Need help with AidLearning-FrameWork?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

4.1K Stars 620 Forks Other 667 Commits 12 Opened issues


🔥🔥AidLearning for Linux-Python-AI code development on Android. AidLearning builds a linux environment supporting GUI, deep learning and visual programming on Android devices..Now Aid supports GPUDelegate by python

Services available


Need anything else?

Contributors list

# 3,479
654 commits
# 115,222
3 commits


Linux+AI+Python+GUI 4in1 Environments Running on the Android . [中文版] [English]

Build Status Coverage Status License Repo size Code size Fork star commit


AidLearning is a powerful mobile AI development platform, which supports almost all frameworks and tools for deep learning neural network development. - It has built a complete Linux OS supporting GUI desktop on Android, built-in the most popular deep learning framework Caffe / mxnet / keras / tensorflow / ncnn / opencv/pytorch... - Built in Visual AI development IDE, built-in most popular programming tools such as VSCode and Jupyter, supporting touch-and-drop ui design, supporting dynamic debugging and running of code. Support the development of AI applications in Python on mobile and pad, and support the package your Python source code into app (APK) to release. - One click installation is supported. You only need to install a 10m app to automatically boot to complete the installation. - At present, it has been online in major app application centers, with more than 1.5 million downloads and visits, and a large number of AI examples and tutorials are built-in. There are also a large number of tutorials on the Internet to facilitate your learning and development. - Finally, on version 0.87,Aid adds CPU acceleration and GPU acceleration, and tflite is built into version 0.87. tflite GPU module can fully release the GPU performance of mobile phone, support CPU + GPU mode, and speed up at the same time! tflite.NNModel (model_ Path, inshape, outshape, 4,0) # - 4 represents 4 CPU threads, 0 represents GPU, - 1 represents CPU, and 1 represents the number of nnapi threads. I set 4 threads on aid. You can flexibly set the number of threads and use GPU + CPU mode

Download the v0.87F2 now!



  • The best Linux environment on android terminal (mobile phone), the only Linux os that supports GUI desktop without vnc running on the android.
  • It is the only simulator that supports AI development environment, built-in deep learning framework with the world's most popular top 7, and a large number of deep learning models, examples and development components
  • The only simulator that supports Python Visual development and debugging supports touch and drag interface design to improve your development efficiency
  • It supports the development of apps that can run on mobile phones with Python, and supports the direct compilation of Python code to generate deployable APK files
  • One click installation, without any dependence, you only need to install a 10m boot app on your mobile phone, and you can automatically complete the installation of all environments.
  • Cross platform development, support cloud desktop (mobile desktop and computer desktop are the same), can not only run on mobile phone or tablet or other embedded motherboard, but also can be directly accessed and developed on the computer side based on Web.
  • It supports openblas, multi thread and multi process, runs smoothly without jamming, and gives full play to the computing power of ARM CPU and GPU


  • Support Tensorflow, Caffe, mxnet, keras, python, ncnn, opencv, Scipy
  • Supports Python 2.7 / Python 3.7.3.
  • Buildin AidCode visual programming IDE, it also supports Jupyter notebook and Microsoft's vscode programming development tool.
  • Built in complete and native cross platform desktop, no need to install third-party VNC support, support the same desktop of computer and mobile phone
  • It supports not only mobile phone/pad, but also industrial arm board
  • The developed program can be deployed on both mobile phone and computer
  • It supports 99.5% of the mobile phones on the market, and has tested a full series of 64 bit mobile phones such as Huawei, vivo, oppo, Samsung and Xiaomi
  • Support Linux native xfce4 desktop, do not need to install VNC and other software
  • Support the development of pyqt5, pyGame, vortex, SDL, etc


  • Aid virtualizes a closed space on the mobile phone. It doesn't need root and won't destroy the content of your mobile phone.
  • Will not collect your personal privacy, all permissions can be set by yourselves ### Easy to use
  • One click installation, automatically download the latest dependency package, automatically configure the environment required for AI development, and reduce the threshold of AI development
  • Built in a large number of AI components, models, examples, tutorials, reduce the threshold of AI development, you can not understand AI algorithm, but you can use this platform to develop AI applications.
  • The built-in sensor control package can easily control various sensors on the mobile phone: sound, gyroscope, position, camera, etc
  • One mobile phone, two systems, Android and Linux co-exist, no restart, free switching; entertainment, development, learning three not wrong
  • It supports the automatic synchronization of mobile phone development and computer development code, supports interface touch and drag design, and automatically generates interface code
  • One click compilation and release of AI enabled apps
  • Extensible support Java, C + +, go... And other languages

support peripherals

  • The built-in sensor control package can easily control various sensors on the mobile phone: sound, gyroscope, position, camera, etc
  • Using OTG USB can support the extension of peripherals and control aduino, which can be programmed in Python
  • Using OTG USB can also support peripheral storage device read and write operations
  • It can be used as the operating system of intelligent robot


Aidlearning framework can be divided into two parts: Linux simulator and AI programming platform.

Linux simulator consists of terminal and desktop. The former builds a complete Linux simulator based on Android underlying Linux kernel and busybox command package, and you can install any dependency package you need with apt command; the latter builds a graphical operating desktop based on Web, which you can control the whole system with touch on your mobile phone. At the same time, the desktop supports cloud desktop, which you can easily access through a website on the computer.

AI programming platform is composed of deep learning framework and python visual programming framework (Python IDE). The former includes almost all the popular deep learning framework, which is responsible for the loading of models and the scheduling of calculation graphs, and includes the memory allocation and op implementation of each calculation. After that, a python visual rapid development platform is built, which can not only run and debug Python code online, but also support touch-pull interface design, and generate the final executable program and output APK file.

Quick start

Buildin Tools







  • VTE (libvte): Terminal emulator widget for GTK+, mainly used in gnome-terminal. Source, Open Issues, and All (including closed) issues.
  • iTerm 2: OS X terminal application. Source, Issues and Documentation (which includes iTerm2 proprietary escape codes).
  • Konsole: KDE terminal application. Source, in particular tests, Bugs and Wishes.
  • hterm: JavaScript terminal implementation from Chromium. Source, including tests, and Google group.
  • xterm: The grandfather of terminal emulators. Source.
  • Connectbot: Android SSH client. Source
  • Android Terminal Emulator: Android terminal app which Termux terminal handling is based on. Inactive. Source.
  • Termux: Android terminal and Linux environment - app repository. Source.
  • remi:Python REMote Interface library. Platform independent. In about 100 Kbytes, perfect for your diet.Source.
  • [Caffe]
  • [Tensorflow]
  • [Mxnet]
  • [Keras]
  • [ncnn]
  • [pytorch]
  • [opencv]

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.