libcpab
libcpab copied to clipboard
CPAB Transformations: finite-dimensional spaces of simple, fast, and highly-expressive diffeomorphisms derived from parametric, continuously-defined, velocity fields in Numpy, Tensorflow and Pytorch
While running example of libcpab: "alignment_demo.py "
Line 61: Tensorflow
While running example of DTan [](https://github.com/BGU-CS-VIL/dtan/tree/master/examples). UCR_alignment.py
I would like to apply different deformation into each input channel, is it possible to do it automatically in the library?
As of my best understanding of the things, there was an error in interpolate2D. The indices for the gather_nd operation generated incorrectly. As a result of that, the transformation on...
Hi. I am YJHong and thanks to your nice work. What I gonna ask is that getting a kind of mapping function after CPAB transformation. For example, I want to...
Hi, As I understand at this moment the library doesn't support mixed-precision calculation. What steps should be done to add it? Thanks
error with gesv function in torch_funcs.py using pytorch 1.2.0 fix: #R, _ = torch.gesv(P, Q) # solve P = Q*R R, _ = torch.solve(P, Q)
Hi, What is the latest stable version that include 3D transformations? I see that in version 1.4 it is not working well. Thanks, Ilya