janas
37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
Open the source (opens github.com) https://github.com/prabanta-dev/janas
What it is
AI repositories from prabanta-dev.
The problem it solves and who it suits haven't been reviewed by InnerVault. The description above is the source's own.
How to get it
Get the code
git clone https://github.com/prabanta-dev/janas.git
Standard git access to a public repository
Setup steps: not provided by the source. Follow the instructions on the source.
Related
-
kimi-k3-in-c
AI repositoriesA 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
git clone https://github.com/FareedKhan-dev/kimi-k3-in-c.git -
jev
AI repositoriesjevos: an open-source alternative to Jev and Microsoft-Decision-1. Yes/no decisions, choices and scores with probabilities, on your laptop's CPU.
git clone https://github.com/feder-cr/jev.git -
llmfit
AI repositoriesHundreds of models & providers. One command to find what runs on your hardware.
git clone https://github.com/AlexsJones/llmfit.git -
handy-ollama
AI repositories动手学Ollama,CPU玩转大模型部署,在线阅读地址:https://datawhalechina.github.io/handy-ollama/
git clone https://github.com/datawhalechina/handy-ollama.git -
mojo-dllm
AI repositoriesLocal inference for diffusion language models (LLaDA, Dream), written in Mojo
git clone https://github.com/vladimirrott/mojo-dllm.git -
Soup
AI repositoriesFine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.
git clone https://github.com/MakazhanAlpamys/Soup.git