In this tutorial, we will build a real-time machine learning web app using Python and MLDB.ai. We will use MLDB's Tensorflow integration to use a deep learning model to embed images in a high dimensional conceptual space, and use that representation as features to do transfer learning. This will allow us to build a real-time image classification endpoint.
François Maillet - @mailletf - http://blog.francoismaillet.com
November 13, 15h10
This repo contains the built web app.
For those who attended the talk, the glitch that caused the built plugin to fail to load is now fixed. The code below will work properly.
- The slides are available in the
talk
folder of the repository. - The screencast of the talk is available on Youtube.
from pymldb import Connection
mldb = Connection()
mldb.put("/v1/plugins/pycon", {
"type": "python",
"params": {
"address": "git://github.com/mldbai/pyconca-2016.git"
}
})
You can then browse to https://<host:port>/v1/plugins/pycon/routes/static/index.html
to access the UI.