Tools / repository

BitNet

bitnet.cpp is Microsoft's inference framework for 1-bit LLMs such as BitNet b1.58, with kernels for ternary-weight models on CPUs and GPUs. It supports BitNet language and embedding models.

URL: https://github.com/microsoft/bitnet
Collection: AI & Agents
License: MIT
Open source: Yes

1-bit-llm inference cpu-inference gpu quantization embeddings microsoft edge-deployment ai-agents agent-tools