Spaces:
Paused
Paused
File size: 1,976 Bytes
05fcd0f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
import torch
import torch.nn as nn
# import devicetorch
# device = devicetorch.get(torch)
backwarp_tenGrid = {} # Cache for grid tensors
def warp(tenInput, tenFlow):
# The key for caching should be based on tenFlow's properties, including its device
k = (str(tenFlow.device), str(tenFlow.size()))
if k not in backwarp_tenGrid:
# Create grid tensors on the same device as tenFlow
flow_device = tenFlow.device
tenHorizontal = (
torch.linspace(-1.0, 1.0, tenFlow.shape[3], device=flow_device)
.view(1, 1, 1, tenFlow.shape[3])
.expand(tenFlow.shape[0], -1, tenFlow.shape[2], -1)
)
tenVertical = (
torch.linspace(-1.0, 1.0, tenFlow.shape[2], device=flow_device)
.view(1, 1, tenFlow.shape[2], 1)
.expand(tenFlow.shape[0], -1, -1, tenFlow.shape[3])
)
# Concatenate; the result will be on flow_device if inputs are.
# No need for an extra .to(device) if flow_device is used for components.
backwarp_tenGrid[k] = torch.cat([tenHorizontal, tenVertical], 1)
# tenFlow is already on its correct device.
# backwarp_tenGrid[k] is now also on that same device.
tenFlow = torch.cat(
[
tenFlow[:, 0:1, :, :] / ((tenInput.shape[3] - 1.0) / 2.0),
tenFlow[:, 1:2, :, :] / ((tenInput.shape[2] - 1.0) / 2.0),
],
1,
)
# Ensure grid used by grid_sample is on the same device as tenInput and tenFlow.
# backwarp_tenGrid[k] is already on tenFlow.device.
# tenFlow is on tenFlow.device.
# If tenInput can be on a different device than tenFlow, that's a separate issue.
# Assuming tenInput and tenFlow are on the same device for grid_sample.
g = (backwarp_tenGrid[k] + tenFlow).permute(0, 2, 3, 1)
return torch.nn.functional.grid_sample(
input=tenInput, grid=g, mode="bilinear", padding_mode="border", align_corners=True
) |