xuebinqin
Dataset for Segmenting Humans
I am using your great work for background removal tasks, see https://huggingface.co/schirrmacher/ormbg.
Here is my dataset: https://huggingface.co/datasets/schirrmacher/humans.
This dataset was created for developing a fully open-source background remover of images with humans. It was crafted with LayerDiffuse, a Stable Diffusion extension for generating transparent images.
You can use it freely. If you have any recommendations for the training process of your architecture, the community appreciates it!
At the moment I am playing around with various augmentation settings. If you have some ideas how to tackle issues people raise here, I would be very happy :)
Here the model is hosted: https://huggingface.co/spaces/schirrmacher/ormbg
It was even featured (although I have to improve it for many edge cases)
Here the model is hosted: https://huggingface.co/spaces/schirrmacher/ormbg这里托管了模型： https://huggingface.co/spaces/schirrmacher/ormbg
How to correctly convert to NCNN model? After my conversion to NCNN model, I can get output layers out0 to out5, but the MASK obtained is incorrect.
