import torch.utils.data as Data
from tqdm import tqdm
class MyDataset(Data.Dataset):
	def __init__(self,filepath):
		number = 0
		with open(filepath,"r") as f:
			# 獲得訓練數(shù)據(jù)的總行數(shù)
			for _ in tqdm(f,desc="load training dataset"):
				number+=1
		self.number = number
		self.fopen = open(filepath,'r')
	def __len__(self):
		return self.number
	def __getitem__(self,index):
		line = self.fopen.__next__()
		# 自定義transform()對訓練數(shù)據(jù)進行預處理
		data = transform(line)
		return data

train_dataset = MyDataset(filepath = "train.txt")
training_data = Data.DataLoader(dataset=train_dataset, batch_size=32,num_workers=1)

import torch
import numpy as np

# 定義GetLoader類，繼承Dataset方法，并重寫__getitem__()和__len__()方法
class GetLoader(torch.utils.data.Dataset):
	# 初始化函數(shù)，得到數(shù)據(jù)
    def __init__(self, data_root, data_label):
        self.data = data_root
        self.label = data_label
    # index是根據(jù)batchsize劃分數(shù)據(jù)后得到的索引，最后將data和對應的labels進行一起返回
    def __getitem__(self, index):
        data = self.data[index]
        labels = self.label[index]
        return data, labels
    # 該函數(shù)返回數(shù)據(jù)大小長度，目的是DataLoader方便劃分，如果不知道大小，DataLoader會一臉懵逼
    def __len__(self):
        return len(self.data)

# 隨機生成數(shù)據(jù)，大小為10 * 20列
source_data = np.random.rand(10, 20)
# 隨機生成標簽，大小為10 * 1列
source_label = np.random.randint(0,2,(10, 1))
# 通過GetLoader將數(shù)據(jù)進行加載，返回Dataset對象，包含data和labels
torch_data = GetLoader(source_data, source_label)

torch.utils.data.DataLoader(dataset,batch_size,shuffle,drop_last，num_workers)

from torch.utils.data import DataLoader

# 讀取數(shù)據(jù)
datas = DataLoader(torch_data, batch_size=6, shuffle=True, drop_last=False, num_workers=2)

for i, data in enumerate(datas):
 # i表示第幾個batch， data表示該batch對應的數(shù)據(jù)，包含data和對應的labels
    print("第 {} 個Batch \n{}".format(i, data))

# 表示輸出數(shù)據(jù)
print(data[0])
# 表示輸出標簽
print(data[1])