Llama.cpp

llama-cpp-python 是 llama.cpp 的 Python 绑定。

它支持对许多 LLM 模型进行推理，这些模型可以在 Hugging Face 上获取。

本笔记将介绍如何在 LangChain 中运行 llama-cpp-python。

注意：新版 llama-cpp-python 使用 GGUF 模型文件（详情请见此处）。

这是一个破坏性变更。

要将现有的 GGML 模型转换为 GGUF，您可以在 llama.cpp 中运行以下命令：

python ./convert-llama-ggmlv3-to-gguf.py --eps 1e-5 --input models/openorca-platypus2-13b.ggmlv3.q4_0.bin --output models/openorca-platypus2-13b.gguf.q4_0.bin

安装

安装 llama-cpp 软件包有不同的选项

• CPU 使用
• CPU + GPU（使用多种 BLAS 后端之一）
• Metal GPU (带 Apple Silicon 芯片的 MacOS)

仅 CPU 安装

%pip install --upgrade --quiet  llama-cpp-python

使用 OpenBLAS / cuBLAS / CLBlast 安装

llama.cpp 支持多种 BLAS 后端以实现更快的处理。使用 FORCE_CMAKE=1 环境变量强制使用 cmake 并为所需的 BLAS 后端安装 pip 包（来源）。

使用 cuBLAS 后端的安装示例

!CMAKE_ARGS="-DGGML_CUDA=on" FORCE_CMAKE=1 pip install llama-cpp-python

重要提示：如果您已经安装了仅 CPU 版本的软件包，则需要从头开始重新安装。请考虑以下命令：

!CMAKE_ARGS="-DGGML_CUDA=on" FORCE_CMAKE=1 pip install --upgrade --force-reinstall llama-cpp-python --no-cache-dir

使用 Metal 安装

llama.cpp 支持 Apple Silicon 的一级公民——通过 ARM NEON、Accelerate 和 Metal 框架进行优化。使用 FORCE_CMAKE=1 环境变量强制使用 cmake 并安装支持 Metal 的 pip 包（来源）。

支持 Metal 的安装示例

!CMAKE_ARGS="-DLLAMA_METAL=on" FORCE_CMAKE=1 pip install llama-cpp-python

重要提示：如果您已经安装了仅 CPU 版本的软件包，则需要从头开始重新安装：请考虑以下命令

!CMAKE_ARGS="-DLLAMA_METAL=on" FORCE_CMAKE=1 pip install --upgrade --force-reinstall llama-cpp-python --no-cache-dir

Windows 安装

通过源代码编译安装 llama-cpp-python 库是稳定的。您可以遵循仓库本身的大部分说明，但有一些特定于 Windows 的说明可能会有用。

安装 llama-cpp-python 的要求，

• git
• python
• cmake
• Visual Studio Community（确保您使用以下设置安装）
  • 使用 C++ 进行桌面开发
  • Python 开发
  • 使用 C++ 进行 Linux 嵌入式开发

1. 递归克隆 git 仓库以同时获取 llama.cpp 子模块

git clone --recursive -j8 https://github.com/abetlen/llama-cpp-python.git

2. 打开命令提示符并设置以下环境变量。

set FORCE_CMAKE=1
set CMAKE_ARGS=-DGGML_CUDA=OFF

如果您有 NVIDIA GPU，请确保 DGGML_CUDA 设置为 ON

编译和安装

现在您可以 cd 进入 llama-cpp-python 目录并安装软件包

python -m pip install -e .

重要提示：如果您已经安装了仅 CPU 版本的软件包，则需要从头开始重新安装：请考虑以下命令

!python -m pip install -e . --force-reinstall --no-cache-dir

使用

确保您遵循所有说明来安装所有必要的模型文件。

您不需要 API_TOKEN，因为您将在本地运行 LLM。

了解哪些模型适合在所需机器上使用是值得的。

TheBloke 的 Hugging Face 模型有一个“Provided files”（提供文件）部分，其中显示了运行不同量化大小和方法模型所需的 RAM（例如：Llama2-7B-Chat-GGUF）。

这个GitHub 问题也与为您的机器找到合适的模型相关。

from langchain_community.llms import LlamaCpp
from langchain_core.callbacks import CallbackManager, StreamingStdOutCallbackHandler
from langchain_core.prompts import PromptTemplate

API 参考：LlamaCpp | CallbackManager | StreamingStdOutCallbackHandler | PromptTemplate

考虑使用适合您模型的模板！请查看 Hugging Face 等上的模型页面以获取正确的提示模板。

template = """Question: {question}

Answer: Let's work this out in a step by step way to be sure we have the right answer."""

prompt = PromptTemplate.from_template(template)

# Callbacks support token-wise streaming
callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])

CPU

使用 LLaMA 2 7B 模型的示例

# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="/Users/rlm/Desktop/Code/llama.cpp/models/openorca-platypus2-13b.gguf.q4_0.bin",
    temperature=0.75,
    max_tokens=2000,
    top_p=1,
    callback_manager=callback_manager,
    verbose=True,  # Verbose is required to pass to the callback manager
)

question = """
Question: A rap battle between Stephen Colbert and John Oliver
"""
llm.invoke(question)

Stephen Colbert:
Yo, John, I heard you've been talkin' smack about me on your show.
Let me tell you somethin', pal, I'm the king of late-night TV
My satire is sharp as a razor, it cuts deeper than a knife
While you're just a british bloke tryin' to be funny with your accent and your wit.
John Oliver:
Oh Stephen, don't be ridiculous, you may have the ratings but I got the real talk.
My show is the one that people actually watch and listen to, not just for the laughs but for the facts.
While you're busy talkin' trash, I'm out here bringing the truth to light.
Stephen Colbert:
Truth? Ha! You think your show is about truth? Please, it's all just a joke to you.
You're just a fancy-pants british guy tryin' to be funny with your news and your jokes.
While I'm the one who's really makin' a difference, with my sat
``````output

llama_print_timings:        load time =   358.60 ms
llama_print_timings:      sample time =   172.55 ms /   256 runs   (    0.67 ms per token,  1483.59 tokens per second)
llama_print_timings: prompt eval time =   613.36 ms /    16 tokens (   38.33 ms per token,    26.09 tokens per second)
llama_print_timings:        eval time = 10151.17 ms /   255 runs   (   39.81 ms per token,    25.12 tokens per second)
llama_print_timings:       total time = 11332.41 ms

"\nStephen Colbert:\nYo, John, I heard you've been talkin' smack about me on your show.\nLet me tell you somethin', pal, I'm the king of late-night TV\nMy satire is sharp as a razor, it cuts deeper than a knife\nWhile you're just a british bloke tryin' to be funny with your accent and your wit.\nJohn Oliver:\nOh Stephen, don't be ridiculous, you may have the ratings but I got the real talk.\nMy show is the one that people actually watch and listen to, not just for the laughs but for the facts.\nWhile you're busy talkin' trash, I'm out here bringing the truth to light.\nStephen Colbert:\nTruth? Ha! You think your show is about truth? Please, it's all just a joke to you.\nYou're just a fancy-pants british guy tryin' to be funny with your news and your jokes.\nWhile I'm the one who's really makin' a difference, with my sat"

使用 LLaMA v1 模型的示例

# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="./ggml-model-q4_0.bin", callback_manager=callback_manager, verbose=True
)

llm_chain = prompt | llm

question = "What NFL team won the Super Bowl in the year Justin Bieber was born?"
llm_chain.invoke({"question": question})

1. First, find out when Justin Bieber was born.
2. We know that Justin Bieber was born on March 1, 1994.
3. Next, we need to look up when the Super Bowl was played in that year.
4. The Super Bowl was played on January 28, 1995.
5. Finally, we can use this information to answer the question. The NFL team that won the Super Bowl in the year Justin Bieber was born is the San Francisco 49ers.
``````output

llama_print_timings:        load time =   434.15 ms
llama_print_timings:      sample time =    41.81 ms /   121 runs   (    0.35 ms per token)
llama_print_timings: prompt eval time =  2523.78 ms /    48 tokens (   52.58 ms per token)
llama_print_timings:        eval time = 23971.57 ms /   121 runs   (  198.11 ms per token)
llama_print_timings:       total time = 28945.95 ms

'\n\n1. First, find out when Justin Bieber was born.\n2. We know that Justin Bieber was born on March 1, 1994.\n3. Next, we need to look up when the Super Bowl was played in that year.\n4. The Super Bowl was played on January 28, 1995.\n5. Finally, we can use this information to answer the question. The NFL team that won the Super Bowl in the year Justin Bieber was born is the San Francisco 49ers.'

GPU

如果 BLAS 后端安装正确，您将在模型属性中看到 BLAS = 1 指示。

两个最重要的 GPU 使用参数是

• n_gpu_layers - 决定有多少层模型被卸载到您的 GPU。
• n_batch - 每次并行处理的 token 数量。

正确设置这些参数将显著提高评估速度（更多详情请参阅封装代码）。

n_gpu_layers = -1  # The number of layers to put on the GPU. The rest will be on the CPU. If you don't know how many layers there are, you can use -1 to move all to GPU.
n_batch = 512  # Should be between 1 and n_ctx, consider the amount of VRAM in your GPU.

# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="/Users/rlm/Desktop/Code/llama.cpp/models/openorca-platypus2-13b.gguf.q4_0.bin",
    n_gpu_layers=n_gpu_layers,
    n_batch=n_batch,
    callback_manager=callback_manager,
    verbose=True,  # Verbose is required to pass to the callback manager
)

llm_chain = prompt | llm
question = "What NFL team won the Super Bowl in the year Justin Bieber was born?"
llm_chain.invoke({"question": question})

1. Identify Justin Bieber's birth date: Justin Bieber was born on March 1, 1994.

2. Find the Super Bowl winner of that year: The NFL season of 1993 with the Super Bowl being played in January or of 1994.

3. Determine which team won the game: The Dallas Cowboys faced the Buffalo Bills in Super Bowl XXVII on January 31, 1993 (as the year is mis-labelled due to a error). The Dallas Cowboys won this matchup.

So, Justin Bieber was born when the Dallas Cowboys were the reigning NFL Super Bowl.
``````output

llama_print_timings:        load time =   427.63 ms
llama_print_timings:      sample time =   115.85 ms /   164 runs   (    0.71 ms per token,  1415.67 tokens per second)
llama_print_timings: prompt eval time =   427.53 ms /    45 tokens (    9.50 ms per token,   105.26 tokens per second)
llama_print_timings:        eval time =  4526.53 ms /   163 runs   (   27.77 ms per token,    36.01 tokens per second)
llama_print_timings:       total time =  5293.77 ms

"\n\n1. Identify Justin Bieber's birth date: Justin Bieber was born on March 1, 1994.\n\n2. Find the Super Bowl winner of that year: The NFL season of 1993 with the Super Bowl being played in January or of 1994.\n\n3. Determine which team won the game: The Dallas Cowboys faced the Buffalo Bills in Super Bowl XXVII on January 31, 1993 (as the year is mis-labelled due to a error). The Dallas Cowboys won this matchup.\n\nSo, Justin Bieber was born when the Dallas Cowboys were the reigning NFL Super Bowl."

Metal

如果 Metal 安装正确，您将在模型属性中看到 NEON = 1 指示。

两个最重要的 GPU 参数是

• n_gpu_layers - 决定有多少层模型被卸载到您的 Metal GPU。
• n_batch - 每次并行处理的 token 数量，默认为 8，可以设置为更大的数字。
• f16_kv - 出于某种原因，Metal 只支持 True，否则您会收到类似 Asserting on type 0 GGML_ASSERT: .../ggml-metal.m:706: false && "not implemented" 的错误。

正确设置这些参数将显著提高评估速度（更多详情请参阅封装代码）。

n_gpu_layers = 1  # The number of layers to put on the GPU. The rest will be on the CPU. If you don't know how many layers there are, you can use -1 to move all to GPU.
n_batch = 512  # Should be between 1 and n_ctx, consider the amount of RAM of your Apple Silicon Chip.
# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="/Users/rlm/Desktop/Code/llama.cpp/models/openorca-platypus2-13b.gguf.q4_0.bin",
    n_gpu_layers=n_gpu_layers,
    n_batch=n_batch,
    f16_kv=True,  # MUST set to True, otherwise you will run into problem after a couple of calls
    callback_manager=callback_manager,
    verbose=True,  # Verbose is required to pass to the callback manager
)

控制台日志将显示以下日志以指示 Metal 已正确启用。

ggml_metal_init: allocating
ggml_metal_init: using MPS
...

您还可以通过查看进程的 GPU 使用情况来检查 Activity Monitor，在开启 n_gpu_layers=1 后，CPU 使用率将显著下降。

对于 LLM 的首次调用，由于 Metal GPU 中的模型编译，性能可能会较慢。

文法

我们可以使用文法来约束模型输出并根据其中定义的规则采样 token。

为了演示这一概念，我们包含了示例文法文件，它们将在下面的示例中使用。

创建 gbnf 文法文件可能很耗时，但如果您有一个输出模式很重要的用例，有两种工具可以提供帮助

• 在线文法生成器应用，将 TypeScript 接口定义转换为 gbnf 文件。
• 用于将 json schema 转换为 gbnf 文件的Python 脚本。例如，您可以创建 pydantic 对象，使用 .schema_json() 方法生成其 JSON schema，然后使用此脚本将其转换为 gbnf 文件。

在第一个示例中，提供指定 json.gbnf 文件的路径以生成 JSON

n_gpu_layers = 1  # The number of layers to put on the GPU. The rest will be on the CPU. If you don't know how many layers there are, you can use -1 to move all to GPU.
n_batch = 512  # Should be between 1 and n_ctx, consider the amount of RAM of your Apple Silicon Chip.
# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="/Users/rlm/Desktop/Code/llama.cpp/models/openorca-platypus2-13b.gguf.q4_0.bin",
    n_gpu_layers=n_gpu_layers,
    n_batch=n_batch,
    f16_kv=True,  # MUST set to True, otherwise you will run into problem after a couple of calls
    callback_manager=callback_manager,
    verbose=True,  # Verbose is required to pass to the callback manager
    grammar_path="/Users/rlm/Desktop/Code/langchain-main/langchain/libs/langchain/langchain/llms/grammars/json.gbnf",
)

%%capture captured --no-stdout
result = llm.invoke("Describe a person in JSON format:")

{
  "name": "John Doe",
  "age": 34,
  "": {
    "title": "Software Developer",
    "company": "Google"
  },
  "interests": [
    "Sports",
    "Music",
    "Cooking"
  ],
  "address": {
    "street_number": 123,
    "street_name": "Oak Street",
    "city": "Mountain View",
    "state": "California",
    "postal_code": 94040
  }}
``````output

llama_print_timings:        load time =   357.51 ms
llama_print_timings:      sample time =  1213.30 ms /   144 runs   (    8.43 ms per token,   118.68 tokens per second)
llama_print_timings: prompt eval time =   356.78 ms /     9 tokens (   39.64 ms per token,    25.23 tokens per second)
llama_print_timings:        eval time =  3947.16 ms /   143 runs   (   27.60 ms per token,    36.23 tokens per second)
llama_print_timings:       total time =  5846.21 ms

我们还可以提供 list.gbnf 以返回一个列表

n_gpu_layers = 1
n_batch = 512
llm = LlamaCpp(
    model_path="/Users/rlm/Desktop/Code/llama.cpp/models/openorca-platypus2-13b.gguf.q4_0.bin",
    n_gpu_layers=n_gpu_layers,
    n_batch=n_batch,
    f16_kv=True,  # MUST set to True, otherwise you will run into problem after a couple of calls
    callback_manager=callback_manager,
    verbose=True,
    grammar_path="/Users/rlm/Desktop/Code/langchain-main/langchain/libs/langchain/langchain/llms/grammars/list.gbnf",
)

%%capture captured --no-stdout
result = llm.invoke("List of top-3 my favourite books:")

["The Catcher in the Rye", "Wuthering Heights", "Anna Karenina"]
``````output

llama_print_timings:        load time =   322.34 ms
llama_print_timings:      sample time =   232.60 ms /    26 runs   (    8.95 ms per token,   111.78 tokens per second)
llama_print_timings: prompt eval time =   321.90 ms /    11 tokens (   29.26 ms per token,    34.17 tokens per second)
llama_print_timings:        eval time =   680.82 ms /    25 runs   (   27.23 ms per token,    36.72 tokens per second)
llama_print_timings:       total time =  1295.27 ms

相关

• LLM 概念指南
• LLM 操作指南