bitnet 开源项目推荐（162 个）

microsoft /unilm

跨任务、语言和模态的大规模自监督预训练

ustcwhy /BitVLA

Official implementation for BitVLA: 1-bit Vision-Language-Action Models for Robotics Manipulation (CoRL 2026, Spotlight)

TilelliLab /atome-lm

A ternary, zero-heap tiny language model that runs inside a $2 microcontroller — bit-exact Python <-> C99 <-> Cortex-M3 (QEMU) parity. Apache-2.0.

ahtx /kindi

Kindi 0.4: local LLM inference on low-cost NVIDIA GPUs, Radeon 780M and CPUs. Bonsai 2B-27B, AH-Nanbeige4.3-33B and experimental Nemotron, with an OpenAI-compatible server.

grctest /Electron-BitNet

Running Microsoft's BitNet via Electron, React & Astro

shifulegend /project-zero

CPU-optimized LLM inference engine (C)

stackblogger /BitNet-Stack

BitNet-Stack - Run a small BitNet model on local machine with one Docker command, and chat in browser.

tomsanbear /bitnet-rs

Implementing the BitNet model in Rust

grctest /FastAPI-BitNet

Running Microsoft's BitNet inference framework via FastAPI, Uvicorn and Docker.

kevbuh /bitnet

pure pytorch implementation of Microsoft's BitNet b1.58 2B4T

Quitetall /tritium

Foundational ternary-model inference & training library — CUDA + CPU, BitNet/TQ ternary kernels

moshix /BITNETServices

All the important tools for BITNET/HNET programs like relay chat, xyzzy, YWAKEUP, IUCVTRAP, CLRSCRN

Zhayr1 /BitMamba-2

Official implementation of BitMamba-2. A scalable 1.58-bit State Space Model (Mamba-2 + BitNet) trained from scratch on 150B tokens. Includes JAX training code and high-performance C++ inference engine.

m96-chan /0xBitNet

Run BitNet b1.58 ternary LLMs with WebGPU — in browsers and native apps

pcdeni /CaSA

Software side of an LLM running inside a DRAM chip via charge-sharing PIM (BitNet b1.58 on DRAM-Bender silicon).

Zhayr1 /bitmamba.cpp

Ultra-lightweight C++ inference engine for BitMamba-2 (1.58-bit SSM). Runs 1B models on consumer CPUs at 50+ tok/s using <700MB RAM. No heavy dependencies.

kamkyu94 /BitNet

BitNet: Learning-Based-Bit-Depth-Expansion

jserv /bitmamba.c

Portable C inference engine for BitMamba-2 models

RightNow-AI /bonsai-turbo

Single-launch batch-1 decode engine for PrismML Bonsai 27B (ternary and 1-bit) on NVIDIA GPUs. 1.76x the vendor llama.cpp fork on H100, same outputs.

eugenehp /bitnet-cpp-rs

Rust bindings for bitnet.cpp based on llama-cpp-4