mlserve-demo

Demo project for mlserve (https://github.com/ml-libs/mlserve) deployed on heroku.

• https://young-ridge-56019.herokuapp.com

Deployed Project

• Prediction of fake instagram users based on public profile.
• Iris species classification, using on dataset from sklearn.
• Toxic comment classification, bases on kaggle dataset.

Deployement

Execute following commands:

$ pip install git+https://github.com/ml-libs/mlserve.git
$ heroku create
$ git push heroku master
$ heroku logs --tail
$ heroku open

Purge caches

Sometimes it is useful to purge pip caches to force app to use new library:

$ heroku plugins:install heroku-repo
$ heroku repo:purge_cache -a appname
$ git commit --allow-empty -m "Purge cache"
$ git push heroku master

https://help.heroku.com/18PI5RSY/how-do-i-clear-the-build-cache