reward ajustments

robot starts balancing without changing phase and farms alive bonus
fix ->external force and bigger height penalty

ml/evaluate.py

@@ -44,17 +44,16 @@ def evaluate(
 
     for ep in range(episodes):
         obs, _ = env.reset()
-        done = False
+        terminated = False
         total_reward = 0.0
         steps = 0
 
         print(f"\n--- Starting Evaluation Episode {ep + 1}/{episodes} ---")
 
-        while not done:
+        while not terminated:
             action, _ = model.predict(obs, deterministic=True)
             obs, reward, terminated, truncated, _ = env.step(action)
             
-            done = terminated or truncated
             total_reward += float(reward)
             steps += 1