亲宝软件园·资讯

展开

PyTorch 中的 torch.utils.data 解析(推荐)

想变厉害的大白菜 人气:0

PyTorch 中的 torch.utils.data 解析

PyTorch 中的 torch.utils.data 解析

在 PyTorch 中,提供了一个处理数据集的工具包 torch.utils.data。这里来简单介绍这个包的结构。以下内容翻译和整理自 PyTorch 官方文档

概述

PyTorch 数据集处理包 torch.utils.data 的核心是 DataLoader 类。该类的构造函数签名为

DataLoader(dataset, batch_size=1, shuffle=False, sampler=None,
           batch_sampler=None, num_workers=0, collate_fn=None,
           pin_memory=False, drop_last=False, timeout=0,
           worker_init_fn=None, *, prefetch_factor=2,
           persistent_workers=False)

它构造一个 可迭代对象 loader,代表经过 “加工” 后的数据集。所谓的 “加工” 过程,是由构造函数参数表指定的,它包括:

一旦构造了 DataLoader 对象 loader,就可以用

for data in loader:
    # data 是数据集中的一组数据,且已转换成 Tensor

来加载数据。缺省情况下,PyTorch 会对数据进行 auto-batching,此时 data 对应一个 batch 的数据。

可以将 loader 理解成一个 生成器,其定义按情况可分为:(出现一些概念之后都会解释)

## 启用 auto-batching

# 对 map-style 数据集
for indices in batch_sampler:
    yield collate_fn([dataset[i] for i in indices])
# 对 iterable-style 数据集
dataset_iter = iter(dataset)
for indices in batch_sampler:
    yield collate_fn([next(dataset_iter) for _ in indices])

## 不启用 auto-batching (设置 batch_size=None 和 batch_sampler=None)

# 对 map-style 数据集
for index in sampler:
    yield collate_fn(dataset[index])
# 对 iterable-style 数据集
for data in iter(dataset):
    yield collate_fn(data)

数据集

DataLoader 构造函数中的必需参数 dataset 代表一个数据集。数据集主要分为两种:

数据加载顺序

手动定义 sampler

可以通过指定 sampler 参数来手动设置加载顺序。一个 sampler 是可迭代对象,其迭代的每一个值表示下一个待加载数据的 key/index。它应当实例化泛型类 torch.utils.data.Sampler[int] 的一个子类,并且重载 __iter____len__ 函数,具体地讲:

构造函数 __init__(self, data_source, *args) 必须提供一个重载了 __len__ 的数据集 data_source 作为参数__iter__ 返回一个整型迭代器,其每迭代一次的返回值为下一个待加载数据的 key/index__len__ 返回要加载的数据总数

但是要注意,只有 map-style 数据集才可定义 sampler,因为 iterable-style 不一定支持随机访问。

使用内置 sampler

在模块 torch.utils.data.sampler 中定义了一些内置的 sampler,通常来说已经够用了。在缺省 sampler 参数的情况下,如果指定参数 shuffle=False 将使用 SequentialSampler,即按顺序加载整个数据集;如果指定 shuffle=True 则使用 RandomSampler,即随机打乱数据后加载整个数据集。但是注意,不允许同时指定 sampler 参数和 shuffle 参数。

另外一些 sampler 可以参见模块源代码。

数据的 batching

在训练神经网络的时候经常需要将数据分成 mini-batch。PyTorch 本身提供了 auto-batching 的功能,也可以通过修改参数 batch_sizebatch_samplerdrop_lastcollate_fn 进行自定义 batching。

使用 batch sampler

在定义了 batching 后,PyTorch 会一次性输入多个(数量为 batch_size)数据。这时候需使用 batch sampler 来取代普通的 sampler。

通过指定 batch_sampler 参数,可以手动实现想要的 batch sampler。一个 batch_samplertorch.utils.data.BatchSampler 的实例。在 PyTorch 源代码中,该类继承了 Sampler[List[int]],并且封装了一个 sampler

请注意,

修改 collate_fn

参数 collate_fn 指定如何对每一 batch 的数据做预处理。在模块 torch.utils.data._utils 中,定义了两个默认的 collate_fn

扩展:PyTorch torch.utils.data.Dataset 介绍与实战案例

一、前言

训练模型一般都是先处理 数据的输入问题预处理问题 。Pytorch提供了几个有用的工具:torch.utils.data.Dataset 类和 torch.utils.data.DataLoader 类 。

流程是先把原始数据转变成 torch.utils.data.Dataset 类,随后再把得到的 torch.utils.data.Dataset 类当作一个参数传递给 torch.utils.data.DataLoader 类,得到一个数据加载器,这个数据加载器每次可以返回一个 Batch 的数据供模型训练使用。

在 pytorch 中,提供了一种十分方便的数据读取机制,即使用 torch.utils.data.DatasetDataloader 组合得到数据迭代器。在每次训练时,利用这个迭代器输出每一个 batch 数据,并能在输出时对数据进行相应的预处理或数据增广操作。

本文我们主要介绍对 torch.utils.data.Dataset 的理解,对 Dataloader 的介绍请参考我的另一篇文章:【PyTorch】torch.utils.data.DataLoader 简单介绍与使用

在本文的最后将给出 torch.utils.data.DatasetDataloader 结合使用处理数据的实战代码。

二、torch.utils.data.Dataset 是什么

1. 干什么用的?

2. 长什么样子?

torch.utils.data.Dataset 的源码:

class Dataset(object):
    """An abstract class representing a Dataset.

    All other datasets should subclass it. All subclasses should override
    ``__len__``, that provides the size of the dataset, and ``__getitem__``,
    supporting integer indexing in range from 0 to len(self) exclusive.
    """

    def __getitem__(self, index):
        raise NotImplementedError

    def __len__(self):
        raise NotImplementedError

    def __add__(self, other):
        return ConcatDataset([self, other])

注释翻译:

表示一个数据集的抽象类。

所有其他数据集都应该对其进行子类化。 所有子类都应该重写提供数据集大小的 __len____getitem__ ,支持从 0 到 len(self) 独占的整数索引。

理解:

就是说,Dataset 是一个 数据集 抽象类,它是其他所有数据集类的父类(所有其他数据集类都应该继承它),继承时需要重写方法 __len____getitem____len__ 是提供数据集大小的方法, __getitem__ 是可以通过索引号找到数据的方法。

三、通过继承 torch.utils.data.Dataset 定义自己的数据集类

torch.utils.data.Dataset 是代表自定义数据集的抽象类,我们可以定义自己的数据类抽象这个类,只需要重写__len__和__getitem__这两个方法就可以。

要自定义自己的 Dataset 类,至少要重载两个方法:__len__, __getitem__

下面将简单实现一个返回 torch.Tensor 类型的数据集:

from torch.utils.data import Dataset
import torch

class TensorDataset(Dataset):
    # TensorDataset继承Dataset, 重载了__init__, __getitem__, __len__
    # 实现将一组Tensor数据对封装成Tensor数据集
    # 能够通过index得到数据集的数据,能够通过len,得到数据集大小

    def __init__(self, data_tensor, target_tensor):
        self.data_tensor = data_tensor
        self.target_tensor = target_tensor

    def __getitem__(self, index):
        return self.data_tensor[index], self.target_tensor[index]

    def __len__(self):
        return self.data_tensor.size(0)    # size(0) 返回当前张量维数的第一维

# 生成数据
data_tensor = torch.randn(4, 3)   # 4 行 3 列,服从正态分布的张量
print(data_tensor)
target_tensor = torch.rand(4)     # 4 个元素,服从均匀分布的张量
print(target_tensor)

# 将数据封装成 Dataset (用 TensorDataset 类)
tensor_dataset = TensorDataset(data_tensor, target_tensor)

# 可使用索引调用数据
print('tensor_data[0]: ', tensor_dataset[0])

# 可返回数据len
print('len os tensor_dataset: ', len(tensor_dataset))

输出结果:

tensor([[ 0.8618,  0.4644, -0.5929],
        [ 0.9566, -0.9067,  1.5781],
        [ 0.3943, -0.7775,  2.0366],
        [-1.2570, -0.3859, -0.3542]])
tensor([0.1363, 0.6545, 0.4345, 0.9928])
tensor_data[0]:  (tensor([ 0.8618,  0.4644, -0.5929]), tensor(0.1363))
len os tensor_dataset:  4

四、为什么要定义自己的数据集类?

因为我们可以通过定义自己的数据集类并重写该类上的方法 实现多种多样的(自定义的)数据读取方式

比如,我们重写 __init__ 实现用 pd.read_csv 读取 csv 文件:

from torch.utils.data import Dataset
import pandas as pd  # 这个包用来读取CSV数据

# 继承Dataset,定义自己的数据集类 mydataset
class mydataset(Dataset):
    def __init__(self, csv_file):   # self 参数必须,其他参数及其形式随程序需要而不同,比如(self,*inputs)
        self.csv_data = pd.read_csv(csv_file)
    def __len__(self):
        return len(self.csv_data)
    def __getitem__(self, idx):
        data = self.csv_data.values[idx]
        return data

data = mydataset('spambase.csv')
print(data[3])
print(len(data))

输出结果:

[0.000e+00 0.000e+00 0.000e+00 0.000e+00 6.300e-01 0.000e+00 3.100e-01
 6.300e-01 3.100e-01 6.300e-01 3.100e-01 3.100e-01 3.100e-01 0.000e+00
 0.000e+00 3.100e-01 0.000e+00 0.000e+00 3.180e+00 0.000e+00 3.100e-01
 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00
 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00
 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00
 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00 0.000e+00
 1.370e-01 0.000e+00 1.370e-01 0.000e+00 0.000e+00 3.537e+00 4.000e+01
 1.910e+02 1.000e+00]
4601

要点:

__init__ 方法里面进行 读取数据文件
__getitem__ 方法里支持通过下标访问数据。
__len__ 方法里返回自定义数据集的大小,方便后期遍历。

五、实战:torch.utils.data.Dataset + Dataloader 实现数据集读取和迭代

实例 1

数据集 spambase.csv 用的是 UCI 机器学习存储库里的垃圾邮件数据集,它一条数据有57个特征和1个标签。

import torch.utils.data as Data
import pandas as pd  # 这个包用来读取CSV数据
import torch


# 继承Dataset,定义自己的数据集类 mydataset
class mydataset(Data.Dataset):
    def __init__(self, csv_file):   # self 参数必须,其他参数及其形式随程序需要而不同,比如(self,*inputs)
        data_csv = pd.DataFrame(pd.read_csv(csv_file))   # 读数据
        self.csv_data = data_csv.drop(axis=1, columns='58', inplace=False)  # 删除最后一列标签
    def __len__(self):
        return len(self.csv_data)
    def __getitem__(self, idx):
        data = self.csv_data.values[idx]
        return data


data = mydataset('spambase.csv')
x = torch.tensor(data[:5])         # 前五个数据
y = torch.tensor([1, 1, 1, 1, 1])  # 标签


torch_dataset = Data.TensorDataset(x, y)  # 对给定的 tensor 数据,将他们包装成 dataset

loader = Data.DataLoader(
    # 从数据库中每次抽出batch size个样本
    dataset = torch_dataset,       # torch TensorDataset format
    batch_size = 2,                # mini batch size
    shuffle=True,                  # 要不要打乱数据 (打乱比较好)
    num_workers=2,                 # 多线程来读数据
)

def show_batch():
    for step, (batch_x, batch_y) in enumerate(loader):
        print("steop:{}, batch_x:{}, batch_y:{}".format(step, batch_x, batch_y))

show_batch()

输出结果:

steop:0, batch_x:tensor([[0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 6.3000e-01, 0.0000e+00,
         3.1000e-01, 6.3000e-01, 3.1000e-01, 6.3000e-01, 3.1000e-01, 3.1000e-01,
         3.1000e-01, 0.0000e+00, 0.0000e+00, 3.1000e-01, 0.0000e+00, 0.0000e+00,
         3.1800e+00, 0.0000e+00, 3.1000e-01, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 1.3500e-01, 0.0000e+00, 1.3500e-01, 0.0000e+00, 0.0000e+00,
         3.5370e+00, 4.0000e+01, 1.9100e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 6.3000e-01, 0.0000e+00,
         3.1000e-01, 6.3000e-01, 3.1000e-01, 6.3000e-01, 3.1000e-01, 3.1000e-01,
         3.1000e-01, 0.0000e+00, 0.0000e+00, 3.1000e-01, 0.0000e+00, 0.0000e+00,
         3.1800e+00, 0.0000e+00, 3.1000e-01, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 1.3700e-01, 0.0000e+00, 1.3700e-01, 0.0000e+00, 0.0000e+00,
         3.5370e+00, 4.0000e+01, 1.9100e+02]], dtype=torch.float64), batch_y:tensor([1, 1])
steop:1, batch_x:tensor([[2.1000e-01, 2.8000e-01, 5.0000e-01, 0.0000e+00, 1.4000e-01, 2.8000e-01,
         2.1000e-01, 7.0000e-02, 0.0000e+00, 9.4000e-01, 2.1000e-01, 7.9000e-01,
         6.5000e-01, 2.1000e-01, 1.4000e-01, 1.4000e-01, 7.0000e-02, 2.8000e-01,
         3.4700e+00, 0.0000e+00, 1.5900e+00, 0.0000e+00, 4.3000e-01, 4.3000e-01,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         7.0000e-02, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 1.3200e-01, 0.0000e+00, 3.7200e-01, 1.8000e-01, 4.8000e-02,
         5.1140e+00, 1.0100e+02, 1.0280e+03],
        [6.0000e-02, 0.0000e+00, 7.1000e-01, 0.0000e+00, 1.2300e+00, 1.9000e-01,
         1.9000e-01, 1.2000e-01, 6.4000e-01, 2.5000e-01, 3.8000e-01, 4.5000e-01,
         1.2000e-01, 0.0000e+00, 1.7500e+00, 6.0000e-02, 6.0000e-02, 1.0300e+00,
         1.3600e+00, 3.2000e-01, 5.1000e-01, 0.0000e+00, 1.1600e+00, 6.0000e-02,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 6.0000e-02, 0.0000e+00, 0.0000e+00,
         1.2000e-01, 0.0000e+00, 6.0000e-02, 6.0000e-02, 0.0000e+00, 0.0000e+00,
         1.0000e-02, 1.4300e-01, 0.0000e+00, 2.7600e-01, 1.8400e-01, 1.0000e-02,
         9.8210e+00, 4.8500e+02, 2.2590e+03]], dtype=torch.float64), batch_y:tensor([1, 1])
steop:2, batch_x:tensor([[  0.0000,   0.6400,   0.6400,   0.0000,   0.3200,   0.0000,   0.0000,
           0.0000,   0.0000,   0.0000,   0.0000,   0.6400,   0.0000,   0.0000,
           0.0000,   0.3200,   0.0000,   1.2900,   1.9300,   0.0000,   0.9600,
           0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,
           0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,
           0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,
           0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,   0.0000,
           0.0000,   0.0000,   0.7780,   0.0000,   0.0000,   3.7560,  61.0000,
         278.0000]], dtype=torch.float64), batch_y:tensor([1])

一共 5 条数据,batch_size 设为 2 ,则数据被分为三组,每组的数据量为:2,2,1。

实例 2:进阶

import torch.utils.data as Data
import pandas as pd  # 这个包用来读取CSV数据
import numpy as np

# 继承Dataset,定义自己的数据集类 mydataset
class mydataset(Data.Dataset):
    def __init__(self, csv_file):   # self 参数必须,其他参数及其形式随程序需要而不同,比如(self,*inputs)
        # 读取数据
        frame = pd.DataFrame(pd.read_csv('spambase.csv'))
        spam = frame[frame['58'] == 1]
        ham = frame[frame['58'] == 0]
        SpamNew = spam.drop(axis=1, columns='58', inplace=False)  # 删除第58列,inplace=False不改变原数据,返回一个新dataframe
        HamNew = ham.drop(axis=1, columns='58', inplace=False)
        # 数据
        self.csv_data = np.vstack([np.array(SpamNew), np.array(HamNew)])  # 将两个N维数组进行连接,形成X
        # 标签
        self.Label = np.array([1] * len(spam) + [0] * len(ham))  # 形成标签值列表y
    def __len__(self):
        return len(self.csv_data)
    def __getitem__(self, idx):
        data = self.csv_data[idx]
        label = self.Label[idx]
        return data, label


data = mydataset('spambase.csv')
print(len(data))

loader = Data.DataLoader(
    # 从数据库中每次抽出batch size个样本
    dataset = data,       # torch TensorDataset format
    batch_size = 460,                # mini batch size
    shuffle=True,                  # 要不要打乱数据 (打乱比较好)
    num_workers=2,                 # 多线程来读数据
)

def show_batch():
    for step, (batch_x, batch_y) in enumerate(loader):
        print("steop:{}, batch_x:{}, batch_y:{}".format(step, batch_x, batch_y))

show_batch()

输出结果:

4601
steop:0, batch_x:tensor([[0.0000e+00, 2.4600e+00, 0.0000e+00,  ..., 2.1420e+00, 1.0000e+01,
         7.5000e+01],
        [0.0000e+00, 0.0000e+00, 1.6000e+00,  ..., 2.0650e+00, 1.2000e+01,
         9.5000e+01],
        [0.0000e+00, 0.0000e+00, 3.6000e-01,  ..., 3.7220e+00, 2.0000e+01,
         2.6800e+02],
        ...,
        [7.7000e-01, 3.8000e-01, 7.7000e-01,  ..., 1.4619e+01, 5.2500e+02,
         9.2100e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0000e+00, 1.0000e+00,
         5.0000e+00],
        [4.0000e-01, 1.8000e-01, 3.2000e-01,  ..., 3.3050e+00, 1.8100e+02,
         1.6130e+03]], dtype=torch.float64), batch_y:tensor([0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1,
        0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0,
        0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0,
        1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
        0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0,
        0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1,
        1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0,
        0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0,
        1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1,
        0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1,
        1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0,
        0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0,
        0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1,
        0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0,
        1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0,
        0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1,
        1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1,
        0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1,
        0, 1, 0, 1])
steop:1, batch_x:tensor([[0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0000e+00, 1.0000e+00,
         2.0000e+00],
        [4.9000e-01, 0.0000e+00, 7.4000e-01,  ..., 3.9750e+00, 4.7000e+01,
         4.8500e+02],
        [0.0000e+00, 0.0000e+00, 7.1000e-01,  ..., 4.0220e+00, 9.7000e+01,
         5.4300e+02],
        ...,
        [0.0000e+00, 1.4000e-01, 1.4000e-01,  ..., 5.3310e+00, 8.0000e+01,
         1.0290e+03],
        [0.0000e+00, 0.0000e+00, 3.6000e-01,  ..., 3.1760e+00, 5.1000e+01,
         2.7000e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.1660e+00, 2.0000e+00,
         7.0000e+00]], dtype=torch.float64), batch_y:tensor([0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
        1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0,
        0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 1, 0,
        1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 0, 0,
        1, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0,
        0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0,
        1, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0,
        0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0,
        1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1,
        1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 1,
        0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0,
        0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1,
        0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 0,
        0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0,
        0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1,
        1, 1, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0, 1,
        1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1,
        0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0,
        0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1,
        1, 0, 0, 0])
steop:2, batch_x:tensor([[0.0000e+00, 0.0000e+00, 1.4700e+00,  ..., 3.0000e+00, 3.3000e+01,
         1.7700e+02],
        [2.6000e-01, 4.6000e-01, 9.9000e-01,  ..., 1.3235e+01, 2.7200e+02,
         1.5750e+03],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.0450e+00, 6.0000e+00,
         4.5000e+01],
        ...,
        [4.0000e-01, 0.0000e+00, 0.0000e+00,  ..., 1.1940e+00, 5.0000e+00,
         1.2900e+02],
        [2.6000e-01, 0.0000e+00, 0.0000e+00,  ..., 1.8370e+00, 1.1000e+01,
         1.5800e+02],
        [5.0000e-02, 0.0000e+00, 1.0000e-01,  ..., 3.7150e+00, 1.0700e+02,
         1.3860e+03]], dtype=torch.float64), batch_y:tensor([1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0,
        0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0,
        1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0,
        0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0,
        0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0,
        0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
        0, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0,
        0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0,
        0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1,
        0, 1, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0,
        1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0,
        0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0,
        0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0,
        1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0,
        1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1,
        0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0,
        0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1,
        1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0,
        1, 1, 0, 0])
steop:3, batch_x:tensor([[2.6000e-01, 0.0000e+00, 5.3000e-01,  ..., 2.6460e+00, 7.7000e+01,
         1.7200e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.4280e+00, 5.0000e+00,
         1.7000e+01],
        [3.4000e-01, 0.0000e+00, 1.7000e+00,  ..., 6.6700e+02, 1.3330e+03,
         1.3340e+03],
        ...,
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0000e+00, 1.0000e+00,
         7.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.7010e+00, 2.0000e+01,
         1.8100e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 4.0000e+00, 1.1000e+01,
         3.6000e+01]], dtype=torch.float64), batch_y:tensor([0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0,
        1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 1,
        0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0,
        1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0,
        0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
        0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0,
        1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0,
        1, 1, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0,
        0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0,
        0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1,
        0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0,
        0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0,
        0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1,
        0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0,
        0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1,
        1, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0,
        1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 0,
        1, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0,
        1, 0, 0, 1])
steop:4, batch_x:tensor([[  0.0000,   0.0000,   0.3100,  ...,   5.7080, 138.0000, 274.0000],
        [  0.0000,   0.0000,   0.3400,  ...,   2.2570,  17.0000, 158.0000],
        [  1.0400,   0.0000,   0.0000,  ...,   1.0000,   1.0000,  17.0000],
        ...,
        [  0.0000,   0.0000,   0.0000,  ...,   4.0000,  12.0000,  28.0000],
        [  0.3300,   0.0000,   0.0000,  ...,   1.7880,   6.0000,  93.0000],
        [  0.0000,  14.2800,   0.0000,  ...,   1.8000,   5.0000,   9.0000]],
       dtype=torch.float64), batch_y:tensor([1, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1,
        0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1,
        0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 0,
        1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1,
        0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0,
        1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0,
        0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0,
        0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1,
        0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0,
        1, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1,
        1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0,
        0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0,
        1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1,
        0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
        0, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1,
        1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0,
        0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1,
        0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1,
        1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0,
        1, 1, 0, 0])
steop:5, batch_x:tensor([[7.0000e-01, 0.0000e+00, 1.0500e+00,  ..., 1.1660e+00, 1.3000e+01,
         1.8900e+02],
        [0.0000e+00, 3.3600e+00, 1.9200e+00,  ..., 6.1370e+00, 1.0700e+02,
         1.7800e+02],
        [5.4000e-01, 0.0000e+00, 1.0800e+00,  ..., 5.4540e+00, 6.8000e+01,
         1.8000e+02],
        ...,
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 3.8330e+00, 9.0000e+00,
         2.3000e+01],
        [6.0000e-02, 6.5000e-01, 7.1000e-01,  ..., 4.7420e+00, 1.1700e+02,
         1.3420e+03],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.6110e+00, 1.2000e+01,
         4.7000e+01]], dtype=torch.float64), batch_y:tensor([1, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 1,
        1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0,
        0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,
        0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0,
        0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 1,
        0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1,
        0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0,
        0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0,
        0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1,
        1, 0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1,
        0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1,
        1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1,
        0, 0, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1,
        0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 0, 1, 0,
        0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 1,
        0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1,
        0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0,
        1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 0,
        0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
        0, 1, 1, 1])
steop:6, batch_x:tensor([[0.0000e+00, 1.4280e+01, 0.0000e+00,  ..., 1.8000e+00, 5.0000e+00,
         9.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.9280e+00, 1.5000e+01,
         5.4000e+01],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0692e+01, 6.5000e+01,
         1.3900e+02],
        ...,
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.5000e+00, 5.0000e+00,
         2.4000e+01],
        [7.6000e-01, 1.9000e-01, 3.8000e-01,  ..., 3.7020e+00, 4.5000e+01,
         1.0700e+03],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.0000e+00, 1.2000e+01,
         8.8000e+01]], dtype=torch.float64), batch_y:tensor([0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1,
        0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1,
        0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0,
        1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1,
        1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0,
        0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1,
        0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0,
        0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
        1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 0,
        0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1,
        0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
        1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0,
        0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
        1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 1,
        0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0,
        0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1,
        1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,
        1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1,
        1, 0, 1, 0])
steop:7, batch_x:tensor([[0.0000e+00, 2.7000e-01, 0.0000e+00,  ..., 5.8020e+00, 4.3000e+01,
         4.1200e+02],
        [0.0000e+00, 3.5000e-01, 7.0000e-01,  ..., 3.6390e+00, 6.1000e+01,
         3.1300e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.5920e+00, 7.0000e+00,
         1.2900e+02],
        ...,
        [8.0000e-02, 1.6000e-01, 8.0000e-02,  ..., 2.7470e+00, 8.6000e+01,
         1.9950e+03],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.6130e+00, 1.1000e+01,
         7.1000e+01],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.9110e+00, 1.5000e+01,
         6.5000e+01]], dtype=torch.float64), batch_y:tensor([0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0,
        0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0,
        1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1,
        0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1,
        0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1,
        0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0,
        0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0,
        0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0,
        1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1,
        1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0,
        0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
        0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 1,
        0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0,
        0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0,
        0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 0, 1, 0,
        1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1,
        0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1,
        0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1,
        1, 0, 0, 0])
steop:8, batch_x:tensor([[1.7000e-01, 0.0000e+00, 1.7000e-01,  ..., 1.7960e+00, 1.2000e+01,
         4.5800e+02],
        [3.7000e-01, 0.0000e+00, 6.3000e-01,  ..., 1.1810e+00, 4.0000e+00,
         1.0400e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0000e+00, 1.0000e+00,
         7.0000e+00],
        ...,
        [2.3000e-01, 0.0000e+00, 4.7000e-01,  ..., 2.4200e+00, 1.2000e+01,
         3.3400e+02],
        [0.0000e+00, 0.0000e+00, 1.2900e+00,  ..., 1.3500e+00, 4.0000e+00,
         2.7000e+01],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.3730e+00, 1.1000e+01,
         1.6900e+02]], dtype=torch.float64), batch_y:tensor([1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1,
        0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0,
        1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 0,
        0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1,
        1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0,
        0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0,
        0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0,
        0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 1,
        0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
        1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1,
        0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 0,
        1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0,
        0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0,
        1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0,
        0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1,
        1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0,
        1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0,
        0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0,
        1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1,
        0, 0, 0, 0])
steop:9, batch_x:tensor([[0.0000e+00, 6.3000e-01, 0.0000e+00,  ..., 2.2150e+00, 2.2000e+01,
         1.1300e+02],
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 1.0000e+00, 1.0000e+00,
         5.0000e+00],
        [0.0000e+00, 0.0000e+00, 2.0000e-01,  ..., 1.1870e+00, 1.1000e+01,
         1.1400e+02],
        ...,
        [0.0000e+00, 0.0000e+00, 0.0000e+00,  ..., 2.3070e+00, 1.6000e+01,
         3.0000e+01],
        [5.1000e-01, 4.3000e-01, 2.9000e-01,  ..., 6.5900e+00, 7.3900e+02,
         2.3330e+03],
        [6.8000e-01, 6.8000e-01, 6.8000e-01,  ..., 2.4720e+00, 9.0000e+00,
         8.9000e+01]], dtype=torch.float64), batch_y:tensor([0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0,
        0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0,
        0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0,
        0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 1,
        1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0,
        0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0,
        0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1,
        1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1,
        0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1,
        0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 1,
        1, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0,
        1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0,
        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0,
        0, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1,
        1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0,
        0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0,
        1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0,
        1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0,
        1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0,
        1, 1, 1, 1])
steop:10, batch_x:tensor([[0.0000e+00, 2.5000e-01, 7.5000e-01, 0.0000e+00, 1.0000e+00, 2.5000e-01,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 2.5000e-01, 2.5000e-01,
         1.2500e+00, 0.0000e+00, 0.0000e+00, 2.5000e-01, 0.0000e+00, 1.2500e+00,
         2.5100e+00, 0.0000e+00, 1.7500e+00, 0.0000e+00, 2.5000e-01, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 2.5000e-01, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00,
         0.0000e+00, 0.0000e+00, 0.0000e+00, 4.2000e-02, 0.0000e+00, 0.0000e+00,
         1.2040e+00, 7.0000e+00, 1.1800e+02]], dtype=torch.float64), batch_y:tensor([0])

一共 4601 条数据,按 batch_size = 460 来分:能划分为 11 组,前 10 组的数据量为 460,最后一组的数据量为 1 。

参考链接

  1. torch.Tensor.size()方法的使用举例
  2. Pytorch笔记05-自定义数据读取方式orch.utils.data.Dataset与Dataloader
  3. pytorch 可训练数据集创建(torch.utils.data)
  4. Pytorch的第一步:(1) Dataset类的使用
  5. pytorch中的torch.utils.data.Dataset和torch.utils.data.DataLoader

加载全部内容

相关教程
猜你喜欢
用户评论