Adventures with tensorflow.js

GuruAtWork
1 min readApr 8, 2018

--

Deep learning fans will know the tensorflow.js package was recently released. This presentation by Nikhil Thorat and Daniel Smilkov at the recent Tensorflow dev summit provided a glimpse of what the google brain team have been up to.

I’ve borrowed the getting started example from js.tensorflow.org

Simple simple demo : single layer prediction from numeric data.

Steps:

  • load tensorflow.js (creates the tf object)
  • explore the object, list methods.
  • creating the model
  • add layers
  • compile model
  • fit
  • predict.

Yay. We have our first demo.

Next >

— — — — — — — — — — — — — — — — — — -

This is a WIP. adding as I work through to a useful demo.

--

--

GuruAtWork
GuruAtWork

Written by GuruAtWork

Machine Learning, data analysis and Artificial Intelligence developer.

No responses yet