Refactor dataset batch-size

[glenn-jocher]

Precursor to #9542

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Refactoring batch size assignment in image classification, object detection, and segmentation scripts.

📊 Key Changes

• Unified the batch size (bs) assignment for webcam and non-webcam sources in predict.py, detect.py, and segment/predict.py.
• Removed redundant batch size setting lines for non-webcam sources.

🎯 Purpose & Impact

• 🧹 Simplifies Code: Eliminates the need to set the batch size separately within conditional blocks, making the code more maintainable.
• ✅ Consistency Across Modes: Ensures batch size is consistently defined whether input is coming from a webcam or individual images.
• 👥 User Experience: Impacts the users by potentially providing more efficient processing and a cleaner, more understandable codebase.