frabob2017

> The most background images I've worked with was a 50/50 single-class dataset, which ended up with normal metrics. If you are trying a 10/90 dataset, you definitely need to...

How to use this function? I thought by doing data augmentation, you only change .yaml file [yolov5/data/hyps/hyp.scratch-low.yaml](https://github.com/ultralytics/yolov5/blob/11640698977724daf7982c9da398c2ee2f2b6e91/data/hyps/hyp.scratch-low.yaml#L25)

> @frabob2017 classify_albumentations function is in yolov5/utils/augmentations.py file. You can use it by changing the parameters from here. it is a great learning platform for Yolo algorithm.

> Use low for nano or small models, medium for medium, and high for larger models. Great to learn.

> Are you talking about 3d detection? Sorry that I did not formulate my question more clearly. Since lesions in medical imaging are usually continuously multiple slices. I found that...

May I ask if this task is difficult?

> I cannot speak about nn.conv3d. Given the confidence is so low, I wonder about the quality of and number of samples used in your dataset for training. Additionally, are...

No one can further answer this question?

It looks that I need to install "Microsoft C++ Build Tools" https://github.com/imartinez/privateGPT/issues/445

I have not figured it out yet. I also use Windows.