技術標籤：資料處理深度學習pytorch

from torch.utils.data import Dataset,DataLoader
from torchvision import datasets,transforms
import os
from PIL import Image
import numpy as np
import torch

class My_dataset(Dataset):
    def __init__(self,root_path,is_train=True,is_miniTrain=False):
        self.is_train=is_train
        super
 
().__init__()
        f=open(root_path,'r',encoding='utf-8')
        data_list=f.readlines()

        self.x=[]
        self.y=[]
        for i,data in enumerate(data_list):
            data=data.rstrip()
            self.x.append(data.split(',')[0])
            self.y.append(data.split(',')[1:])

        if
 
 is_miniTrain:
            self.x=self.x[:700]
            self.y=self.y[:700]

    def __len__(self):
        return len(self.x)

    def __getitem__(self, index):
        x=self.x[index]     #'./img/my_data/TRAIN/2586_paste.png'
        y=self.y[index]
        # img=Image.open('./img/my_data/TRAIN/2586_paste.png')
 

        # img.show()
        # exit()
        img=self.train_transform(Image.open(x)) if self.is_train \
            else self.others_transform(Image.open(x))       # #(224, 224, 3)

        lable=[]
        for i in y:
            lable.append(int(i))

        # lable=np.array(lable).reshape(5,-1)

        return img,lable

    def train_transform(self,x):
        return transforms.Compose([
            transforms.RandomCrop(224,padding=28),
            transforms.RandomRotation((0.5)),
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.485,0.456,0.486],std=[0.485,0.456,0.486])
        ])(x)

    def others_transform(self,x):
        return  transforms.Compose([
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.485,0.456,0.486],std=[0.485,0.456,0.486])
        ])(x)