The following code attempts to handle potential exceptions during Gemini Nano model inference. What is wrong with this approach?
Android interview question for Advanced practice.
Answer
The code lacks resource release in the finally block.
Explanation
The primary issue is the lack of resource cleanup in a finally block. Even if an exception occurs, the model object needs to be explicitly released using model.close() to free up resources. While catching specific exceptions (option A) is good practice, it's not the most critical error in this snippet. Option C is also a valid point for improvement, but less crucial than the resource leak.