add Lora config to arg list in Neo sharding script& its integ test change

[HappyAmazonian]

Description

• pass Lora config to arg list in Neo sharding script
• also add tinyllama fast model loading for related integ code

[siddvenk]

if we're changing to the string "False", we don't need to wrap this in a str() conversion.

Let's use the following:

enforce_eager = self.properties.get("option.enforce_eager", "true").lower() == "true"

I believe the default value should be true for this to keep in line with our existing behavior (i should have caught that in the original PR)

[siddvenk]

I think we should only be passing lora configs to the engine if the user specifies them. I'd prefer we rewrite this section as

lora_kwargs = {}
if self.properties.get("option.enable_lora"):
    lora_kwargs["enable_lora"] = self.properties.get("option.enable_lora").lower() == "true"
if self.properties.get("option.fully_sharded_loras"):
    lora_kwargs["fully_shareded_loras"] = option.get("option.fully_sharded_loras").lower() == "true"

engine_args = VllmEngineArgs(
    <prior configs>
    **lora_kwargs
)
...

and so on. This way we're only providing the lora configs to the engine if they have been specified