AutoML Benchmark Wave App

Dashboard to compare AutoML frameworks by exploring the results from runs of the OpenML AutoML Benchmark.

App Goal: View & compare performance of different AutoML frameworks, as benchmarked on the OpenML AutoML Benchmark

Target Audience: Data scientists

Industry: Any

GitHub: https://github.com/h2oai/wave-amlb

Actively Being Maintained: Yes

Comes with Demo Mode (pre-loaded data, models, results, etc.): Yes

Allows for Uploading and Using New Data: Yes

System Requirements

Python 3.6+
pip3

Installation

1. Run the Wave Server

Follow the instructions here to download and run the latest Wave Server, a requirement for apps.

Setting up Wave for the first time

Go to download the wave SDK for your platform
- Make sure to scroll down on the page to Assets. For MacOs select "darwin".

wave-0.12.0-darwin-amd64.tar.gz

Extract your download:
- Make sure to run this command in the directory you downloaded the .tar.gz

tar -xzf wave-0.12.0-darwin-amd64.tar.gz

Move it to a convenient location:

mv wave-0.12.0-darwin-amd64 <pathname>/wave

check your $home/wave directory, the output should look like the following below:

.
├── demo
├── examples
├── readme.txt
├── test
├── waved
└── www

Go to your wave directory:

cd wave

Start the wave server by running:

./waved

When runninng the command, you may get an issue about "cannot open file because identity of developer can't be verified: - go to system preferences - security & privacy - general - you'll see a message about wave. select 'allow anyway' - attempt to run ./waved again

If step 6 worked you should be able to go to the following link:
- it's a spinning circle waiting for contact
- To run any Wave app, you need the Wave server up and running at all times.

2. Setup Your Python Environment

$ git clone https://github.com/h2oai/wave-amlb
$ cd wave-amlb
$ make setup
$ source venv/bin/activate

3. Run the App

After you set-up the python environment you can start the app by running

Note! this is in a separate terminal tab then where you ran './waved' command to start the wave server:

wave run src.app

Note! If you did not activate your virtual environment this will be:

./venv/bin/wave run src.app

4. View the App

Point your web browser to localhost:10101

h2oai / wave-amlb