slrbl / human-in-the-loop-machine-learning-tool-tornado

Tornado is an open source Human-in-the-loop machine learning tool. It helps you label your dataset on the fly while training your model through a simple web user interface. It supports all data types: structured, text and image.

GNU Affero General Public License v3.0

61 stars 10 forks source link

active-learning active-learning-module activelearning artificial-intelligence automl data-annotation data-annotation-machine-learning data-annotation-tools dataannotations human-in-the-loop-machine-learning interactive-machine-learning keras machine-learning machine-learning-api machinelearning natural-language-processing python ruby-on-rails sklearn

readme

TornadoAi

A zero coding human in the loop (HITL) collaborative data annotation plateform

What is TornadoAi

Tornado is an open source Human-in-the-loop machine learning tool. It helps you label your dataset on the fly while training your model through a simple web user interface. It supports all data types: structured, text and image.

Live demo

In this demo https://www.youtube.com/watch?v=zBe6b_vxs_I the user collaborates with the computer to label a set of 1300 pump impeller images through 4 human in the loop machine learning iterations. The goal of labelling the images is the put in place a classifier to automatize the metal casting quality process within a manufactory.

Data flow

Technology stack

Usage

Build Tornado Docker image

$ docker-compose build

Launch Tornado image

$ docker-compose up

Tornado will then be available at http://localhost:3000

Details and how to use

All information are available at:
https://www.linkedin.com/pulse/tornado-zero-coding-active-learning-ml-tool-walid-daboubi/