No reward showing on tensorboard when decision interval is greater than 1

[guillefix]

Describe the bug
I have two behaviours in the environemnt, and whenever the decision interval is set to be >1 for either of them, the reward of that one doesn't show on Tensorboard (and indeed it isn't present on the summaries folder). But when the decision interval is set to 1, it does show.

Environment (please complete the following information):

• OS + version: Ubuntu 18.10
• ML-Agents version: develop branch at commit ebad90f. EDIT: I've tested on latest dev branch, and it seems to have the same issue
• TensorFlow version: 1.14.0

[surfnerd]

Hi @guillefix,
Thanks for the report. This, indeed, sounds like a bug. I've logged this in our internal tracker as MLA-490. Any updates to the issue will be posted here.

[ervteng]

Hi @guillefix, this bug should have been fixed by PR #3181 and it should work now in both master and the new 0.13.0 release.

FYI the develop branch has been deprecated; master contains the latest changes. You can use tags to get the official versioned releases.

[github-actions]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.