xuyige/BERT4doc-Classification
A research implementation investigating optimal methods for fine-tuning BERT on text classification.
Not currently ranked — collecting fresh signals.
star history
This repository contains the code and experiments for a paper studying how to fine-tune BERT for text classification tasks. The work explores different fine-tuning strategies and hyperparameters for BERT, a pre-trained language model, when adapting it for downstream classification tasks. It supports multiple datasets including Sogou News and other text classification benchmarks, and provides checkpoints compatible with the fastNLP library’s BertEmbedding module.
Frequently asked
- What is xuyige/BERT4doc-Classification?
- A research implementation investigating optimal methods for fine-tuning BERT on text classification.
- Is BERT4doc-Classification open source?
- Yes — xuyige/BERT4doc-Classification is open source, released under the Apache-2.0 license.
- What language is BERT4doc-Classification written in?
- xuyige/BERT4doc-Classification is primarily written in Python.
- How popular is BERT4doc-Classification?
- xuyige/BERT4doc-Classification has 640 stars on GitHub.
- Where can I find BERT4doc-Classification?
- xuyige/BERT4doc-Classification is on GitHub at https://github.com/xuyige/BERT4doc-Classification.