Copper V
838 Elo
- #10,083 on the ladder
- +65 today · +72 this week
- Peak 838
Run models larger than your RAM or VRAM by streaming MoE experts from SSD. Built on llama.cpp.
wici-ai/inferssd is #10,083 on the bestofgithub Elo ladder, in Copper V, with a rating of 838 (+72 this week). It has 1,179 stars.
838 Elo
1,179
git clone https://github.com/wici-ai/inferssd.gitRun 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).
git clone https://github.com/synchopate/llama-modchip.gitDistribute and run LLMs with a single file.
git clone https://github.com/mozilla-ai/llamafile.gitFast key-value DB in Go.
git clone https://github.com/dgraph-io/badger.gitGemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook
git clone https://github.com/drumih/turbo-fieldfare.gitHundreds of models & providers. One command to find what runs on your hardware.
brew install llmfitFastest and only macOS Dictation app with on-device STT and custom trained AI enhancement model. Windows, iOS testflight! Android, Linux waitlist open! A local Wispr Flow alternative. Follow https://x.com/fluidvoiceapp for more
brew install --cask fluidvoiceLoading README…