Prateek Gupta

@hammad2008 you should use HankAI's darknet/yolov4 instead of other versions for small object detection. Check this video- [https://www.youtube.com/watch?v=861LvUXvJmA&t=14s](url) The benefit of using darknet/yolov4 is smaller training time and better results.

@akraylite Share your training command or change value of copy_paste parameter in hyp.scratch-high.yaml file to 0.0 and then try.

@Raj2032 for object detection YOLO is one of the most accurate and near real time solution. You must try it.

@JacksonWang YOLO model resizes the images to it's network dimensions before starting the training. So no matter what image sizes you have, it will be resized to it's network size(...

@JacksonWang selection of models always depend on the dataset complexity. If you have very small objects (less than 16x16 pixels for example) to detect in your 600x600 images, you should...

@Harshchitaliya You must raise your issue in the ultralytics/yolov8 repo not here since it is YOLOv7 repo.

It seems author is busy with the MIT licensed YOLOV7 and v9 implementation. Please check this- [https://github.com/WongKinYiu/yolov9mit](url)

@fukingbus how have you converted the gelan-c segmentation model to the tflite format?

Thanks @fukingbus . It seems the repo [https://github.com/PINTO0309/onnx2tf](url) is the silver bullet for tf conversion. Author has done the great work.

I also faced the installation issue in kaggle/colab with or without GPU. Error logs- `Collecting autoxgb Downloading autoxgb-0.2.2-py3-none-any.whl.metadata (6.2 kB) Collecting fastapi==0.70.0 (from autoxgb) Downloading fastapi-0.70.0-py3-none-any.whl.metadata (23 kB) Collecting loguru==0.5.3...