In this workshop, we would learn how to automatically style ( bold , Italics, etc. ) a word according to context.
We learn styling from html files automatically and apply them to raw text.
This project is used mainly to demonstrate deep-learning implementation of named-entity-recognition (NER) models.
- Google Colab notebooks (and other resources) are located in Google Drive under
Colab Notebooks
directory.
If you are using Colab for the first time, open Colab and save one of the example notebooks. The notebook will be saved toColab Notebooks
directory. - Upload folder (repo content + data zip file) to your Google Drive. Make sure
nlp_ner_workshop
folder is located in yourColab Notebooks
folder.
Due to Google Drive quota issues make sure not to unzip the data file. - Open one of the example notebooks, change the
GOOGLE_COLAB
toTrue
, and run all to test it. - You might need to configure your
Runtime type
toPython 3
and set theHardware accelerator
toGPU
. Both located in Runtime=>Change runtime type.
- Make sure Python3 is installed.
- You can create a virtual environment (recommended) using
python3 -m virtualenv ner_ws
- To activate your virtual env, run:
source ner_ws/bin/activate
- Now install all of the requirements:
pip3 install -r requirements.txt
- Download data from our Google drive
- Save the
.zip
file in thedata/
folder. - Run
style_learn.py
to train an NER model. - Run
server.py
to evaluate your model in the browser.
For more details, contact me at goren.ml .