Closed Godley closed 5 years ago
Don't take the fruit thing too seriously - we'll probably use road signs instead.
Here is a working example: https://github.com/dotmesh-io/dotscience-roadsigns/blob/master/roadsigns.ipynb
It exports a SavedModel
with structure:
`-- 1
|-- saved_model.pb
`-- variables
|-- variables.data-00000-of-00001
`-- variables.index
2 directories, 3 files
This should be able to be deployed using Tensorflow serving, according to docs here: https://www.tensorflow.org/tfx/tutorials/serving/rest_simple
Done sufficiently for the demo at GTC which is in the past now!
As a user, I want to be able to see a "real-world" use to a model so that I can see the value of deploying it automatically.
Based on the notes in https://docs.google.com/document/d/1pFzOb3EBbK4vOfjFqnALAzKTEcZY8lS4dK91BHxhC0c/edit# this should be a fruit detector. It should live in an account on dotscience production in its own project