Marathon Environments

I implemented one kind of Reinforcement Learning Physics-Based Character Interaction with this development environment: hit and bounce.

Please check this video:

Name		Name	Last commit message	Last commit date
Latest commit History 411 Commits
README		README
UnitySDK		UnitySDK
com.unity.ml-agents		com.unity.ml-agents
config		config
images		images
marathon-envs		marathon-envs
unity-volume		unity-volume
.gitignore		.gitignore
1.11.0		1.11.0
Changes.md		Changes.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
Reinforcement Learning of Physics-Based Character Interaction.mp4		Reinforcement Learning of Physics-Based Character Interaction.mp4
Reinforcement_Learning_of_Physics-Based_Character_Interaction.pdf		Reinforcement_Learning_of_Physics-Based_Character_Interaction.pdf
Training.md		Training.md
run-MarathonManBackflip_MarathonManBackflip-v0-tag-Environment_Cumulative Reward (1).csv		run-MarathonManBackflip_MarathonManBackflip-v0-tag-Environment_Cumulative Reward (1).csv
stable_baselines_sac_run.py		stable_baselines_sac_run.py
stable_baselines_sac_train.py		stable_baselines_sac_train.py

Provide feedback