Skip to content

Masterthesis: Reinforcement learning for playing the wire-loop game with a real robot and a camera

Notifications You must be signed in to change notification settings

chrisbitter/cense

Repository files navigation

CENSE

This project uses Machine Learning to learn the Loop-Wire Game

Git Conventions

###Branch Names

Convention: topic_short/initials/branchname

####Topics

Topic topic_short Description
ro Robot direct interfaces with the hardware
ag Agent learning algorithms
en Environment -
si Simulation -
mi Middleware managing settings and training
ui User Interface -
vr Virtual Reality -

####Initials

WM: First letter of first and last name

  • Max Muster → mm

HiWi: First two letters of first and last name

  • Max Muster → mamu

####branch name

Requirements

  • Python 3.5

Setup

  • install wheels in Resources/python_wheels
    • in Console: pip install path_to_wheel

Hacks

  • in VideoCapture/init.py, change im = Image.fromstring to im = Image.frombytes
    • there will be an error when running pointing to the right line of code.
  • in keras/utils/generic_utils change line 175 to "code = marshal.dumps(func.code).replace(b'\',b'/').decode('raw_unicode_escape')"
    • needed because of lam

Remote Access

  • copy Resources/credentials_template.json file
  • rename to credentials.json
  • enter valid credentials
  • YOU OBVIOUSLY MUST NOT DISTRIBUTE credentials.json to Git!
    • Note: credentials.json is in .gitignore

About

Masterthesis: Reinforcement learning for playing the wire-loop game with a real robot and a camera

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages