Update app.py

app.py

@@ -9,8 +9,7 @@ from stable_baselines3.common.env_util import make_atari_env
 from stable_baselines3.common.vec_env import VecFrameStack
 from stable_baselines3.common.env_util import make_atari_env
 
-
-st.title("Atari Environments Live Model")
+st.subheader("Atari 2600 Deep RL Environments Live AI")
 
 # @st.cache This is not cachable :(
 def load_env(env_name):
@@ -18,7 +17,6 @@ def load_env(env_name):
     env = VecFrameStack(env, n_stack=4)
     return env
 
-
 # @st.cache This is not cachable :(
 def load_model(env_name):
     custom_objects = {
@@ -26,20 +24,15 @@ def load_model(env_name):
         "lr_schedule": lambda _: 0.0,
         "clip_range": lambda _: 0.0,
     }
-
     checkpoint = load_from_hub(
         f"ThomasSimonini/ppo-{env_name}",
         f"ppo-{env_name}.zip",
     )
-
     model = PPO.load(checkpoint, custom_objects=custom_objects)
-
     return model
     
-st.write("In game theory and optimization Nash Equilibrium loss minimization starts playing randomly but then by understanding ratios of action success to action-reward with an action (observe, decide/predict, act and then observe outcome the Deep RL agents go from 50% efficiency to 98-99% efficiency based on quality of decision without making mistakes.")
-
-st.write("list of agent environments https://github.com/DLR-RM/rl-baselines3-zoo/blob/master/benchmark.md")
-st.write("Deep RL models: https://huggingface.co/sb3")
+st.write("In game theory and optimization Nash Equilibrium loss minimization starts playing randomly but then by understanding ratios of action success to action-reward with an action (observe, decide/predict, act and then observe outcome the Deep RL agents go from 50% efficiency to 98-99% efficiency based on quality of decision without making mistakes. A good reference to environments is here https://github.com/DLR-RM/rl-baselines3-zoo/blob/master/benchmark.md")
+#st.write("Deep RL models: https://huggingface.co/sb3")
 
 env_name = st.selectbox(
     "Select environment",