如何在PyTorch中显示单个图像？

我想显示一个图像。它使用a加载ImageLoader并存储在PyTorch中Tensor。

当我尝试通过显示它时plt.imshow(image)，我得到：

TypeError: Invalid dimensions for image data

该.shape张量是：

torch.Size([3, 244, 244])

如何显示此PyTorch张量中包含的图像？

给定一个Tensor代表图像的图像，请使用.permute()：

plt.imshow(  tensor_image.permute(1, 2, 0)  )

注意：permute不会复制或分配内存，也不会。 from_numpy()

如您所见，matplotlib即使不转换为numpy数组也可以正常工作。但是PyTorch张量（“图像张量”）是第一个通道，因此要与它们一起使用，matplotlib您需要对其进行重塑：

码：

from scipy.misc import face
import matplotlib.pyplot as plt
import torch

np_image = face()
print(type(np_image), np_image.shape)
tensor_image = torch.from_numpy(np_image)
print(type(tensor_image), tensor_image.shape)
# reshape to channel first:
tensor_image = tensor_image.view(tensor_image.shape[2], tensor_image.shape[0], tensor_image.shape[1])
print(type(tensor_image), tensor_image.shape)

# If you try to plot image with shape (C, H, W)
# You will get TypeError:
# plt.imshow(tensor_image)

# So we need to reshape it to (H, W, C):
tensor_image = tensor_image.view(tensor_image.shape[1], tensor_image.shape[2], tensor_image.shape[0])
print(type(tensor_image), tensor_image.shape)

plt.imshow(tensor_image)
plt.show()

输出：

 (768, 1024, 3)
 torch.Size([768, 1024, 3])
 torch.Size([3, 768, 1024])
 torch.Size([768, 1024, 3])