ICVL_DATA

[XiaoZhang-NN]

您好，请问一下在，这里的2187和558和576以及4095是指什么呢？
高光谱图片切块训练，按照14414431.这个高光谱图片512*512应该只能有25块附近吧？self.HR = torch.zeros([num, channels, fis, fis])怎么batchsize会是2187 或者558？
def init(self,path,label,s=4,channels=31,fis=144):
# num 31 512 512
if label == 'train':
num = 2187
elif label == 'val':
num = 558
else:
num = 576
self.HR = torch.zeros([num, channels, fis, fis])

    count = 0
    for i in range(len(path)):

        img = h5py.File(path[i], 'r')['rad']
        img = np.array(img)
        img /= 4095.0