Skip to content

Latest commit

 

History

History
59 lines (41 loc) · 1.34 KB

README.md

File metadata and controls

59 lines (41 loc) · 1.34 KB

spaCy tuTorial

Instructions

This is built to run in Jupyter notebooks on Google Colab. See the slides at https://derwen.ai/s/d5c7 for details.

Google Colab

Environment dependencies

If you're not running on Google Colab, first for GCP set up for Jupyter and Python 3.x:

sudo bash
apt-get update
apt-get install python-dev python3-dev

pushd /usr/bin/
rm -r python
ln -s python3 python

rm -f /etc/boto.cfg

exit

That assumes the Python binary is located at /usr/bin/python3 -- change it as needed.

Then install conda from its site.

conda install -c conda-forge spacy

python -m spacy download en_core_web_sm
#python -m spacy download en_core_web_md
#python -m download en_core_web_lg
python -m spacy validate

conda install -c conda-forge scattertext
conda install -c anaconda beautifulsoup4

pip install spacy-wordnet

#pip install spacy-transformers
#python -m spacy download en_trf_bertbaseuncased_lg

Jupyter launch

Be sure to:

source ~/.bashrc

Then to launch a headless Jupyter instance (e.g., running on a remote server in the cloud):

nohup ~/anaconda3/bin/jupyter notebook --no-browser --port=8888 --NotebookApp.token='' --ip='0.0.0.0' &