Adapted gym env to only del a visualizer if the environment name is in the...

Adapted gym env to only del a visualizer if the environment name is in the dictionary, as this is not the case during intialization of the rl env.

cooperative_cuisine/reinforcement_learning/config/model/PPO.yaml

@@ -8,10 +8,10 @@ model_type_inference:
   _partial_: true
   _target_: stable_baselines3.PPO.load
 total_timesteps:  3_000_000 # hendric sagt eher so 300_000_000 schritte
-number_envs_parallel:  64
+number_envs_parallel:  16
 learning_rate: 0.0003
 n_steps: 2048
-batch_size: 64
+batch_size: 16
 n_epochs: 10
 gamma: 0.99
 gae_lambda: 0.95

cooperative_cuisine/reinforcement_learning/gym_env.py

@@ -155,7 +155,7 @@ class EnvGymWrapper(Env):
         super().__init__()
 
         self.randomize_counter_placement = False
-        self.use_rgb_obs = True  # if False uses simple vectorized state
+        self.use_rgb_obs = False  # if False uses simple vectorized state
         self.full_vector_state = True
         config_env = OmegaConf.to_container(config.environment, resolve=True)
         config_item_info = OmegaConf.to_container(config.item_info, resolve=True)
@@ -258,7 +258,8 @@ class EnvGymWrapper(Env):
         """
         Resets the environment according to the configs
         """
-        del visualizer.surface_cache_dict[self.env.env_name]
+        if self.env.env_name in visualizer.surface_cache_dict:
+            del visualizer.surface_cache_dict[self.env.env_name]
         self.env: Environment = Environment(
             env_config=deepcopy(self.config_env),
             layout_config=self.layout,