- New upload dialog. The dialog allows to set the video title and to pick the video language so that Speech to Text can correctly process a non-US English video.
-
Dark Vision now processes audio from videos too. It uses Watson Speech to Text to extract the audio track. Then the audio is processed by Natural Language Understanding to extract concepts, entities, emotion.
-
New visual design for the web interface.
- A new toolchain to deploy Dark Vision in two steps! Head over to the README for details.
-
No more deploy-darkvision.sh. It has been replaced by deploy.js in the root folder. The goal is to support deployment of the actions from Windows too. It was a bit too challenging to maintain shell scripts and Windows commands so instead Dark Vision relies on a node script. Node.js version 6.7.0 minimum is required.
-
The local.env file has changed location too. It is now in the root directory.