213d64f52eb9f33f6d127e5033e529c6694d3c89,torchgeometry/homography_warper.py,,create_meshgrid,#Any#Any#Any#,14
Before Change
else:
xs = torch.linspace(0, width - 1, width)
ys = torch.linspace(0, height - 1, height)
return torch.stack(torch.meshgrid([ys, xs])).view(1, 2, -1)[:, (1, 0), :]
// layer api
After Change
def warp_grid(self, H):
:param H: Homography or homographies (stacked) to transform all points
in the grid.
:retur ns: Tensor[1, Height, Width, 2] containing transformed points in
normalized images space.
batch_size = H.shape[0] // expand grid to match the input batch size
In pattern: SUPERPATTERN
Frequency: 3
Non-data size: 4
Instances Project Name: arraiy/torchgeometry
Commit Name: 213d64f52eb9f33f6d127e5033e529c6694d3c89
Time: 2018-10-07
Author: edgar.riba@gmail.com
File Name: torchgeometry/homography_warper.py
Class Name:
Method Name: create_meshgrid
Project Name: maciejkula/spotlight
Commit Name: 70e4d7fe60a9658bb27b9f5fb67592a1222b2ec3
Time: 2017-07-06
Author: maciej.kula@gmail.com
File Name: spotlight/sequence/representations.py
Class Name: CNNNet
Method Name: user_representation
Project Name: maciejkula/spotlight
Commit Name: 70e4d7fe60a9658bb27b9f5fb67592a1222b2ec3
Time: 2017-07-06
Author: maciej.kula@gmail.com
File Name: spotlight/sequence/representations.py
Class Name: PoolNet
Method Name: user_representation