Min total reward logging to tensorboard is consistently 0

[PaoloGinefra]

Description

I am developing a PPO agent in isaac lab. The robot is meant to follow some waypoints and it gets a +0.1 reward every time it gets to one. I have placed a waypoint at the robot spawn position and, by printing, I know for a fact that the min total reward for each episode is 0.1. On tensorboard this value is constantly 0 tho. I tested with training with just one environment and it correctly shows 0.1 but as soon as I move to 2 envs it gets back to 0. I suspect there is something wrong with the resetting of the buffer for the total reward but I could be wrong

What skrl version are you using?

1.4.3

What ML framework/library version are you using?

IsaacLab

Additional system information

No response

[Toni-SM]

Hi @PaoloGinefra

Mmmm, could you share a link to your task or a way to repro it?