* QWEN-2 models added to resources to download
    * LLM vision implementation files added
    * LLM encode function updated to use EncodePayload instead of string
    * Include supported input/output modalities in LLM

Signed-off-by: Eanna O Cathain <eanna.ocathain@arm.com>
Change-Id: I9859cc7971fe75ad20a961f6eb15fe0aed57d0e1