enhance model loading logic in BaseModelPool and CLIPVisionModelPool for better configuration handling

[tanganke]

This pull request should fix issue #149

[Copilot]

Pull Request Overview

This pull request enhances the model loading logic in the BaseModelPool and CLIPVisionModelPool classes to improve configuration handling and error reporting. The changes address issue #149 by providing more robust model loading mechanisms.

• Enhanced the load_model method in both base and CLIP vision model pools to handle different configuration types more systematically
• Replaced conditional if-else chains with pattern matching for better type handling and code readability
• Added comprehensive error handling and logging for better debugging experience

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
fusion_bench/modelpool/base_pool.py	Refactored load_model method to use pattern matching and improved error handling for different model configuration types
fusion_bench/modelpool/clip_vision/modelpool.py	Enhanced load_model method with pattern matching, better documentation, and improved logging messages
fusion_bench/method/opcm/opcm.py	Added assertion to ensure model loading doesn't return None

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

[nitpick] Using str() as a pattern in match statements is unusual syntax. Consider using case model_path if isinstance(model_path, str): or simply case str(model_path): for better readability and consistency.

Suggested change

	case str() as model_path:
	case model_path if isinstance(model_path, str):

[Copilot]

[nitpick] Similar to the string pattern, nn.Module() as a pattern is unconventional. Consider using case model if isinstance(model, nn.Module): for better code clarity.

Suggested change

	case nn.Module() as model:
	case model if isinstance(model, nn.Module):