Hello and thank you for this great Fast-nnUNet project.
I am a PhD student in a lab where we develop open source softwares (i.e. axondeepseg, spinalcordtoolbox). A few years ago we started packaging nnunetv2 models for segmentation tasks, but inference speed is always a concern for us, as most of our users don't have GPUs.
This is why your project is of great interest to us. I understand Fast-nnUNet is partially patent-protected. Is it possible to use your distillation module and distribute the lighter models in open-source softwares without commiting patent infringement?
Hello and thank you for this great Fast-nnUNet project.
I am a PhD student in a lab where we develop open source softwares (i.e. axondeepseg, spinalcordtoolbox). A few years ago we started packaging nnunetv2 models for segmentation tasks, but inference speed is always a concern for us, as most of our users don't have GPUs.
This is why your project is of great interest to us. I understand Fast-nnUNet is partially patent-protected. Is it possible to use your distillation module and distribute the lighter models in open-source softwares without commiting patent infringement?