sutantodadang/MLz

MLz is a Zig implementation for running LLaMA language models, primarily utilizing fine-tuned bindings to llama.cpp for high-performance inference. It supports hardware acceleration via CUDA and Vulkan, provides a modern chat interface, and includes an OpenAI-compatible API server.

语言：Assembly · GitHub 星标：3 · Fork：2

话题：inference-engine · llama-cpp · ziglang

在 GitHub 上查看原仓库