85 GB DeepSeek-V4-Flash at ~3 tok/s on a 12 GB RTX 3060 + 64 GB DDR5 RAM - Overspill for FreeToken, inspired by Colibri

I've been experimenting with ways to run MoE models that don't fit comfortably in RAM, and I ended up making Overspill, a disk tier for FreeToken. The basic idea came from looking at how Colibri handles experts across disk/RAM/VRAM so I took inspiration from the general approach. Repo: https://githu

中文全文翻译正在补齐，可先查看下方英文原文。

I've been experimenting with ways to run MoE models that don't fit comfortably in RAM, and I ended up making Overspill, a disk tier for FreeToken. The basic idea came from looking at how Colibri handles experts across disk/RAM/VRAM so I took inspiration from the general approach. Repo: https://githu