Machine Learning In Javascript

Materials for the Machine Learning in Javascript Workshop by Expero inc

Given at the M3 Machine Learning Conference, London, Oct 2017

For an overview of the content see the course outline

Getting Setup

The first part of the course will be run in Jupyter notebooks. Jupyter is a REPL like interface embedded in a web page which is good for experimentation and exploratory development.

It's a standard tool in python for data science. Using it means we install a python distribution even though we are working in javascript. If you are learning ML some python is likely in your future anyways so we can maybe live with that.

Installing python and the notebook kernel

Install an IPython distribution

I recomend installing Anaconda 3.6 it is heayweight but that is the distibution I have tested with and if you are interested in moving on apply some of the things you have learnt after the conference in python too, then this is a good starting point.

Install node.js we are using v8.5.x.

If you already have node installed and don't want ot mess with your local environement install nvm and install node v8.5.x from there.
- on Mac nvm
- on Windows nvm-windows

Install the jp-babel variant of the IJavascript kernel
- npm install -g jp-babel
- if you are running on linux check out the instructions on this page
Install Javascript dependencies

run npm install in the repo root folder :)

Start the notebook

run jp-babel-notebook in the repo root folder
a web browser should pop up showing the Jupyter browser / tree view of the current folder.

From Jupyter, open the smoke test notebook

When the notebook opens you should be able to see the jp-babel javascript kernel loaded in the top left corner.
If not, go to Kernel Menu > Change Kernel and select the jp-babel option and follow the prompts.
In the Cell Menu, select Run All. Each cell should execute and if all dependencies are properly installed we shouldn't see any errors.

Setting up to run Keras.js models in the browser

...

Name		Name	Last commit message	Last commit date
Latest commit History 34 Commits
datasets		datasets
images		images
imutils		imutils
keras-js-demos-data @ cf4bad3		keras-js-demos-data @ cf4bad3
presentations		presentations
.babelrc		.babelrc
.gitignore		.gitignore
.gitmodules		.gitmodules
0.0_smoke_test.ipynb		0.0_smoke_test.ipynb
0.1_hello_notebook.ipynb		0.1_hello_notebook.ipynb
0.2_hello_plotly.ipynb		0.2_hello_plotly.ipynb
0.3_hello-datasets.ipynb		0.3_hello-datasets.ipynb
1.0_vectors.ipynb		1.0_vectors.ipynb
1.1_distances_similarity_and_cost.ipynb		1.1_distances_similarity_and_cost.ipynb
1.2_exploring_some_data.ipynb		1.2_exploring_some_data.ipynb
1.3_my_first_classifier.ipynb		1.3_my_first_classifier.ipynb
1.4_regression.ipynb		1.4_regression.ipynb
2.1_unsupervised_kmeans.ipynb		2.1_unsupervised_kmeans.ipynb
2.2_dimensionality.ipynb		2.2_dimensionality.ipynb
2.3_dimensionality_reduction.ipynb		2.3_dimensionality_reduction.ipynb
2.4_hyperparameter_tuning.ipynb		2.4_hyperparameter_tuning.ipynb
2.5_gaussian_mixture_models.ipynb		2.5_gaussian_mixture_models.ipynb
3.0_supervised-knn.ipynb		3.0_supervised-knn.ipynb
3.0_supervised.ipynb		3.0_supervised.ipynb
3.1_accuracy.ipynb		3.1_accuracy.ipynb
3.2_cross_validation.ipynb		3.2_cross_validation.ipynb
4.0_neural_networks.ipynb		4.0_neural_networks.ipynb
4.1_understanding_capacity.ipynb		4.1_understanding_capacity.ipynb
4.2_neural_networks_playground.ipynb		4.2_neural_networks_playground.ipynb
4.3_classifying_wine.ipynb		4.3_classifying_wine.ipynb
5.0_running_forward_models_with_keras.ipynb		5.0_running_forward_models_with_keras.ipynb
OUTLINE.md		OUTLINE.md
README.md		README.md
ml_dev.png		ml_dev.png
ml_prod2.png		ml_prod2.png
ml_production.png		ml_production.png
package-lock.json		package-lock.json
package.json		package.json
simple_ml_cycle.png		simple_ml_cycle.png
utils.js		utils.js
yarn.lock		yarn.lock

experoinc/machine-learning-in-javascript

Folders and files

Latest commit

History

Repository files navigation

Machine Learning In Javascript

Getting Setup

Installing python and the notebook kernel

Setting up to run Keras.js models in the browser

About

Resources

Stars

Watchers

Forks

Languages