pybullet a bit more refactoring, moving around files.

pybullet: move data to pybullet_data package, with getDataPath() method

examples/pybullet/gym/pybullet_envs/baselines/__init__.py

@@ -0,0 +1,10 @@
+from . import enjoy_kuka_grasping
+from . import enjoy_pybullet_cartpole
+from . import enjoy_pybullet_racecar
+from . import enjoy_pybullet_zed_racecar
+from . import train_kuka_cam_grasping
+from . import train_kuka_grasping
+from . import train_pybullet_cartpole
+from . import train_pybullet_racecar
+from . import train_pybullet_zed_racecar
+

examples/pybullet/gym/pybullet_envs/baselines/enjoy_kuka_grasping.py

@@ -0,0 +1,32 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.kukaGymEnv import KukaGymEnv
+
+from baselines import deepq
+
+
+def main():
+    
+    env = KukaGymEnv(renders=True)
+    act = deepq.load("kuka_model.pkl")
+    print(act)
+    while True:
+        obs, done = env.reset(), False
+        print("===================================")        
+        print("obs")
+        print(obs)
+        episode_rew = 0
+        while not done:
+            env.render()
+            obs, rew, done, _ = env.step(act(obs[None])[0])
+            episode_rew += rew
+        print("Episode reward", episode_rew)
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/enjoy_pybullet_cartpole.py

@@ -0,0 +1,35 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+
+from baselines import deepq
+from pybullet_envs.bullet.cartpole_bullet import CartPoleBulletEnv
+
+def main():
+    env = gym.make('CartPoleBulletEnv-v0')
+    act = deepq.load("cartpole_model.pkl")
+
+    while True:
+        obs, done = env.reset(), False
+        print("obs")
+        print(obs)
+        print("type(obs)")
+        print(type(obs))
+        episode_rew = 0
+        while not done:
+            env.render()
+           
+            o = obs[None]
+            aa = act(o)
+            a = aa[0]
+            obs, rew, done, _ = env.step(a)
+            episode_rew += rew
+        print("Episode reward", episode_rew)
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/enjoy_pybullet_racecar.py

@@ -0,0 +1,32 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.racecarGymEnv import RacecarGymEnv
+
+from baselines import deepq
+
+
+def main():
+    
+    env = RacecarGymEnv(renders=True,isDiscrete=True)
+    act = deepq.load("racecar_model.pkl")
+    print(act)
+    while True:
+        obs, done = env.reset(), False
+        print("===================================")        
+        print("obs")
+        print(obs)
+        episode_rew = 0
+        while not done:
+            env.render()
+            obs, rew, done, _ = env.step(act(obs[None])[0])
+            episode_rew += rew
+        print("Episode reward", episode_rew)
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/enjoy_pybullet_zed_racecar.py

@@ -0,0 +1,32 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.racecarZEDGymEnv import RacecarZEDGymEnv
+
+from baselines import deepq
+
+
+def main():
+    
+    env = RacecarZEDGymEnv(renders=True)
+    act = deepq.load("racecar_zed_model.pkl")
+    print(act)
+    while True:
+        obs, done = env.reset(), False
+        print("===================================")        
+        print("obs")
+        print(obs)
+        episode_rew = 0
+        while not done:
+            env.render()
+            obs, rew, done, _ = env.step(act(obs[None])[0])
+            episode_rew += rew
+        print("Episode reward", episode_rew)
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/train_kuka_cam_grasping.py

@@ -0,0 +1,50 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.kukaCamGymEnv import KukaCamGymEnv
+
+from baselines import deepq
+
+import datetime
+
+
+
+def callback(lcl, glb):
+    # stop training if reward exceeds 199
+    total = sum(lcl['episode_rewards'][-101:-1]) / 100
+    totalt = lcl['t']
+    #print("totalt")
+    #print(totalt)
+    is_solved = totalt > 2000 and total >= 10
+    return is_solved
+
+
+def main():
+  	
+    env = KukaCamGymEnv(renders=True)
+    model = deepq.models.cnn_to_mlp(
+        convs=[(32, 8, 4), (64, 4, 2), (64, 3, 1)],
+        hiddens=[256],
+        dueling=False
+    )
+    act = deepq.learn(
+        env,
+        q_func=model,
+        lr=1e-3,
+        max_timesteps=10000000,
+        buffer_size=50000,
+        exploration_fraction=0.1,
+        exploration_final_eps=0.02,
+        print_freq=10,
+        callback=callback
+    )
+    print("Saving model to kuka_cam_model.pkl")
+    act.save("kuka_cam_model.pkl")
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/train_kuka_grasping.py

@@ -0,0 +1,46 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.kukaGymEnv import KukaGymEnv
+
+from baselines import deepq
+
+import datetime
+
+
+
+def callback(lcl, glb):
+    # stop training if reward exceeds 199
+    total = sum(lcl['episode_rewards'][-101:-1]) / 100
+    totalt = lcl['t']
+    #print("totalt")
+    #print(totalt)
+    is_solved = totalt > 2000 and total >= 10
+    return is_solved
+
+
+def main():
+  	
+    env = KukaGymEnv(renders=False)
+    model = deepq.models.mlp([64])
+    act = deepq.learn(
+        env,
+        q_func=model,
+        lr=1e-3,
+        max_timesteps=10000000,
+        buffer_size=50000,
+        exploration_fraction=0.1,
+        exploration_final_eps=0.02,
+        print_freq=10,
+        callback=callback
+    )
+    print("Saving model to kuka_model.pkl")
+    act.save("kuka_model.pkl")
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/train_pybullet_cartpole.py

@@ -0,0 +1,39 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.cartpole_bullet import CartPoleBulletEnv
+
+from baselines import deepq
+
+
+def callback(lcl, glb):
+    # stop training if reward exceeds 199
+    is_solved = lcl['t'] > 100 and sum(lcl['episode_rewards'][-101:-1]) / 100 >= 199
+    return is_solved
+
+
+def main():
+	
+    env = CartPoleBulletEnv(renders=False)
+    model = deepq.models.mlp([64])
+    act = deepq.learn(
+        env,
+        q_func=model,
+        lr=1e-3,
+        max_timesteps=100000,
+        buffer_size=50000,
+        exploration_fraction=0.1,
+        exploration_final_eps=0.02,
+        print_freq=10,
+        callback=callback
+    )
+    print("Saving model to cartpole_model.pkl")
+    act.save("cartpole_model.pkl")
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/train_pybullet_racecar.py

@@ -0,0 +1,44 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.racecarGymEnv import RacecarGymEnv
+
+from baselines import deepq
+
+import datetime
+
+
+
+def callback(lcl, glb):
+    # stop training if reward exceeds 199
+    total = sum(lcl['episode_rewards'][-101:-1]) / 100
+    totalt = lcl['t']
+    is_solved = totalt > 2000 and total >= -50
+    return is_solved
+
+
+def main():
+  
+    env = RacecarGymEnv(renders=False,isDiscrete=True)
+    model = deepq.models.mlp([64])
+    act = deepq.learn(
+        env,
+        q_func=model,
+        lr=1e-3,
+        max_timesteps=10000,
+        buffer_size=50000,
+        exploration_fraction=0.1,
+        exploration_final_eps=0.02,
+        print_freq=10,
+        callback=callback
+    )
+    print("Saving model to racecar_model.pkl")
+    act.save("racecar_model.pkl")
+
+
+if __name__ == '__main__':
+    main()

examples/pybullet/gym/pybullet_envs/baselines/train_pybullet_zed_racecar.py

@@ -0,0 +1,47 @@
+#add parent dir to find package. Only needed for source code build, pip install doesn't need it.
+import os, inspect
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(os.path.dirname(currentdir))
+os.sys.path.insert(0,parentdir)
+
+import gym
+from pybullet_envs.bullet.racecarZEDGymEnv import RacecarZEDGymEnv
+
+from baselines import deepq
+
+import datetime
+
+
+
+def callback(lcl, glb):
+    # stop training if reward exceeds 199
+    total = sum(lcl['episode_rewards'][-101:-1]) / 100
+    totalt = lcl['t']
+    is_solved = totalt > 2000 and total >= -50
+    return is_solved
+
+def main():
+  
+    env = RacecarZEDGymEnv(renders=False, isDiscrete=True)
+    model = deepq.models.cnn_to_mlp(
+        convs=[(32, 8, 4), (64, 4, 2), (64, 3, 1)],
+        hiddens=[256],
+        dueling=False
+    )
+    act = deepq.learn(
+        env,
+        q_func=model,
+        lr=1e-3,
+        max_timesteps=10000,
+        buffer_size=50000,
+        exploration_fraction=0.1,
+        exploration_final_eps=0.02,
+        print_freq=10,
+        callback=callback
+    )
+    print("Saving model to racecar_zed_model.pkl")
+    act.save("racecar_zed_model.pkl")
+
+
+if __name__ == '__main__':
+    main()