Implements deterministic evaluation for skrl's multi-agent algorithms (#1972)

## Description

Implement deterministic evaluation for skrl's multi-agent algorithms in
`play.py` script (fix
#1935).
The current application only takes into account single-agent algorithms.

## Type of change

- Bug fix (non-breaking change which fixes an issue)

## Checklist

- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./isaaclab.sh --format`
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] I have updated the changelog and the corresponding version in the
extension's `config/extension.toml` file
- [x] I have added my name to the `CONTRIBUTORS.md` or my name already
exists there

Author: Toni-SM

scripts/reinforcement_learning/skrl/play.py

@@ -179,7 +179,12 @@ def main():
         with torch.inference_mode():
             # agent stepping
             outputs = runner.agent.act(obs, timestep=0, timesteps=0)
-            actions = outputs[-1].get("mean_actions", outputs[0])
+            # - multi-agent (deterministic) actions
+            if hasattr(env, "possible_agents"):
+                actions = {a: outputs[-1][a].get("mean_actions", outputs[0][a]) for a in env.possible_agents}
+            # - single-agent (deterministic) actions
+            else:
+                actions = outputs[-1].get("mean_actions", outputs[0])
             # env stepping
             obs, _, _, _, _ = env.step(actions)
         if args_cli.video: