added ogail script and splitting up feasability on next line to avoid calculating it unnecessarily (though it might be fine as is)

ogail-ppo-options-setobs2.py

@@ -0,0 +1,174 @@
+# %%
+import os
+
+import gym
+from src.safe_options.options import gail_ppo, Buffer
+from src.core.value import SetValue
+from src.safe_options.policy import SetMaskedDiscretePolicy
+from src.core.discriminator import DeepsetDiscriminator
+import torch
+
+from intersim.envs import IntersimpleLidarFlatRandom
+from intersim.envs.intersimple import speed_reward
+import functools
+from src.util.wrappers import CollisionPenaltyWrapper, TransformObservation, Setobs
+import numpy as np
+from src.safe_options.options import SafeOptionsEnv
+from torch.utils.tensorboard import SummaryWriter
+from ray import tune
+
+DIR = os.path.dirname(os.path.abspath(__file__))
+option_list = [[(vel, time) for vel in [0, 1, 2, 4, 6, 8, 10] for time in [5]],
+                [(vel, time) for vel in [0, 1, 2, 5, 7.5, 10] for time in [5, 20]],
+                [(vel, time) for vel in [0, 1, 2, 4, 6, 8, 10] for time in [5, 10, 20]], # was the best in training with single hidden layer, but very slow
+                [(vel, time) for vel in [0, 1, 2, 5, 7.5, 10] for time in [5, 20, 40]], 
+                [(vel, time) for vel in [0, 2, 5, 10] for time in [5, 10, 20]],
+                [(vel, time) for vel in [0, 3, 10] for time in [5, 20, 40]]
+]
+
+obs_min = np.array([
+    [-1000, -1000, 0, -np.pi, -1e-1, 0.],
+    [0, -np.pi, -20, -20, -np.pi, -1e-1],
+    [0, -np.pi, -20, -20, -np.pi, -1e-1],
+    [0, -np.pi, -20, -20, -np.pi, -1e-1],
+    [0, -np.pi, -20, -20, -np.pi, -1e-1],
+    [0, -np.pi, -20, -20, -np.pi, -1e-1],
+]).reshape(-1)
+
+obs_max = np.array([
+    [1000, 1000, 20, np.pi, 1e-1, 0.],
+    [50, np.pi, 20, 20, np.pi, 1e-1],
+    [50, np.pi, 20, 20, np.pi, 1e-1],
+    [50, np.pi, 20, 20, np.pi, 1e-1],
+    [50, np.pi, 20, 20, np.pi, 1e-1],
+    [50, np.pi, 20, 20, np.pi, 1e-1],
+]).reshape(-1)
+
+def training_function(config):
+    np.random.seed(config['seed'])
+    torch.manual_seed(config['seed'])
+
+    envs = sum([[SafeOptionsEnv(Setobs(
+        TransformObservation(CollisionPenaltyWrapper(IntersimpleLidarFlatRandom(
+            n_rays=5,
+            reward=functools.partial(
+                speed_reward,
+                collision_penalty=0
+            ),
+            stop_on_collision=config['stop_on_collision'], track=track,
+        ), collision_distance=6, collision_penalty=100), lambda obs: (obs - obs_min) / (obs_max - obs_min + 1e-10))
+    ), options=option_list[config['policy']['option']], safe_actions_collision_method=None, 
+    abort_unsafe_collision_method=None) for _ in range(20)] for track in range(4)],[])
+
+    env_fn = lambda i: envs[i]
+
+    policy = SetMaskedDiscretePolicy(env_fn(0).action_space.n, 
+        n_hidden_layers=config['policy']['n_hidden_layers'],
+        hidden_layer_size=config['policy']['hidden_layer_size'],
+        activation=config['policy']['activation'] ) # config net architecture
+    pi_opt = torch.optim.Adam(policy.parameters(), lr=config['policy']['learning_rate'])
+    pi_lr_scheduler = torch.optim.lr_scheduler.ExponentialLR(pi_opt, gamma=config['policy']['learning_rate_decay'])
+
+    value = SetValue() # config net architecture
+    v_opt = torch.optim.Adam(value.parameters(), lr=config['value']['learning_rate'])
+
+    discriminator = DeepsetDiscriminator() # config net architecture
+    disc_opt = torch.optim.Adam(discriminator.parameters(), lr=config['discriminator']['learning_rate'], weight_decay=config['discriminator']['weight_decay'])
+
+    expert_data = [
+        torch.load(os.path.join(DIR, 'intersimple-expert-data-setobs2-loc0-track0.pt')),
+        torch.load(os.path.join(DIR, 'intersimple-expert-data-setobs2-loc0-track1.pt')),
+        torch.load(os.path.join(DIR, 'intersimple-expert-data-setobs2-loc0-track2.pt')),
+        torch.load(os.path.join(DIR, 'intersimple-expert-data-setobs2-loc0-track3.pt')),
+    ]
+    d0 = [d[0] for d in expert_data]
+    d1 = [d[1] for d in expert_data]
+    d2 = [d[2] for d in expert_data]
+    d3 = [d[3] for d in expert_data]
+    expert_data = (torch.cat(d0), torch.cat(d1), torch.cat(d2), torch.cat(d3))
+    expert_data = Buffer(*expert_data)
+
+    def callback(info):
+        tune.report(gen_mean_reward_per_episode=info['gen/mean_reward_per_episode'],
+                    disc_mean_reward_per_episode=info['disc/mean_reward_per_episode'], 
+                    mean_episode_length=info['gen/mean_episode_length'])
+        
+        # save model checkpoints
+        ep = info['epoch'] + 1
+        if (ep % 25 == 0):
+            torch.save(info['policy'].state_dict(), f'policy_epoch{ep}.pt')
+
+    value, policy = gail_ppo(
+        env_fn=env_fn,
+        expert_data=expert_data,
+        discriminator=discriminator,
+        disc_opt=disc_opt,
+        disc_iters=config['discriminator']['iterations_per_epoch'],
+        policy=policy,
+        value=value,
+        v_opt=v_opt,
+        v_iters=config['value']['iterations_per_epoch'],
+        epochs=200,
+        rollout_episodes=60,
+        rollout_steps=60,
+        gamma=0.99,
+        gae_lambda=0.9,
+        clip_ratio=config['policy']['clip_ratio'],
+        pi_opt=pi_opt,
+        pi_iters=config['policy']['iterations_per_epoch'],
+        logger=SummaryWriter(comment='sgail-ppo-options-setobs2'),
+        callback=callback,
+        lr_schedulers=[pi_lr_scheduler],
+    )
+
+    # save model
+    torch.save(policy.state_dict(), 'policy_final.pt')
+
+analysis = tune.run(
+    training_function,
+    config={
+        'stop_on_collision': tune.grid_search([True, False]),
+        'policy': {
+            'learning_rate': 3e-4, # tune.grid_search([3e-4]),
+            'learning_rate_decay': 1.0, #tune.grid_search([1.0]),
+            'clip_ratio': 0.2, #tune.grid_search([0.2]),
+            'iterations_per_epoch': 100, #tune.grid_search([100]),
+            'hidden_layer_size': tune.grid_search([10, 20, 40]),
+            'n_hidden_layers': tune.grid_search([2, 3, 4]),
+            'activation':tune.grid_search([torch.nn.LeakyReLU, torch.nn.Tanh]),
+            'option': tune.grid_search(list(range(len(option_list))))
+        },
+        'value': {
+            'learning_rate': 1e-3, # tune.grid_search([1e-3]),
+            'iterations_per_epoch': 1000, #tune.grid_search([1000]),
+        },
+        'discriminator': {
+            'learning_rate': 1e-3, #tune.grid_search([1e-3]),
+            'weight_decay': 1e-4, #tune.grid_search([1e-4]),
+            'iterations_per_epoch': 100, #tune.grid_search([100]),
+        },
+        'seed': 0,
+    }
+)
+
+print('Best config: ', analysis.get_best_config(metric='gen_mean_reward_per_episode', mode='max'))
+
+# %%
+# policy = SetMaskedDiscretePolicy(env_fn(0).action_space.n)
+# policy(torch.zeros(env_fn(0).observation_space['observation'].shape), torch.zeros(env_fn(0).observation_space['safe_actions'].shape))
+# policy.load_state_dict(torch.load('sgail-ppo-options-setobs2.pt'))
+
+# env = env_fn(0)
+# obs = env.reset()
+# env.render(mode='post')
+# for i in range(300):
+#     action = policy.sample(policy(
+#         torch.tensor(obs['observation'], dtype=torch.float32),
+#         torch.tensor(obs['safe_actions'], dtype=torch.float32),
+#     ))
+#     obs, reward, done, _ = env.step(action, render_mode='post')
+#     print('step', i, 'reward', reward, 'safe actions', obs['safe_actions'])
+#     if done:
+#         break
+# env.close()
+# %%

src/safe_options/options.py

@@ -249,8 +249,8 @@ class SafeOptionsEnv(OptionsEnv):
             if d:
                 break
 
-            if self.abort_unsafe_collision_method is not None and \
+            if self.abort_unsafe_collision_method is not None:
-                not feasible(self.env, plan[k:], method=self.abort_unsafe_collision_method):
+                if not feasible(self.env, plan[k:], method=self.abort_unsafe_collision_method):
                     break
         
         n_steps = k + 1