dodox/metastable-baselines
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Show confidence-ellipsoid for supported envs

Browse Source
This commit is contained in:
Dominik Moritz Roth 2022-07-17 00:48:17 +02:00
parent 49f9acff3e
commit 9133ecd61b
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
replay.py
View File

@ -28,6 +28,8 @@ def main(load_path, n_eval_episodes=0):
model = PPO.load(load_path, env=env) model = PPO.load(load_path, env=env)
show_chol = env_name.startswith('Columbus')
if n_eval_episodes: if n_eval_episodes:
mean_reward, std_reward = evaluate_policy( mean_reward, std_reward = evaluate_policy(
model, env, n_eval_episodes=n_eval_episodes, deterministic=False) model, env, n_eval_episodes=n_eval_episodes, deterministic=False)
@ -41,7 +43,10 @@ def main(load_path, n_eval_episodes=0):
time.sleep(1/30) time.sleep(1/30)
action, _ = model.predict(obs, deterministic=False) action, _ = model.predict(obs, deterministic=False)
obs, reward, done, info = env.step(action) obs, reward, done, info = env.step(action)
env.render() if show_chol:
env.render(chol=model.policy.chol)
else:
env.render()
episode_reward += reward episode_reward += reward
if done: if done:
episode_reward = 0.0 episode_reward = 0.0