Coder Social home page Coder Social logo

mobimaliangsdk's Introduction


目录

  1. 安装步骤
  2. 模型下载
  3. 快速上手
  4. 线上摩笔马良

1. 安装步骤

1.1. 克隆项目

安装好 Ubuntu20.04 之后,在桌面点击鼠标右键,选择 New Folder 新建文件夹,命名为 mtai_workspace,双击文件夹进入

再次点击右键,选择 Open in Terminal 打开命令行,可以看到此时命令行前缀为:~/Desktop/mtai_workspace$ (如果系统语言为中文,则为 ~/桌面/mtai_workspace$,下同)

输入以下命令并回车,将会克隆项目并进入项目文件夹,此时命令行前缀变为:~/Desktop/mtai_workspace/MobiMaliangSDK$

git clone https://github.com/MooreThreads/MobiMaliangSDK;cd MobiMaliangSDK

1.2. 安装驱动

输入以下命令,将会安装驱动 musa_2.1.1-Ubuntu-dev_amd64.deb

sudo bash install_driver.sh

如果sudo要求输入用户密码则请输入;如果出现 Configuring lightdm 的UI,请将 Default display manager 选成 lightdm

如果看到类似以下的驱动信息,则说明驱动安装成功;如果显示器分辨率没有变得更大,最好重启以确保驱动生效

Package: musa
Status: install ok installed
Priority: optional
Section: graphics
Installed-Size: 258062
Maintainer: Moore Threads
Architecture: amd64
Version: 2.1.1-Ubuntu-dev
Replaces: mtsnd, mtgpu, musa-wine, musa_all-in-one, linux-firmware
Depends: libdrm2 (>= 2.4.99), lightdm, dkms
Conflicts: mtsnd, mtgpu, musa-wine, musa_all-in-one
Description: Moore Threads MUSA driver
Homepage: https://developer.mthreads.com/

1.3. 其他安装

如果重启了,请打开 MobiMaliangSDK 文件夹,点击右键,选择 Open in Terminal,以重新进入命令行,前缀为:~/Desktop/mtai_workspace/MobiMaliangSDK$

输入以下命令,以完成后续安装

sudo bash install.sh
以上安装文件将依次完成以下内容
sudo dpkg -i mtml_1.5.0.deb sgpu-dkms_1.1.1.deb mt-container-toolkit_1.5.0.deb
(cd /usr/bin/musa && sudo ./docker setup $PWD)

具体参考 摩尔线程容器运行时套件

  • 拉取 docker 镜像 registry.mthreads.com/mcconline/musa-pytorch-release-public:latest
sudo docker pull registry.mthreads.com/mcconline/musa-pytorch-release-public:latest

运行以下命令,创建容器 mtai_workspace

sudo docker run -id --name mtai_workspace --privileged -e MTHREADS_VISIBLE_DEVICES=all -p 1001:1001 -p 1002:1002 -p 1003:1003 -p 1004:1004 -p 1005:1005 -v ~/Desktop/mtai_workspace:/mtai_workspace:rw --shm-size 16G registry.mthreads.com/mcconline/musa-pytorch-release-public:latest

如果系统语言为中文,则运行以下命令

sudo docker run -id --name mtai_workspace --privileged -e MTHREADS_VISIBLE_DEVICES=all -p 1001:1001 -p 1002:1002 -p 1003:1003 -p 1004:1004 -p 1005:1005 -v ~/桌面/mtai_workspace:/mtai_workspace:rw --shm-size 16G registry.mthreads.com/mcconline/musa-pytorch-release-public:latest

输入以下命令,进入容器 mtai_workspace

sudo docker exec -it mtai_workspace /bin/bash

此时命令行前缀变成类似 (py38) root@xxxx,说明已经进入容器了,运行以下两行命令完成安装,其中第一行进入到项目目录,第二行完成容器内安装任务

cd /mtai_workspace/MobiMaliangSDK/
bash install_inside_docker.sh

1.4. 启动 WebUI

如果在容器外面,输入以下命令以进入容器

sudo docker exec -it mtai_workspace /bin/bash

如果已经在容器里了,输入以下命令,其中第一行进入到项目目录,第二行启动 WebUI 服务

cd /mtai_workspace/MobiMaliangSDK/
streamlit run frontend/main.py --server.port 1001

在浏览器地址中输入,localhost:1001 即可访问 WebUI,也可右键点击命令行中的 Network URL 并选择 Open Link

如果要停止 WebUI 服务,按 Ctrl + C 即可;停止服务后,如果要退出容器,输入 exit 即可

1.5. 常用 docker 命令

需要在容器外使用

  • sudo docker images -a:查看全部已下载的镜像
  • sudo docker ps -a:查看全部容器
  • sudo docker start container_name:启动名为 container_name 的容器
  • sudo docker stop container_name:停止名为 container_name 的容器

2. 模型下载

请您在使用推荐的模型前,仔细阅读并理解以下事宜:

  1. 推荐的模型为第三方软件或代码库,客户在下载使用推荐的模型时,需要了解并遵守该模型的用户协议或开源协议要求
  2. 如果您计划使用推荐的模型,请确保您的使用符合相应的用户协议或开源协议要求。某些开源协议可能对商业化使用有特定的限制或要求您公开源代码等。您需要自行承担了解和遵守这些规定的责任
  3. 公司不对推荐的模型作任何明示或暗示的保证或承诺。在使用这些模型时,您应自行承担风险和责任。公司不承担因使用这些模型而导致的任何直接或间接损失或损害

3. 快速上手

文生图
# python examples/example_txt2img.py --checkpoint_path <your model path> --prompt <your prompt> --output_path <result saved path>
python examples/example_txt2img.py --checkpoint_path models/Stable-diffusion/v1-5-pruned-emaonly.ckpt --prompt "A photo of an astronaut riding a horse on mars." --output_path outputs
图生图
# python examples/example_img2img.py --checkpoint_path <your model path> --image_path <your initial image path> --prompt <your prompt> --output_path <result saved path>
python examples/example_img2img.py --checkpoint_path models/Stable-diffusion/v1-5-pruned-emaonly.ckpt --image_path data/person.jpg --prompt "a pretty girl" --output_path outputs
边缘控制
# python examples/controlnets/example_canny2img.py --checkpoint_path <your model path> --canny_controlnet_path <your canny controlnet model path> --image_path <your image path> --prompt <your prompt> --output_path <result saved path>
python examples/controlnets/example_canny2img.py --checkpoint_path models/Stable-diffusion/v1-5-pruned-emaonly.ckpt --canny_controlnet_path models/controlnet/canny_v11 --image_path data/person.jpg --prompt "a pretty girl" --output_path outputs

更多的 controlnets 请移步 examples/controlnets 文件夹

LoRA 生图
# python examples/example_lora2img.py --checkpoint_path <your model path> --lora_path <your lora model path> --prompt <your prompt> --output_path <result saved path>
python examples/example_lora2img.py --checkpoint_path models/Stable-diffusion/v1-5-pruned-emaonly.ckpt --lora_path models/lora/moore_threads_lora.safetensors --prompt "a pretty girl" --output_path outputs
Text Inversion 生图
# python examples/example_textinversion2img.py --checkpoint_path <your model path> --text_inversion_path <your text inversion model path> --prompt <your prompt> --negative_prompt <your negative text inversion trigger word> --output_path <result saved path>
python examples/example_textinversion2img.py --checkpoint_path models/Stable-diffusion/v1-5-pruned-emaonly.ckpt --text_inversion_path models/text_embeddings/negative/moore_threads_easy_negative.safetensors --prompt "a pretty girl" --negative_prompt "moore_threads_easy_negative" --output_path outputs
图生文反推词
# python examples/example_tagger.py --checkpoint_path <your model path> --image_path <your image path>
python examples/example_tagger.py --checkpoint_path models/tools/deepbooru/model-resnet_custom_v3.pt --image_path data/person.jpg
超分
# python examples/example_super_resolution.py --checkpoint_path <your model path> --image_path <your image path> --output_path <result saved path>
python examples/example_super_resolution.py --checkpoint_path models/tools/realesrgan/RealESRGAN_x4plus.pth --image_path data/person.jpg --output_path outputs
中英翻译
# python examples/example_translator.py --checkpoint_path <your model path> --chinese_text <your Chinese text>
python examples/example_translator.py --checkpoint_path models/tools/zh2en --chinese_text "一个漂亮的小姑娘"

4. 线上摩笔马良

点击 链接,体验摩笔马良线上网站

或 扫描下方二维码,体验摩笔马良线上小程序

5. FAQ

如果您有任何问题,可以通过邮箱地址 [email protected] 与我们取得联系,我们将尽力解答您的问题

mobimaliangsdk's People

Contributors

honglunzhang-mt avatar yangshiyu-mt avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.