Feature Request: Speculative Prefill

[endyjasmi]

Prerequisites

• I am running the latest code. Mention the version if possible as well.
• I carefully followed the README.md.
• I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• I reviewed the Discussions, and have a new and useful enhancement to share.

Feature Description

Basically use to improve TTFT. This is very important especially for Ryzen AI Max+ 395 where the prefill stage is slow and unsuitable to be used in Agentic workflow. Hope it can be implemented with

Motivation

This is very important especially for Ryzen AI Max+ 395 where the prefill stage is slow and unsuitable to be used in Agentic workflow. Hope it can be implemented with GLM 4.7 flash

Possible Implementation

https://github.com/[Jingyu6/speculative_prefill](https://github.com/Jingyu6/speculative_prefill)