Enable AIPCC builds of GuideLLM

[dbutenhof]

Before we can integrate GuideLLM with downstream RHAI, we need to

• Understand the AIPCC onboarding process
• Make sure all dependencies are built (or request they be added)
• Set up the AIPCC build of GuideLLM artifacts
• Determine how to integrate those artifacts into our build/release process

This is also a dependency for Eval Hub support (#653)

[dbutenhof]

Interestingly, AIPCC has already "onboarded" GuideLLM and is building a guidellm wheel (I believe only with [recommended] extras), including v0.6.0, under rocm, cuda, and cpu variations. The wheels seem to be publicly accessible (outside the firewall) although without being certain of the proprieties I won't post a URL on GitHub.

The next step is figuring out what's involved in getting a container built in the AIPCC pipeline.