bestofgithub.dev

← Back to the directory

synchopate/llama-modchip

View on GitHub

Run 284B DeepSeek-V4-Flash (<59GiB RAM: 41.1 cold pp, 9.1 N=1 tg, up to 25.3 spec tok/s) & 30B MoEs (22.4 N=1 / 111.6 spec tok/s) on 64GB CPU (0 GPUs).

synchopate/llama-modchip is #10,057 on the bestofgithub Elo ladder, in Copper V, with a rating of 885 (+258 this week). It has 1,267 stars.

Copper V

885 Elo

  • #10,057 on the ladder
  • +18 today · +258 this week
  • Peak 885
66 Elo to Copper IV

Stars

1,267

  • +32 today
  • – this week
  • – this month

Repository

Forks
112
Language
Python
Last push
Category
AI & Agents

Achievements

bestofgithub badgeMaintainer? Show your league in your README

Card: your league's emblem, from Copper to Legendary

[![bestofgithub](https://bestofgithub.dev/badge/synchopate/llama-modchip.svg?style=card)](https://bestofgithub.dev/repo/synchopate/llama-modchip)

Flat: fits beside your other badges

[![bestofgithub](https://bestofgithub.dev/badge/synchopate/llama-modchip.svg)](https://bestofgithub.dev/repo/synchopate/llama-modchip)

Install

git
git clone https://github.com/synchopate/llama-modchip.git

Similar repositories

wici-ai/inferssd

Open on GitHubC++

Run models larger than your RAM or VRAM by streaming MoE experts from SSD. Built on llama.cpp.

  • 838+65 Elo, change today
  • 1.1k stars
  • 99 forks
git
git clone https://github.com/wici-ai/inferssd.git
  • apple-silicon
  • cuda
  • gguf
  • inference

README

Loading README…