General

• Managing Python version: asdf
• Python dependancy/project management: poetry
• Data/model versioning/management: dvc
• Auto-linting: black

tfbert

Go away

Working with it

https://www.kaggle.com/aditya08/tf2-2-fine-tuning-bert-for-seq-classification/

It uses poetry. After cloning,

poetry install

If you do not have python 3.8 as the default python or a virtual environment based on python 3.8, you might run into some python version issues. In such cases, follow the below steps -

# install virtualenv and pyenv
brew install pyenv pyenv-virtualenv
# install python3.8.0
pyenv install 3.8.0
# create a virtualenv with python3.8.0
pyenv virtualenv 3.8.0 dl_env
# activate the environment
source ~/.pyenv/versions/dl_env/bin/activate
# install dependencies
poetry install

To train the model,

poetry run python main.py

Currently stuck at current_error.png

Later

https://www.kaggle.com/aditya08/tf2-2-k-fold-bert-fine-tuning