Devotd Github

marco-lu2508/fllm

完全用 Rust 编写的本地 LLM 推理引擎 fLLM,为 Q2_K·Q3_K·IQ4_XS 量化格式的共用内核添加了内存预取。

GitHub 项目说明: Local LLM inference engine in pure Rust: GGUF, OpenAI/Anthropic/Ollama-compatible API, CUDA (incl. NVFP4), Metal, Vulkan, CPU

在 GitHub 查看 ↗
星标
★ 0
复刻
0
语言
Rust
许可证
Apache-2.0

文章

1