ggml-org/llama.cpp

来源：https://github.com/ggml-org/llama.cpp

llama.cpp

llama 许可证：MIT (https://opensource.org/licenses/MIT)
发布版本 (https://github.com/ggml-org/llama.cpp/releases)
服务器 (https://github.com/ggml-org/llama.cpp/actions/workflows/server.yml)
宣言 (https://github.com/ggml-org/llama.cpp/discussions/205) / ggml (https://github.com/ggml-org/ggml) / 操作 (https://github.com/ggml-org/llama.cpp/blob/master/docs/ops.md)

用 C/C++ 实现 LLM 推理

近期 API 变更

• libllama API 变更日志 (https://github.com/ggml-org/llama.cpp/issues/9289)
• llama-server REST API 变更日志 (https://github.com/ggml-org/llama.cpp/issues/9291)

热门话题

• Hugging Face 缓存迁移：使用 -hf 下载的模型现在存储在标准的 Hugging Face 缓存目录中，从而实现与其他 HF 工具的共享。
• 指南：使用 llama.cpp 的新 WebUI (https://github.com/ggml-org/llama.cpp/discussions/16938)
• 指南：使用 llama.cpp 运行 gpt-oss (https://github.com/ggml-org/llama.cpp/discussions/15396)
• [反馈] 改进 llama.cpp 的打包以支持下游消费者 🤗
• 已添加对原生 MXFP4 格式的 gpt-oss 模型的支持 | PR (https://github.com/ggml-org/llama.cpp/pull/15091) | 与 NVIDIA 的合作 (https://blogs.nvidia.com/blog/rtx-ai-garage-openai-oss) | 评论 (https://github.com/ggml-org/llama.cpp/discussions/15095)
• llama-server 已支持多模态：#12898 (https://github.com/ggml-org/llama.cpp/pull/12898) | 文档
• 用于 FIM 补全的 VS Code 扩展：https://github.com/ggml-org/llama.vscode
• 用于 FIM 补全的 Vim/Neovim 插件：https://github.com/ggml-org/llama.vim
• Hugging Face Inference Endpoints 现已原生支持 GGUF！https://github.com/ggml-org/llama.cpp/discussions/9669
• Hugging Face GGUF 编辑器：讨论 (https://github.com/ggml-org/llama.cpp/discussions/9268) | 工具 (https://huggingface.co/spaces/CISCai/gguf-editor)

快速入门

开始使用 llama.cpp 非常简单。以下是在您的机器上安装它的几种方法：

• 使用 brew, nix 或 winget 安装 llama.cpp
• 使用 Docker 运行——请参阅我们的 Docker 文档
• 从发布页面 (https://github.com/ggml-org/llama.cpp/releases) 下载预编译的二进制文件
• 通过克隆此仓库从源码构建——请查阅 我们的构建指南

安装后，您需要一个模型来使用。转到“获取和量化模型”部分了解更多信息。

示例命令：

# 使用本地模型文件  
llama-cli -m my_model.gguf  

# 或者直接从 Hugging Face 下载并运行模型  
llama-cli -hf ggml-org/gemma-3-1b-it-GGUF  

# 启动兼容 OpenAI 的 API 服务器  
llama-server -hf ggml-org/gemma-3-1b-it-GGUF  

描述

llama.cpp 的主要目标是以最少的设置和最先进的性能在广泛的硬件上（本地和云端）实现 LLM 推理。

• 纯 C/C++ 实现，无需任何依赖
• Apple silicon 是一等公民——通过 ARM NEON、Accelerate 和 Metal 框架优化
• AVX、AVX2、AVX512 和 AMX 支持 x86 架构
• RVV、ZVFH、ZFH、ZICBOP 和 ZIHINTPAUSE 支持 RISC-V 架构
• 1.5 位、2 位、3 位、4 位、5 位、6 位和 8 位整数量化，用于更快的推理和减少内存使用
• 自定义 CUDA 内核，用于在 NVIDIA GPU 上运行 LLM（通过 HIP 支持 AMD GPU，通过 MUSA 支持 Moore Threads GPU）
• Vulkan 和 SYCL 后端支持
• CPU+GPU 混合推理，用于部分加速超出总 VRAM 容量的模型

llama.cpp 项目是开发 ggml (https://github.com/ggml-org/ggml) 库新功能的主要试验场。

模型

通常，以下基础模型的微调版本也受支持。添加新模型支持的说明：HOWTO-add-model.md

纯文本模型

• LLaMA 🦙
• LLaMA 2 🦙🦙
• LLaMA 3 🦙🦙🦙
• Mistral 7B (https://huggingface.co/mistralai/Mistral-7B-v0.1)
• Mixtral MoE (https://huggingface.co/models?search=mistral-ai/Mixtral)
• DBRX (https://huggingface.co/databricks/dbrx-instruct)
• Jamba (https://huggingface.co/ai21labs)
• Falcon (https://huggingface.co/models?search=tiiuae/falcon)
• 中文 LLaMA / Alpaca (https://github.com/ymcui/Chinese-LLaMA-Alpaca) 和中文 LLaMA-2 / Alpaca-2 (https://github.com/ymcui/Chinese-LLaMA-Alpaca-2)
• Vigogne（法语）(https://github.com/bofenghuang/vigogne)
• BERT (https://github.com/ggml-org/llama.cpp/pull/5423)
• Koala (https://bair.berkeley.edu/blog/2023/04/03/koala/)
• Baichuan 1 & 2 (https://huggingface.co/models?search=baichuan-inc/Baichuan) + 衍生版本 (https://huggingface.co/hiyouga/baichuan-7b-sft)
• Aquila 1 & 2 (https://huggingface.co/models?search=BAAI/Aquila)
• Starcoder 模型 (https://github.com/ggml-org/llama.cpp/pull/3187)
• Refact (https://huggingface.co/smallcloudai/Refact-1_6B-fim)
• MPT (https://github.com/ggml-org/llama.cpp/pull/3417)
• Bloom (https://github.com/ggml-org/llama.cpp/pull/3553)
• Yi 模型 (https://huggingface.co/models?search=01-ai/Yi)
• StableLM 模型 (https://huggingface.co/stabilityai)
• Deepseek 模型 (https://huggingface.co/models?search=deepseek-ai/deepseek)
• Qwen 模型 (https://huggingface.co/models?search=Qwen/Qwen)
• PLaMo-13B (https://github.com/ggml-org/llama.cpp/pull/3557)
• Phi 模型 (https://huggingface.co/models?search=microsoft/phi)
• PhiMoE (https://github.com/ggml-org/llama.cpp/pull/11003)
• GPT-2 (https://huggingface.co/gpt2)
• Orion 14B (https://github.com/ggml-org/llama.cpp/pull/5118)
• InternLM2 (https://huggingface.co/models?search=internlm2)
• CodeShell (https://github.com/WisdomShell/codeshell)
• Gemma (https://ai.google.dev/gemma)
• Mamba (https://github.com/state-spaces/mamba)
• Grok-1 (https://huggingface.co/keyfan/grok-1-hf)
• Xverse (https://huggingface.co/models?search=xverse)
• Command-R 模型 (https://huggingface.co/models?search=CohereForAI/c4ai-command-r)
• SEA-LION (https://huggingface.co/models?search=sea-lion)
• GritLM-7B (https://huggingface.co/GritLM/GritLM-7B) + GritLM-8x7B (https://huggingface.co/GritLM/GritLM-8x7B)
• OLMo (https://allenai.org/olmo)
• OLMo 2 (https://allenai.org/olmo)
• OLMoE (https://huggingface.co/allenai/OLMoE-1B-7B-0924)
• Granite 模型 (https://huggingface.co/collections/ibm-granite/granite-code-models-6624c5cec322e4c148c8b330)
• GPT-NeoX (https://github.com/EleutherAI/gpt-neox) + Pythia (https://github.com/EleutherAI/pythia)
• Snowflake-Arctic MoE (https://huggingface.co/collections/Snowflake/arctic-66290090abe542894a5ac520)
• Smaug (https://huggingface.co/models?search=Smaug)
• Poro 34B (https://huggingface.co/LumiOpen/Poro-34B)
• Bitnet b1.58 模型 (https://huggingface.co/1bitLLM)
• Flan T5 (https://huggingface.co/models?search=flan-t5)
• Open Elm 模型 (https://huggingface.co/collections/apple/openelm-instruct-models-6619ad295d7ae9f868b759ca)
• ChatGLM3-6b (https://huggingface.co/THUDM/chatglm3-6b) + ChatGLM4-9b (https://huggingface.co/THUDM/glm-4-9b) + GLMEdge-1.5b (https://huggingface.co/THUDM/glm-edge-1.5b-chat) + GLMEdge-4b (https://huggingface.co/THUDM/glm-edge-4b-chat)
• GLM-4-0414 (https://huggingface.co/collections/THUDM/glm-4-0414-67f3cbcb34dd9d252707cb2e)
• SmolLM (https://huggingface.co/collections/HuggingFaceTB/smollm-6695016cad7167254ce15966)
• EXAONE-3.0-7.8B-Instruct (https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct)
• FalconMamba 模型 (https://huggingface.co/collections/tiiuae/falconmamba-7b-66b9a580324dd1598b0f6d4a)
• Jais (https://huggingface.co/inceptionai/jais-13b-chat)
• Bielik-11B-v2.3 (https://huggingface.co/collections/speakleash/bielik-11b-v23-66ee813238d9b526a072408a)
• RWKV-7 (https://huggingface.co/collections/shoumenchougou/rwkv7-gxx-gguf)
• RWKV-6 (https://github.com/BlinkDL/RWKV-LM)
• QRWKV-6 (https://huggingface.co/recursal/QRWKV6-32B-Instruct-Preview-v0.1)
• GigaChat-20B-A3B (https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct)
• Trillion-7B-preview (https://huggingface.co/trillionlabs/Trillion-7B-preview)
• Ling 模型 (https://huggingface.co/collections/inclusionAI/ling-67c51c85b34a7ea0aba94c32)
• LFM2 模型 (https://huggingface.co/collections/LiquidAI/lfm2-686d721927015b2ad73eaa38)
• Hunyuan 模型 (https://huggingface.co/collections/tencent/hunyuan-dense-model-6890632cda26b19119c9c5e7)
• BailingMoeV2 (Ring/Ling 2.0) 模型 (https://huggingface.co/collections/inclusionAI/ling-v2-68bf1dd2fc34c306c1fa6f86)

多模态模型

• LLaVA 1.5 模型 (https://huggingface.co/collections/liuhaotian/llava-15-653aac15d994e992e2677a7e), LLaVA 1.6 模型 (https://huggingface.co/collections/liuhaotian/llava-16-65b9e40155f60fd046a5ccf2)
• BakLLaVA (https://huggingface.co/models?search=SkunkworksAI/Bakllava)
• Obsidian (https://huggingface.co/NousResearch/Obsidian-3B-V0.5)
• ShareGPT4V (https://huggingface.co/models?search=Lin-Chen/ShareGPT4V)
• MobileVLM 1.7B/3B 模型 (https://huggingface.co/models?search=mobileVLM)
• Yi-VL (https://huggingface.co/models?search=Yi-VL)
• Mini CPM (https://huggingface.co/models?search=MiniCPM)
• Moondream (https://huggingface.co/vikhyatk/moondream2)
• Bunny (https://github.com/BAAI-DCAI/Bunny)
• GLM-EDGE (https://huggingface.co/models?search=glm-edge)
• Qwen2-VL (https://huggingface.co/collections/Qwen/qwen2-vl-66cee7455501d7126940800d)
• LFM2-VL (https://huggingface.co/collections/LiquidAI/lfm2-vl-68963bbc84a610f7638d5ffa)

绑定

• Python: ddh0/easy-llama (https://github.com/ddh0/easy-llama)
• Python: abetlen/llama-cpp-python (https://github.com/abetlen/llama-cpp-python)
• Go: go-skynet/go-llama.cpp (https://github.com/go-skynet/go-llama.cpp)
• Node.js: withcatai/node-llama-cpp (https://github.com/withcatai/node-llama-cpp)
• JS/TS（llama.cpp 服务器客户端）：lgrammel/modelfusion (https://modelfusion.dev/integration/model-provider/llamacpp)
• JS/TS（可编程提示引擎 CLI）：offline-ai/cli (https://github.com/offline-ai/cli)
• JavaScript/Wasm（可在浏览器中运行）：tangledgroup/llama-cpp-wasm (https://github.com/tangledgroup/llama-cpp-wasm)
• Typescript/Wasm（更友好的 API，可在 npm 上获取）：ngxson/wllama (https://github.com/ngxson/wllama)
• Ruby: yoshoku/llama_cpp.rb (https://github.com/yoshoku/llama_cpp.rb)
• Ruby: docusealco/rllama (https://github.com/docusealco/rllama)
• Rust（功能更丰富）：edgenai/llama_cpp-rs (https://github.com/edgenai/llama_cpp-rs)
• Rust（更友好的 API）：mdrokz/rust-llama.cpp (https://github.com/mdrokz/rust-llama.cpp)
• Rust（更直接的绑定）：utilityai/llama-cpp-rs (https://github.com/utilityai/llama-cpp-rs)
• Rust（从 crates.io 自动构建）：ShelbyJenkins/llm_client (https://github.com/ShelbyJenkins/llm_client)
• C#/.NET: SciSharp/LLamaSharp (https://github.com/SciSharp/LLamaSharp)
• C#/VB.NET（功能更丰富——社区许可）：LM-Kit.NET (https://docs.lm-kit.com/lm-kit-net/index.html)
• Scala 3: donderom/llm4s (https://github.com/donderom/llm4s)
• Clojure: phronmophobic/llama.clj (https://github.com/phronmophobic/llama.clj)
• React Native: mybigday/llama.rn (https://github.com/mybigday/llama.rn)
• Java: kherud/java-llama.cpp (https://github.com/kherud/java-llama.cpp)
• Java: QuasarByte/llama-cpp-jna (https://github.com/QuasarByte/llama-cpp-jna)
• Zig: deins/llama.cpp.zig (https://github.com/Deins/llama.cpp.zig)
• Flutter/Dart: netdur/llama_cpp_dart (https://github.com/netdur/llama_cpp_dart)
• Flutter: xuegao-tzx/Fllama (https://github.com/xuegao-tzx/Fllama)
• PHP（基于 llama.cpp 构建的 API 绑定和功能）：distantmagic/resonance (https://github.com/distantmagic/resonance)（更多信息）(https://github.com/ggml-org/llama.cpp/pull/6326)
• Guile Scheme: guile_llama_cpp (https://savannah.nongnu.org/projects/guile-llama-cpp)
• Swift srgtuszy/llama-cpp-swift (https://github.com/srgtuszy/llama-cpp-swift)
• Swift ShenghaiWang/SwiftLlama (https://github.com/ShenghaiWang/SwiftLlama)
• Delphi Embarcadero/llama-cpp-delphi (https://github.com/Embarcadero/llama-cpp-delphi)
• Go（无需 CGo）：hybridgroup/yzma (https://github.com/hybridgroup/yzma)
• Android: llama.android

用户界面

（要在此处列出项目，应明确声明依赖于 llama.cpp）

• AI Sublime Text 插件 (https://github.com/yaroslavyaroslav/OpenAI-sublime-text) (MIT)
• BonzAI 应用 (https://apps.apple.com/us/app/bonzai-your-local-ai-agent/id6752847988) (专有)
• cztomsik/ava (https://github.com/cztomsik/ava) (MIT)
• Dot (https://github.com/alexpinel/Dot) (GPL)
• eva (https://github.com/ylsdamxssjxxdd/eva) (MIT)
• iohub/collama (https://github.com/iohub/coLLaMA) (Apache-2.0)
• janhq/jan (https://github.com/janhq/jan) (AGPL)
• johnbean393/Sidekick (https://github.com/johnbean393/Sidekick) (MIT)
• KanTV (https://github.com/zhouwg/kantv?tab=readme-ov-file) (Apache-2.0)
• KodiBot (https://github.com/firatkiral/kodibot) (GPL)
• llama.vim (https://github.com/ggml-org/llama.vim) (MIT)
• LARS (https://github.com/abgulati/LARS) (AGPL)
• Llama Assistant (https://github.com/vietanhdev/llama-assistant) (GPL)
• LlamaLib (https://github.com/undreamai/LlamaLib) (Apache-2.0)
• LLMFarm (https://github.com/guinmoon/LLMFarm?tab=readme-ov-file) (MIT)
• LLMUnity (https://github.com/undreamai/LLMUnity) (MIT)
• LMStudio (https://lmstudio.ai/) (专有)
• LocalAI (https://github.com/mudler/LocalAI) (MIT)
• LostRuins/koboldcpp (https://github.com/LostRuins/koboldcpp) (AGPL)
• MindMac (https://mindmac.app) (专有)
• MindWorkAI/AI-Studio (https://github.com/MindWorkAI/AI-Studio) (FSL-1.1-MIT)
• Mobile-Artificial-Intelligence/maid (https://github.com/Mobile-Artificial-Intelligence/maid) (MIT)
• Mozilla-Ocho/llamafile (https://github.com/Mozilla-Ocho/llamafile) (Apache-2.0)
• nat/openplayground (https://github.com/nat/openplayground) (MIT)
• nomic-ai/gpt4all (https://github.com/nomic-ai/gpt4all) (MIT)
• ollama/ollama (https://github.com/ollama/ollama) (MIT)
• oobabooga/text-generation-webui (https://github.com/oobabooga/text-generation-webui) (AGPL)
• PocketPal AI (https://github.com/a-ghorbani/pocketpal-ai) (MIT)
• psugihara/FreeChat (https://github.com/psugihara/FreeChat) (MIT)
• ptsochantaris/emeltal (https://github.com/ptsochantaris/emeltal) (MIT)
• pythops/tenere (https://github.com/pythops/tenere) (AGPL)
• ramalama (https://github.com/containers/ramalama) (MIT)
• semperai/amica (https://github.com/semperai/amica) (MIT)
• withcatai/catai (https://github.com/withcatai/catai) (MIT)
• Autopen (https://github.com/blackhole89/autopen) (GPL)

工具

• akx/ggify (https://github.com/akx/ggify) – 从 Hugging Face Hub 下载 PyTorch 模型并将其转换为 GGML
• akx/ollama-dl (https://github.com/akx/ollama-dl) – 从 Ollama 库下载模型，直接用于 llama.cpp
• crashr/gppm (https://github.com/crashr/gppm) – 启动利用 NVIDIA Tesla P40 或 P100 GPU 的 llama.cpp 实例，并降低空闲功耗
• gpustack/gguf-parser (https://github.com/gpustack/gguf-parser-go/tree/main/cmd/gguf-parser) – 检查/查看 GGUF 文件并估算内存使用量
• Styled Lines (https://marketplace.unity.com/packages/tools/generative-ai/styled-lines-llama-cpp-model-292902)（专有许可，用于 Unity3d 游戏开发的推理部分异步包装器，附带预构建的移动和 Web 平台包装器及模型示例）
• unslothai/unsloth (https://github.com/unslothai/unsloth) – 🦥 将微调和训练的模型导出/保存为 GGUF (Apache-2.0)

基础设施

• Paddler (https://github.com/intentee/paddler) – 开源 LLMOps 平台，用于在您的环境中托管和扩展 AI