Gabri95
commented
2 years ago Hi @Coronal-Halo
Sorry for the late reply. I am trying to gradually move to the new version of the library escnn. Please, take a look at it!
Someone else requested this feature and I will soon integrate it via this pull request.
I will include this feature here too for the moment, but I am not generally planning to support many more features on this library, so I would recommend the users to move to the new one
Best, Gabriele
Currently the e2cnn.R2Upsampling class only has four parameters: in_type, scale_factor, mode and align_corners. However, I think it would be helpful to add another "size" parameter, to be passed to the torch.nn.functional.interpolate method used inside the R2Upsampling class. Sometimes we need to up-sample the input to a specific dimension, especially when dealing with even and odd dimension issues