比较与torchvision.transforms.Normalize的差异

torchvision.transforms.Normalize

class torchvision.transforms.Normalize(mean, std, inplace=False)

更多内容详见torchvision.transforms.Normalize。

mindspore.dataset.vision.Normalize

class mindspore.dataset.vision.Normalize(mean, std, is_hwc=True)

更多内容详见mindspore.dataset.vision.Normalize。

差异对比

PyTorch：根据均值和标准差对输入图像进行归一化，不支持指定图像的格式。

MindSpore：根据均值和标准差对输入图像进行归一化，不支持原地修改。

分类	子类	PyTorch	MindSpore	差异
参数	参数1	mean	mean	-
	参数2	std	std	-
	参数3	inplace	-	是否对Tensor进行原地修改
	参数4	-	is_hwc	指定图像的格式是否为HWC或CHW格式

代码示例

from download import download
from PIL import Image

url = "https://mindspore-website.obs.cn-north-4.myhuaweicloud.com/notebook/datasets/flamingos.jpg"
download(url, './flamingos.jpg', replace=True)
orig_img = Image.open('flamingos.jpg')

# PyTorch
import torchvision.transforms as T

normalize = T.Normalize(mean=[0, 0, 0], std=[1, 1, 1])
to_tensor = T.ToTensor()
img_torch = T.Compose([to_tensor, normalize])((orig_img))
print(img_torch.shape)
# Torch tensor is in C,H,W format
# Out: torch.Size([3, 292, 471])

# MindSpore
import mindspore.dataset.vision as vision
import mindspore.dataset.transforms as transforms

normalize = vision.Normalize(mean=[0, 0, 0], std=[1, 1, 1], is_hwc=False)
to_tensor = vision.ToTensor()
img_ms = transforms.Compose([to_tensor, normalize])((orig_img))
print(img_ms[0].shape)
# vision.ToTensor change the format from HWC to CHW, so normalize have to specify `is_hwc=False`
# Out: (3, 292, 471)