GitHub - real-itu/IR2L: Safe Reinforcement Learning through Transferable Instinct Networks

real-itu / IR2L Public

forked from djole/IR2L

Safe Reinforcement Learning through Transferable Instinct Networks

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 230 Commits
a2c_ppo_acktr		a2c_ppo_acktr
safety_gym_mod		safety_gym_mod
utils		utils
visualisations		visualisations
.gitignore		.gitignore
arguments.py		arguments.py
double_rl_loop_main.py		double_rl_loop_main.py
env_setup.sh		env_setup.sh
env_util.py		env_util.py
exp_dir_util.py		exp_dir_util.py
instinct_training_loop.py		instinct_training_loop.py
main.py		main.py
main_es.py		main_es.py
main_scratch.py		main_scratch.py
model.py		model.py
modular_safety.yml		modular_safety.yml
navigation_2d.py		navigation_2d.py
ob_rms.p		ob_rms.p
policy_training_loop.py		policy_training_loop.py
policy_training_loop_no_instinct.py		policy_training_loop_no_instinct.py
pretrained_baseline.pt		pretrained_baseline.pt
pretrained_instinct.pt		pretrained_instinct.pt
pretrained_instinct_h10.pt		pretrained_instinct_h10.pt
pretrained_instinct_h100.pt		pretrained_instinct_h100.pt
pretrained_policy.pt		pretrained_policy.pt
requirements.txt		requirements.txt
run_evolution.sh		run_evolution.sh
simpleES.py		simpleES.py
simpleGA.py		simpleGA.py
sweep_results.py		sweep_results.py
train_test_model.py		train_test_model.py
visualise_results.py		visualise_results.py