model-releases

Released pre-trained models for use with RPA Tomorrow

Installing with pip

Install the models with pip, with the correct tag/release name

pip install https://github.com/rpa-tomorrow/model-releases/releases/download/<TAG>.tar.gz

Loading models

Once installed, you can load the models in python

import spacy

nlp = spacy.load(<TAG>)
doc = nlp("send an email to Robert")

Manual install

Clone the repo or download the release. Then you can manually load the model by referring to the directory

import spacy

nlp = spacy.load("/path/to/model")

(Maintainers) Creating a release

This is a guide for those who want to push new model releases to this repository.

1. After training and saving the model, check that the meta.json is correct otherwise update it with the correct information.
2. Make a PR with the model's metadata meta.json and place it in this repository inside meta/
3. To create a release package run

python -m spacy package /path/to/your/trained/model /path/to/store/release

4. Build the model package by running the following inside /path/to/store/release

python setup.py sdist

5. The created tarball can be used to create a release here on Github