Skip to content

universal environment for learning neural networks for robots with graphic visualization

Notifications You must be signed in to change notification settings

RemGo95/project-b512

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 

Repository files navigation

project-b512

Universal machine learning environment for robots with graphic visualization. Add your robot in CAD files with special point for axis of every rotation. Connect via pipe to machine learning based on Python.

screen1

Demo version of app for android - .apk file Just send file to your phone and install.

Permissions for application:

-Communication(TCP Socket connection for learning on PC and send data to andorid phone)

-Storage

screen2

Software and packages you will need to run neural network and simulation:

For host with Unity Engine:

  • Python (==3.7.8),
  • Tensorflow (==2.6.0),
  • Visual Studio 2019,
  • Keras (==2.6.0),
  • Opencv-Python (minimum version ==4.1.0.25),
  • tqdm (==4.31.1),
  • Unity Engine (2020.3.4f1),
  • Unity ML Agents (==1.0.8 - important, different version doesnt work in this case),

For virtual environment (python venv):

  • first, you will need to upgrade PIP to the newest version,
  • Pytorch (==1.7.0) - this is little bit tricki, because we can already have pytorch on our venv, be carefull dont install 2 versions of torch,
  • ML Agents,
  • Pillow,
  • Numpy,
  • tqdm. (try to install the same versions)

Optional (if we want to use our graphic card):

  • NVIDIA cuDNN.

About

universal environment for learning neural networks for robots with graphic visualization

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published