[bug-fix] Use correct agent_ids for demo loader

[ervteng]

Proposed change(s)

When the agent_id changes, the demo loader tries to use the old agent_id to access the agent in the BatchedAgentStep. This caused a crash in 0.14.0 onwards. We now just always use the first agent so that changes in agent_id don't cause a crash.

Note that this fix will have to be coupled with a fix on the C# DemonstrationRecorder to properly log the done signals in the demo. After the C# fix happens, we will add new test demos so that we can catch similar issues in the future.

Useful links (Github issues, JIRA tickets, ML-Agents forum threads etc.)

Github issue: #3455
Unity Forum: https://forum.unity.com/threads/imitation-learning-error-message.829188/

Types of change(s)

• Bug fix
• New feature
• Code refactor
• Breaking change
• Documentation update
• Other (please describe)

Checklist

• I have added tests that prove my fix is effective or that my feature works - will re-record demos after C# fix is implemented
• I have added updated the changelog (if applicable)
• I have added necessary documentation (if applicable)
• I have updated the migration guide (if applicable)