Deployment of agent container is always attempted

[bgoelTT]

Describe the bug
In app/api/docker_control/views.py we also attempt to deploy the agent container for every model. This does not make sense for non-LLM models, causing the API call to "fail" but the actual model deployment is successful.

To Reproduce
Steps to reproduce the behavior:

1. Go to 'deploy YOLOv4 inference server'
2. Deploy button will fail and turn red, but inference server was launched successfully

Expected behavior
We should launch the agent container depending on the model type. This is an exact scenario where implementing #167 will fix all of our problems. I think the solution from #167 should be when deciding whether to deploy the agent container.

[bgoelTT]

@mvanniasingheTT which models are compatible for using with the agent container? I assume its all the LLMs?

[mvanniasingheTT]

@mvanniasingheTT which models are compatible for using with the agent container? I assume its all the LLMs?

All the llama models except the small ones like 3B should work.

[bgoelTT]

I don't mean work in terms of accuracy. I mean do they raise errors, if any? I expect the small LLMs like 3.1-1B to still deploy properly but are kind of buggy.

[mvanniasingheTT]

Yes, that's correct. All the LLMs, even the small ones, still deploy properly.