Support speculative decoding with llama.cpp

[kerthcet]

What would you like to be added:

See comment here: ggerganov/llama.cpp#5877 (comment), llama.cpp use another command llama-speculative to process the technology.

Why is this needed:

Completion requirements:

This enhancement requires the following artifacts:

• Design doc
• API change
• Docs update

The artifacts should be linked in subsequent comments.