[FEATURE] Add setting to control enable/disable local model

[ylwu-amzn]

Is your feature request related to a problem?
We have a feature setting to enable/disable remote inference plugins.ml_commons.remote_inference.enabled but we don't have dedicated setting for local model

What solution would you like?
We can add a new setting for enable/disable local model.
Once it disabled, user can't upload/deploy/predict model

What alternatives have you considered?
N/A

Do you have any additional context?
Add any other context or screenshots about the feature request here.