goliaro/ms_thesis

ExpertFlow: Enabling Low-Latency Asynchronous Inference for Mixture of Expert Models

This repository contains the LaTeX source and PDF of my Tsinghua MS thesis, submitted in May 2023.

You can download the PDF here.

For more information, check out the project page here.

The code is available here.

Citation

Please cite as:

@masterthesis{oliaro2023expertflow,
    title        = {ExpertFlow: Enabling Low-Latency Asynchronous Inference for Mixture of Expert Models},
    author       = {Gabriele Oliaro},
    year         = 2023,
    month        = {May},
    address      = {Beijing, China},
    school       = {Tsinghua University},
    type         = {Master's thesis}
}