import osimport numpy as npimport cv2from skimagefeature import local_binary_patternfrom sklearndecomposition import PCAimport mindsporenn as nnimport mindsporedataset as dsimport mindsporedatasetvisi
可以使用mindspore.dataset.vision.py_transforms中的ToPIL()函数将numpy数组转换为PIL image，如下所示：
from mindspore.dataset.vision.py_transforms import ToPIL
prepare transforms
resize = transforms.Resize((100, 100)) to_tensor = transforms.ToTensor() to_pil = ToPIL() dataset = dataset.map(input_columns='data', operations=resize) dataset = dataset.map(input_columns='data', operations=to_pil) dataset = dataset.map(input_columns='data', operations=to_tensor)
原文地址: https://www.cveoy.top/t/topic/bh9y 著作权归作者所有。请勿转载和采集!