Streaming Classification support

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Expanded YOLOv5 classification inference to support new input sources and improved handling of results.

📊 Key Changes

• Added support for YouTube URLs, webcam and generic streams (RTSP, RTMP, HTTP) as input sources.
• Implemented various new options like view_img, save_txt, and update flags.
• Improved image processing to handle different sizes and batch processing.
• Streamlined result saving, enabling both image and text output.
• Introduced inference speed improvements through batch processing and cuDNN benchmarking.
• Added code to process and visualize top prediction results.
• Enhanced the command-line interface with more options for image size, saving, and model updates.
• Included checks and assertions to handle possible errors and misconfigurations.

🎯 Purpose & Impact

• 🎥 The new input capabilities allow users to process data from a wider variety of sources, ranging from live streams to online videos, making YOLOv5 more versatile.
• 💾 The additional flags and output options improve user control over the inference process and the handling of results, facilitating integration into different workflows and systems.
• ⚡️ Batch processing and cuDNN benchmarking can result in faster inference, especially for video or stream processing.
• 🖼️ Enhanced visualization and saving of results help users better interpret the model's predictions.
• ✅ The improved interface and error handling offer a smoother and more robust user experience, reducing potential frustration and troubleshooting time.

[glenn-jocher]

@AyushExel this adds full cls inference support for all sources and also image annotation, i.e.

Usage - sources:
    $ python classify/predict.py --weights yolov5s-cls.pt --source 0                               # webcam
                                                                   img.jpg                         # image
                                                                   vid.mp4                         # video
                                                                   path/                           # directory
                                                                   'path/*.jpg'                    # glob
                                                                   'https://youtu.be/Zgi9g1ksQHc'  # YouTube
                                                                   'rtsp:https://example.com/media.mp4'  # RTSP, RTMP, HTTP stream

[AyushExel]

PERFECT!!