refactor(ppo): optimize early stopping with for-else loop

   - Replace dual break checks with Pythonic for-else structure
   - Improve code readability while maintaining original logic
   - Related to issue haosulab#830

Author: songyuc

examples/baselines/ppo/ppo.py

@@ -436,9 +436,9 @@ def clip_action(action: torch.Tensor):
                 loss.backward()
                 nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
                 optimizer.step()
-
-            if args.target_kl is not None and approx_kl > args.target_kl:
-                break
+            else:
+                continue
+            break
 
         update_time = time.time() - update_time