feat: `cortex run model(:gguf)` part 2

[gabrielle-ong]

Problem

• Followup to feat: cortex run model(:gguf) #1076
• currently, cortex run/chat model_id:option requires stating option eg :gguf in every command
• [to be discussed] For better UX, users can run cortex run model_id without the additional option

Solution brainstorm (WIP)

1. We will likely need a main branch on the Hugging Face Cortex Model repo to support this use case
2. In the long term, this will likely be replaced by something that automatically detects users' hardware and picks the correct engine.

Success Criteria

• cortex run model_id

Tasklist

• [ ]

[gabrielle-ong]

Closing, this is solved by #1521 where users can run models with a substring (regex search) and it returns local models
Also will be linked to hardware compatibility API (recommend models based on hardware specs)