Closed tvercaut closed 5 years ago
This also extends to making recommendations for the licence of any linked model / weight file.
For simplicity, I would probably go for the same licence as NiftyNet for the repo / models and a cc-by or cc0 for the weights.
I't unclear however how we should keep track of the licence in each of the models and weights.
See also https://github.com/NifTK/NiftyNet/issues/180
Also note that license for pre-trained weights might be slightly challenging as it might depend on the license of the original data use to obtain the weights (derived work?)
This also extends to making recommendations for the licence of any linked model / weight file.
For simplicity, I would probably go for the same licence as NiftyNet for the repo / models and a cc-by or cc0 for the weights.
I't unclear however how we should keep track of the licence in each of the models and weights.