retro-rl

Some experimental RL algorithms for Retro Contest. Mostly reusing code from OpenAI repositories.

Name		Name	Last commit message	Last commit date
Latest commit History 111 Commits
batched_env		batched_env
confs		confs
dqn		dqn
exploration		exploration
models/dqn-softx_noisy_net_discount0.99_expert-policy_init-advantage_state-embedder-f64_explorelocal-celly10-w10_exploreglobal-celly10-w10_discover100K_cool100K_expert0.01		models/dqn-softx_noisy_net_discount0.99_expert-policy_init-advantage_state-embedder-f64_explorelocal-celly10-w10_exploreglobal-celly10-w10_discover100K_cool100K_expert0.01
ppo2		ppo2
rollouts		rollouts
scripts		scripts
vec_env		vec_env
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
autoencode.py		autoencode.py
behavioral_clone.py		behavioral_clone.py
collect_policy_targets.py		collect_policy_targets.py
custom_agent.py		custom_agent.py
dqn.docker		dqn.docker
dqn_agent.py		dqn_agent.py
jerk.docker		jerk.docker
jerk_agent.py		jerk_agent.py
ppo2.docker		ppo2.docker
ppo2_agent.py		ppo2_agent.py
random.docker		random.docker
random_agent.py		random_agent.py
replay_policy_targets.py		replay_policy_targets.py
sonic_util.py		sonic_util.py

Provide feedback