Dmitry Kolesnikov

@Burhan-Q @glenn-jocher Thank you very much for your help. I have just two small questions left. What should we do next or just wait for the PR to be merged...

Unfortunately, at the moment, only classic inference through ultralytics models is supported (pytorch), but there is such a wide variety of models that you can choose any with high inference...

@mohammdkaraca Yes, I understand your question. The classic YOLO can achieve around 30-50 fps depending on the chosen architecture and the hardware of the computer on which the calculations are...

@mohammdkaraca Here's an example of how you can perform inference on a neural network using a library. If you specifically want to use a patch-based approach for inference rather than...

@mohammdkaraca For the most part, this function allows you to flexibly configure the final visualization of the results of the neural network inference. I even have a video tutorial on...

I advise you to put ```model = YOLO("yolov8m.pt")``` And you can experiment with the parameters imgsz, conf, iou to increase the quality for your unique task. > thanks alot at...

> its detecting some objects multiple times instead of once is this a normal behaviour? Decrease iou parameter (iou=0.3). Does it help?

@hanbangzou How are your results? Did it help your problem? PS: All usage examples are provided in [Google Colab](https://colab.research.google.com/drive/1eM4o1e0AUQrS1mLDpcgK9HKInWEvnaMn?usp=sharing)

@mohammdkaraca Important information: The latest update to the library has added the capability to input any converted ultralytics detection model into TensorRT, which further increases fps by 1.5 times. ```python...

This library allows you to work with all ultralytics models, so yolov5 is no exception. Here's an example of how to load this model (https://docs.ultralytics.com/models/yolov5/#supported-tasks-and-modes): ```python from ultralytics import YOLO...