Add best.pt PyTorch Hub inference to CI

[glenn-jocher]

Test for #8023

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhanced CI testing for custom model loading with additional test case.

📊 Key Changes

• Added a second model loading and inference test to the Continuous Integration (CI) workflow.

🎯 Purpose & Impact

• 🎯 Purpose: To ensure that not just one, but multiple custom model paths are correctly working when models are loaded via torch.hub.
• 💥 Impact: Increases the robustness of CI testing by verifying additional scenarios, which helps prevent future releases from breaking model loading functionality. This change benefits users by providing more reliable software updates.
  🔍 Users can expect thorough testing of model loading, leading to more stable releases.

[glenn-jocher]

Tests all green. No issues detected.