PyTorchEMD/test_emd_loss.py

46 lines
1.2 KiB
Python
Raw Normal View History

2019-09-10 08:05:04 +00:00
import numpy as np
2023-04-24 09:45:30 +00:00
import torch
2019-09-10 08:05:04 +00:00
from emd import earth_mover_distance
# gt
p1 = torch.from_numpy(np.array([[[1.7, -0.1, 0.1], [0.1, 1.2, 0.3]]], dtype=np.float32)).cuda()
p1 = p1.repeat(3, 1, 1)
p2 = torch.from_numpy(np.array([[[0.3, 1.8, 0.2], [1.2, -0.2, 0.3]]], dtype=np.float32)).cuda()
p2 = p2.repeat(3, 1, 1)
print(p1)
print(p2)
p1.requires_grad = True
p2.requires_grad = True
2023-04-24 09:45:30 +00:00
gt_dist = (
(((p1[0, 0] - p2[0, 1]) ** 2).sum() + ((p1[0, 1] - p2[0, 0]) ** 2).sum()) / 2
+ (((p1[1, 0] - p2[1, 1]) ** 2).sum() + ((p1[1, 1] - p2[1, 0]) ** 2).sum()) * 2
+ (((p1[2, 0] - p2[2, 1]) ** 2).sum() + ((p1[2, 1] - p2[2, 0]) ** 2).sum()) / 3
)
print("gt_dist: ", gt_dist)
2019-09-10 08:05:04 +00:00
gt_dist.backward()
print(p1.grad)
print(p2.grad)
# emd
p1 = torch.from_numpy(np.array([[[1.7, -0.1, 0.1], [0.1, 1.2, 0.3]]], dtype=np.float32)).cuda()
p1 = p1.repeat(3, 1, 1)
p2 = torch.from_numpy(np.array([[[0.3, 1.8, 0.2], [1.2, -0.2, 0.3]]], dtype=np.float32)).cuda()
p2 = p2.repeat(3, 1, 1)
print(p1)
print(p2)
p1.requires_grad = True
p2.requires_grad = True
d = earth_mover_distance(p1, p2, transpose=False)
print(d)
loss = d[0] / 2 + d[1] * 2 + d[2] / 3
print(loss)
loss.backward()
print(p1.grad)
print(p2.grad)