Feature request : Support for PHI3 mini

[raymond-infinitecode]

Prerequisites

Before submitting your issue, please ensure the following:

• I am running the latest version of PowerInfer. Development is rapid, and as of now, there are no tagged versions.
• I have carefully read and followed the instructions in 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).

Feature Description

PHI3 mini is currently the most powerful SLM yet, but can we relu it to make it fast so a single Xeon server can serve hundreds of concurrent users with relu implementation ?

Motivation

Please provide a detailed written description of reasons why this feature is necessary and how it is useful to PowerInfer users.

Possible Implementation

Convert the Phi3 model to relu model