {16a2b2555e5c90895faeec9650304453},
intrahash = {00a330c32642b180b2e4091ec42f88c6},
isbn = {978-3-030-32381-3},
keywords = {bert deep fine learning transfer tuning},
pages = {194--206},
publisher = {Springer International Publishing},
timestamp = {2020-03-02T10:19:24.000+0100},
title = {How to Fine-Tune BERT for Text Classification?},
year = 2019
}