A complete end-to-end demonstration in which we collect training data in Unity and use that data to train a deep neural network to predict the pose of a cube. This model is then deployed in a simulated robotic pick-and-place task.
We are going to stop allowing people to target a Bokken image without specifying a tag.
Only a handful of people are using it, and it gets in the way of using sourcegraph insights to estimate usage.
Currently, no tag equals using the stable version of the image.
I used a pinned version instead of stable because that tag is also going to disappear and I want to prevent you from having to do another maintenance in a few months.
Useful links (GitHub issues, JIRA tickets, forum threads, etc.)
Proposed change(s)
We are going to stop allowing people to target a Bokken image without specifying a tag. Only a handful of people are using it, and it gets in the way of using sourcegraph insights to estimate usage. Currently, no tag equals using the stable version of the image. I used a pinned version instead of stable because that tag is also going to disappear and I want to prevent you from having to do another maintenance in a few months.
Useful links (GitHub issues, JIRA tickets, forum threads, etc.)
JIRA Ticket https://jira.unity3d.com/browse/DSBKN-1034 Image version in the Bokken image catalogue http://images.bokken.cloud:8000/#/image/package-ci/mac/673349491555729414
Types of change(s)
Testing and Verification
Run affected CI pipeline on Yamato.
Test Configuration: Not applicable (CI environment)
Checklist
Other comments