fix: Resolve remaining numpy array comparison errors in MultiDongle

- Fix ambiguous truth value error in get_latest_inference_result method
- Fix ambiguous truth value error in postprocess function
- Replace direct array evaluation with explicit length checks
- Use proper None checks instead of truthy evaluation on numpy arrays

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

cluster4npu_ui/core/functions/Multidongle.py

@@ -426,7 +426,7 @@ class MultiDongle:
                     retrieval_successful = False
                     break
 
-            if retrieval_successful and inf_node_output_list:
+            if retrieval_successful and len(inf_node_output_list) > 0:
                 # Process output nodes
                 if output_descriptor.header.num_output_node == 1:
                     raw_output_array = inf_node_output_list[0].flatten()
@@ -690,7 +690,7 @@ def postprocess(raw_model_output: list) -> float:
     Post-processes the raw model output.
     Assumes the model output is a list/array where the first element is the desired probability.
     """
-    if raw_model_output and len(raw_model_output) > 0:
+    if raw_model_output is not None and len(raw_model_output) > 0:
         probability = raw_model_output[0]
         return float(probability)
     return 0.0 # Default or error value