OpenAI gym framework for Blender Game Engine
OpenAI gym framework for Blender Game Engine
Extract Blend RL to [blend_rl_folder].
Extract blender_2_79b.tar.bz2 to [blender_executable_folder].
Open a terminal, run export PATH=$PATH:[blender_executable_folder] to add blender to path temporarily. You may wish to prepend it to ~/.bashrc to add blender to path permanently.
Run chmod +x [blend_rl_folder]/install to make install executable.
Extract [blend_rl_folder]/minimal_working_example_project.zip to [project_folder]; any folder you have read and write access.
Run cd [project_folder] to change your current working directory to your project folder.
Run blender main.blend to open your project.
Click the “start standalone player” button.
You should see a moving cube and a static plane.
Create a [project_folder].
Run cd [project_folder] to change your current working directory to your project folder.
Run [blend_rl_folder]/install to install Blend RL into your project.
Create the agent and environment objects in their respective .blend file.
Create the agent and environment scripts (for mutation) in their respective .py file.
NOTE: Make a group instance of the agent object. Move the agent object to an inactive layer.
Create your reinforcement learning algorithm in [project_folder]/main.py.
Run blender main.blend to open your project.
Click the “start standalone player” button.