Fixed Maniskill state-based training.

rl_games/algos_torch/network_builder.py

@@ -189,7 +189,6 @@ def _build_value_layer(self, input_size, output_size, value_type='legacy'):
             raise ValueError('value type is not "default", "legacy" or "two_hot_encoded"')
 
 
-
 class A2CBuilder(NetworkBuilder):
     def __init__(self, **kwargs):
         NetworkBuilder.__init__(self)
@@ -471,7 +470,7 @@ def forward(self, obs_dict):
                     else:
                         sigma = self.sigma_act(self.sigma(out))
                     return mu, mu*0 + sigma, value, states
-                    
+
         def is_separate_critic(self):
             return self.separate
 

rl_games/common/a2c_common.py

@@ -416,7 +416,6 @@ def get_action_values(self, obs):
             'obs' : processed_obs,
             'rnn_states' : self.rnn_states
         }
-
         with torch.no_grad():
             res_dict = self.model(input_dict)
             if self.has_central_value:
@@ -757,6 +756,7 @@ def play_steps(self):
                 res_dict = self.get_masked_action_values(self.obs, masks)
             else:
                 res_dict = self.get_action_values(self.obs)
+
             self.experience_buffer.update_data('obses', n, self.obs['obs'])
             self.experience_buffer.update_data('dones', n, self.dones)
 
@@ -782,7 +782,7 @@ def play_steps(self):
             self.current_lengths += 1
             all_done_indices = self.dones.nonzero(as_tuple=False)
             env_done_indices = all_done_indices[::self.num_agents]
-     
+
             self.game_rewards.update(self.current_rewards[env_done_indices])
             self.game_shaped_rewards.update(self.current_shaped_rewards[env_done_indices])
             self.game_lengths.update(self.current_lengths[env_done_indices])

rl_games/configs/maniskill/maniskill.yaml

@@ -20,13 +20,13 @@ params:
           val: 0
         fixed_sigma: True
     mlp:
-      units: [512, 256, 128]
+      units: [256, 128, 64]
       activation: elu
       initializer:
         name: default
 
   config:
-      name: Maniskill
+      name: PickCube_state
       env_name: maniskill
       normalize_input: True
       normalize_value: True
@@ -48,11 +48,13 @@ params:
       use_smooth_clamp: True
       bound_loss_type: regularisation
       bounds_loss_coef: 0.0005
-      max_epochs: 2000
+      max_epochs: 1000
+      save_best_after: 25
+      save_frequency: 100
       num_actors: 4096
-      horizon_length: 64
+      horizon_length: 16
       minibatch_size: 16384
-      mini_epochs: 5
+      mini_epochs: 4
       critic_coef: 2
 
       env_config:

rl_games/envs/maniskill.py

@@ -148,7 +148,11 @@ def __init__(self, config_name, num_envs, **kwargs):
         # else:
         #     self.observation_space = gym.spaces.Box(-self._clip_obs, self._clip_obs, policy_obs_space.shape)
 
-        self.action_space = self.env.unwrapped.single_action_space
+        self._clip_actions = 1.0
+
+        action_space = self.env.unwrapped.single_action_space
+        print("Single action apace:", action_space)
+        self.action_space = gym.spaces.Box(-self._clip_actions, self._clip_actions, action_space.shape)
 
     def step(self, action):
         # # move actions to sim-device
@@ -163,12 +167,14 @@ def step(self, action):
         extras["time_outs"] = truncated #truncated.to(device=self._rl_device)
         # process observations and states
         #obs_and_states = self._process_obs(obs_dict)
-        obs_and_states = obs_dict
+
+        obs_and_states = {'obs': obs_dict}
         # move buffers to rl-device
         # note: we perform clone to prevent issues when rl-device and sim-device are the same.
         #rew = rew.to(device=self._rl_device)
         #dones = (terminated | truncated).to(device=self._rl_device)
-        dones = (terminated | truncated).any() # stop if any environment terminates/truncates
+        dones = (terminated | truncated) # stop if any environment terminates/truncates
+
         # extras = {
         #     k: v.to(device=self._rl_device, non_blocking=True) if hasattr(v, "to") else v for k, v in extras.items()
         # }
@@ -197,10 +203,12 @@ def step(self, action):
 
     def reset(self):
         obs = self.env.reset()
+        # print(obs)
+        print("obs reset shape:", obs[0].shape)
         # if self.flatten_obs:
         #     obs = flatten_dict(obs)
 
-        return obs
+        return {'obs': obs[0]}
 
     def get_number_of_agents(self):
         return 1
@@ -209,6 +217,7 @@ def get_env_info(self):
         info = {}
         info['action_space'] = self.action_space
         info['observation_space'] = self.observation_space
+        print("info:", info)
         return info