3 Star 18 Fork 3

Gitee 极速下载/fastllm

加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
此仓库是为了提升国内下载速度的镜像仓库,每日同步一次。 原始仓库: https://github.com/ztxz16/fastllm
克隆/下载
贡献代码
同步代码
取消
提示: 由于 Git 不支持空文件夾,创建文件夹后会生成空的 .keep 文件
Loading...
README
Apache-2.0

fastllm

| 快速开始 | 部署DeepSeek | 部署Qwen3 | 版本日志 | English Document

介绍

fastllm是c++实现自有算子替代Pytorch的高性能全功能大模型推理库,可以推理Qwen, Llama, Phi等稠密模型,以及DeepSeek, Qwen-moe等moe模型

  • 具有优良的兼容性,支持M40, K80到5090全系列N卡,支持MI50,7900等A卡,支持天数,沐曦等国产卡,支持ThinkForce NPU推理
  • 支持任意显卡的FP8推理
  • 任意显卡只需要显存 > 10G就可以支持单卡推理满血DeepSeek R1 671B模型
  • 双路9004/9005服务器 + 单显卡部署DeepSeek R1 671B - FP8原版模型,单并发速度可达20左右,部署INT4模型单并发速度可达30左右,最高并发速度可达60+

部署交流QQ群: 831641348

微信群:目前群聊超过200人,请添加小助手微信号fastllmxzs加群:

亮点功能

  • 🚀 安装使用简单方便,一条命令就能成功安装,一条命令就能成功运行。
  • 🚀 支持CPU + GPU混合推理MOE大参数模型(单显卡即可推理DEEPSEEK 671B)。
  • 🚀 使用C++实现自有底层算子,不依赖PyTorch。
  • 🚀 兼容性好,PIP安装支持可以支持到P100、MI50等老卡,源码安装支持更多设备。
  • 🚀 支持多卡张量并行推理,支持3、5、7等奇数张卡。
  • 🚀 支持GPU + CPU混合张量并行推理
  • 🚀 支持CPU和显卡实现FP8运算,老设备也可以运行
  • 🚀 支持多CPU加速,且只占用1份内存
  • 🚀 支持ROCM,AMD GPU;支持天数,沐曦,燧原;支持华为昇腾。
  • 🚀 支持动态Batch,流式输出;前后端分离设计,可跨平台移植,可在安卓上直接编译。
  • 🚀 支持Python自定义模型结构

快速开始

安装

  • PIP安装(目前仅支持Nvidia GPU和AMD GPU,其余GPU请使用源码安装

Linux系统 + Nvidia GPU:

由于目前PyPI限制库大小,安装包中不含CUDA依赖,安装ftllm之前建议先手动安装CUDA 12以上版本 (已安装cuda可跳过)

wget https://developer.download.nvidia.com/compute/cuda/12.8.1/local_installers/cuda_12.8.1_570.124.06_linux.run
sudo sh cuda_12.8.1_570.124.06_linux.run

然后用pip安装,命令如下:

pip install ftllm -U

Linux系统 + AMD GPU:

由于目前PyPI限制库大小,安装包中不含ROCM依赖,安装ftllm之前建议先手动安装ROCM 6.3.3 (若已安装ROCM可跳过)

wget wget https://repo.radeon.com/amdgpu-install/6.3.3/ubuntu/jammy/amdgpu-install_6.3.60303-1_all.deb
apt install ./amdgpu-install_6.3.60303-1_all.deb -y
amdgpu-install --usecase=hiplibsdk,rocm,dkms -y

然后用pip安装,命令如下:

pip install ftllm-rocm -U

Windows系统 (仅支持Nvidia GPU):

第一次安装前需要安装依赖库:

pip install https://hf-mirror.com/fastllm/fastllmdepend-windows/resolve/main/ftllmdepend-0.0.0.1-py3-none-win_amd64.whl

然后用pip安装,命令如下:

pip install ftllm -U
  • Hint

Conda下安装有时候会出现环境错误,如果出现可以尝试在Conda外或使用venv等虚拟环境尝试

(若使用时报错,可参考ftllm报错 )

运行demo程序

可以运行一个较小模型测试安装是否成功, 以Qwen/Qwen3-0.6B模型为例

命令行聊天:

ftllm run Qwen/Qwen3-0.6B

WebUI:

ftllm webui Qwen/Qwen3-0.6B

API Server (OpenAI 风格):

ftllm server Qwen/Qwen3-0.6B

使用指南

1. 如何启动模型

基本的启动命令格式如下:

ftllm run Qwen/Qwen3-0.6B # 启动本地对话
ftllm webui Qwen/Qwen3-0.6B # 启动WebUI
ftllm server Qwen/Qwen3-0.6B # 启动API Server

根据你需要开启的服务,选择相应的命令。以 server 命令为例,格式如下:

ftllm server model

这里的model可以是:

  • Huggingface上的模型,例如 Qwen/Qwen3-0.6B 代表 千问3-0.6B模型
  • 本地模型路径。例如/mnt/Qwen3-0.6B,高速下载模型可以参考 模型下载

无论是在线模型还是本地模型,目前支持以下几种格式 (均以在线模型举例,可以在Huggingface上搜到对应模型):

  • FP16, BF16格式的原始模型,例如Qwen/Qwen3-0.6B
  • FP8格式的模型,例如Qwen/Qwen3-0.6B-FP8
  • AWQ格式的模型,例如Qwen/Qwen3-14B-AWQ
  • Fastllm格式的模型,例如fastllm/DeepSeek-V3-0324-INT4。也可以下载原始模型后通过 模型导出 中的命令导出
  • 目前暂不支持 GGUF 格式的模型

如果您是第一次使用ftllm,建议直接使用基本的启动命令启动,所有的参数都会自动设置。如果您希望继续调参,请参照下面的参数设置说明

2. 如何设定推理精度

当启动的模型为浮点精度时(BF16, FP16, FP8)时,可以通过以下参数来设置模型的推理精度:

  • --dtype:

    • 描述: 指定模型的数据类型。
    • 可选值: int4g int4 int8 fp8 float16 或其他支持的数据类型。
    • 示例: --dtype int4
  • --moe_dtype:

    • 描述: 指定模型MOE层的数据类型。
    • 可选值: int4g int4 int8 fp8 float16 或其他支持的数据类型。
    • 示例: --moe_dtype int4
    • 说明: 如果指定的模型不是moe结构的模型,这个参数不会生效

命令示例:

ftllm server Qwen/Qwen3-0.6B --dtype int8 
# 上面的命令会读取原始模型(这个模型是BF16精度),并在线量化为INT8精度推理

ftllm server deepseek-ai/DeepSeek-V3-0324 --dtype fp8 --moe_dtype int4
# 上面的命令会读取原始模型(这个模型是FP8精度),并使用FP8 + INT4的混合精度推理

若不设定这些参数,ftllm会使用模型中设定的精度来进行推理

若使用的模型已经是量化好的模型(例如AWQ模型,Fastllm导出的量化模型等),建议不指定这些参数

3. 如何设定运行设备

可以通过以下参数来设定执行推理的设备

  • --device:

    • 描述: 指定模型运行的计算设备。
    • 示例: --device cpu, --device cuda
    • 常用值说明:
      • cpu 使用cpu推理
      • cuda 使用gpu推理
      • numa 使用多路numa节点加速推理,在多CPU的机器才会有提升
      • multicuda 使用多设备张量并行推理
        • 使用多显卡: --device multicuda:0,1
        • 使用多显卡+CPU: --device multicuda:0,cpu
        • 按比例使用多显卡+CPU: --device multicuda:0:4,1:5,cpu:1 (cuda:0计算4/10, cuda:1计算5/10, cpu计算1/10)
  • --moe_device:

    • 描述: 指定 MOE(Mixture of Experts)层的计算设备。
    • 示例: --moe_device cpu, --moe_device numa
    • 常用值说明:
      • cpu 使用cpu推理
      • numa 使用多路numa节点加速推理,在多CPU的机器才会有提升
      • cuda 使用gpu推理 (MOE层需要大量显存,一般不建议指定为cuda
    • 说明: 一般和device指定为不同的设备实现混合推理,例如 --device cuda --moe_device cpu来实现MOE模型的单卡+CPU混合推理。 --device cuda --moe_device numa 来实现MOE模型的单卡+多NUMA节点加速推理 如果指定的模型不是moe结构的模型,这个参数不会生效

若不设定这些参数,会使用默认配置来推理,默认配置如下:

模型类型 device moe_device
稠密模型 cuda 不生效
MOE模型 cuda cpu

对于发烧友而言,如果想更进一步榨干硬件,可以参考 混合推理指南

4. 如何设定运行参数

可以通过下列参数设置运行参数。

需要注意的是,速度和参数设置并不一定正相关,如果对性能要求高,可以多方向尝试一下

  • -t--threads:
    • 描述: 设置使用的CPU线程数。
      • devicecpu时,这个参数决定了推理使用的线程数
      • devicenuma时,推理线程数主要由环境变量FASTLLM_NUMA_THREADS决定,threads参数请设得小一点(推荐设为1)
    • 示例: -t 27

例如我们在多CPU设备上用GPU + 多CPU混合部署一个MOE模型fastllm/DeepSeek-V3-0324-INT4,可以尝试这些命令:

export FASTLLM_NUMA_THREADS=27 && ftllm server fastllm/DeepSeek-V3-0324-INT4 --device cuda --moe_device numa -t 1 
# 使用多numa推理,每个numa节点使用27个线程

export FASTLLM_NUMA_THREADS=16 && ftllm server fastllm/DeepSeek-V3-0324-INT4 --device cuda --moe_device numa -t 1 
# 使用多numa推理,每个numa节点使用16个线程

numactl -C 0-31 -m 0 ftllm server fastllm/DeepSeek-V3-0324-INT4 --device cuda --moe_device cpu -t 27 
# 绑定单numa节点,使用CPU推理,使用27线程

不同硬件上,不同参数发挥出的性能有很大不同。一般而言,CPU上使用的线程数不建议超过物理核数

5. 其它参数

  • --moe_experts:

    • 描述: 指定 MOE(Mixture of Experts)层使用的专家数。不设定则根据模型配置设定。减少专家数可以提高推理速度,但可能降低推理准确度
    • 示例: --moe_experts 6
  • --port:

    • 描述: 指定服务运行的端口号。
    • 示例: --port 8080
  • --help:

    • 描述: 查看模块参数详细信息。
    • 示例: ftllm server --help
  • --version-v:

    • 描述: 查看ftllm版本号。
    • 示例: ftllm -v
  • --hide_input:

    • 描述: server模式隐藏日志中的请求信息。
    • 示例: ftllm server --hide_input
  • --api_key:

    • 描述: server模式设定api_key。
    • 示例: ftllm server --api_key xxxxxxxx
  • --max_token:

    • 描述: webui模式指定最大输出。
    • 示例: ftllm webui --max_token
  • --think:

    • 描述: 强制思考。
    • 示例: ftllm webui --think
  • --cache_dir:

    • 描述: 指定在线Huggingface模型的缓存目录
    • 示例: ftllm --cache_dir /mnt

模型获取

模型下载

可以使用如下命令将模型下载到本地(使用高速镜像,无需科学上网)

ftllm download deepseek-ai/DeepSeek-R1

模型导出

如果使用量化加载模型(如--dtype int4),那么每次读取模型时会在线量化,读取速度较慢。

ftllm export 是一个用于导出和转换模型权重的工具。它支持将模型权重转换为不同的数据类型。以下是如何使用 ftllm export 的详细说明。

命令格式

ftllm export <模型路径> -o <输出路径> --dtype <数据类型> -t <线程数>

示例命令

ftllm export /mnt/DeepSeek-V3 -o /mnt/DeepSeek-V3-INT4 --dtype int4 -t 16

混合精度

可以通过指定--moe_dtype来实现混合精度,例如

ftllm export /mnt/DeepSeek-V3 -o /mnt/DeepSeek-V3-FP16INT4 --dtype float16 --moe_dtype int4 -t 16

加载导出后的模型

导出后的模型使用方法和原始模型类似,使用导出模型时--dtype参数将被忽略

例如

ftllm run /mnt/DeepSeek-V3-INT4/

支持的模型

如果需要运行更多早期的模型,请参考支持模型列表

源码安装

若pip安装失败或有其它特殊需求,可以用源码编译安装 源码安装后如果需要卸载,方法和PIP安装一样

pip uninstall ftllm

建议使用cmake编译,需要提前安装gcc,g++ (建议9.4以上), make, cmake (建议3.23以上)

GPU编译需要提前安装好CUDA编译环境,建议使用尽可能新的CUDA版本

使用如下命令编译

bash install.sh -DUSE_CUDA=ON -D CMAKE_CUDA_COMPILER=$(which nvcc) # 编译GPU版本
# bash install.sh -DUSE_CUDA=ON -DCUDA_ARCH=89 -D CMAKE_CUDA_COMPILER=$(which nvcc) # 可以指定CUDA架构,如4090使用89架构
# bash install.sh # 仅编译CPU版本
其他平台编译

其他不同平台的编译可参考文档

TFACC平台
ROCm平台

编译中遇到问题可参考 FAQ文档

Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS APPENDIX: How to apply the Apache License to your work. To apply the Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives. Copyright [yyyy] [name of copyright owner] Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

简介

fastllm 是纯 C++ 实现的全平台 llm 加速库 展开 收起
README
Apache-2.0
取消

发行版

暂无发行版

贡献者

全部

近期动态

不能加载更多了
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化
C/C++
1
https://gitee.com/mirrors/fastllm.git
git@gitee.com:mirrors/fastllm.git
mirrors
fastllm
fastllm
master

搜索帮助