Universal Data Tool Update: On-Premise Data Labeling

If you haven’t heard of the Universal Data Tool yet, it’s an open-source web or desktop program to collaborate, build and edit text, image, video, and audio datasets with labels and annotations. 

This is our eighth community update! By releasing these videos, we hope to engage the community and encourage new contributors. 

Full Video:

https://www.youtube.com/watch?v=IBWOaw0jMmM

Join us for a Hacktoberfest Hackathon

We are super excited to announce our 8-hour Hacktoberfest Hackathon on October 29th. You can get all the information as well as register at this link 

https://organize.mlh.io/participants/events/5164-universal-data-tool-livestream-mini-hackathon 

New Feature: Asynchronous Datasets Loading

You can now add millions of samples in a single dataset when using an external server.

Easier On-Premise Usage

The Universal Data Tool along with a collaboration server can now be run as a single docker container. See the on-premise documentation for more details.

You can provide different options to the docker container to control various aspects of the Universal Data Tool. You'll know if a configuration value is used because it will appear in the starting logs as shown below. This allows you to configure things like S3 integrations without needing any user input.

Other Updates

• Fixed reload config state in Audio Transcription: Configuration in Audio Transcription is updating now (thanks @anaplian!)
• Fixed Portuguese translation: Now you can see the special characters in Portuguese translation (thanks, @miguelcarvalho13)
• Introducing Manage Plugins:  You might have noticed the “Manage Plugins” option introduced in version 13 of UTD. It helps to add an ES6 module to add more functionalities. This is a work-in-progress as of now. However, there will be new features in this option soon. 

That’s it for our eighth community update, be sure to follow us on Twitter or join our Slack to hear more!