How I Trained a 193K-Pair MoE Model on a 4-Core Laptop Without a GPU (Introducing QNME-Omega)
中文摘要
作者展示如何仅用4核笔记本且无GPU，训练出193K参数量的MoE模型QNME-Omega。
English Summary
Susanta Banik demonstrates training the 193K-pair MoE model QNME-Omega on a 4-core laptop without a GPU.
原文节选
By Susanta Banik Continue reading on Medium »