uds-lsv / bert-stable-fine-tuning Public

Notifications You must be signed in to change notification settings
Fork 21
Star 133

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

arxiv.org/abs/2006.04884

Apache-2.0 license

133 stars 21 forks Branches Tags Activity

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
docker		docker
docs		docs
examples		examples
model_cards		model_cards
notebooks		notebooks
src/transformers		src/transformers
templates		templates
tests		tests
utils		utils
CONTRIBUTING.md		CONTRIBUTING.md
LICENSE		LICENSE
MANIFEST.in		MANIFEST.in
Makefile		Makefile
README.md		README.md
README_transformers.md		README_transformers.md
deploy_multi_version_doc.sh		deploy_multi_version_doc.sh
hubconf.py		hubconf.py
setup.cfg		setup.cfg
setup.py		setup.py
transformers-cli		transformers-cli
valohai.yaml		valohai.yaml

Repository files navigation

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

Marius Mosbach, Maksym Andriushchenko, Dietrich Klakow

Saarland University and EPFL

This repository contains code for the paper On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines. It is a fork of the Huggingface Transformers repository (v2.5.1).

Reproducing our results

See /examples/bert_stable_fine_tuning/README.md for how to setup Docker and run our models.

About

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

arxiv.org/abs/2006.04884

nlp bert fine-tuning

Apache-2.0 license

Custom properties

Report repository

Releases

No releases published

Packages

No packages published

Languages