PyTorch源码解读之torch.utils.data.DataLoader
阅读原文时间:2021年04月20日阅读:1

1. torch.utils.data.DataLoader类:

class torch.utils.data.DataLoader(dataset, batch_size=1, shuffle=False, sampler=None, batch_sampler=None, num_workers=0, collate_fn=, pin_memory=False, drop_last=False, timeout=0, worker_init_fn=None)
  • 作用:是加载数据的核心,返回可迭代的数据。
  • PyTorch中数据读取的一个重要接口torch.utils.data.DataLoader,该接口定义在dataloader.py脚本中,只要是用PyTorch来训练模型基本都会用到该接口。
  • 该接口主要用来将自定义的数据读取接口的输出或者PyTorch已有的数据读取接口的输入按照batch size封装成Tensor,后续只需要再包装成V;ariable即可作为模型的输入,因此该接口有点承上启下的作用,比较重要。

  • 参数:

    • dataset (Dataset): 加载数据的数据集
    • batch_size (int, optional): 每批加载多少个样本
    • shuffle (bool, optional): 设置为“真”时,在每个epoch对数据打乱.(默认:False)
    • sampler (Sampler, optional): 定义从数据集中提取样本的策略,返回一个样本
    • batch_sampler (Sampler, optional): like sampler, but returns a batch of indices at a time 返回一批样本. 与atch_size, shuffle, sampler和 drop_last互斥.
    • num_workers (int, optional): 用于加载数据的子进程数。0表示数据将在主进程中加载​​。(默认:0)
    • collate_fn (callable, optional): 合并样本列表以形成一个 mini-batch. # callable可调用对象
    • pin_memory (bool, optional): 如果为 True, 数据加载器会将张量复制到 CUDA 固定内存中,然后再返回它们.
    • drop_last (bool, optional): 设定为 True 如果数据集大小不能被批量大小整除的时候, 将丢掉最后一个不完整的batch,(默认:False).
    • timeout (numeric, optional): 如果为正值,则为从工作人员收集批次的超时值。应始终是非负的。(默认:0)
    • worker_init_fn (callable, optional): If not None, this will be called on each worker subprocess with the worker id (an int in [0, num_workers - 1]) as input, after seeding and before data loading. (default: None).

    num_workers 为0的话表示:数据导入在主进程中进行;其他大于0的数表示:通过多个进程来导入数据,可以加快数据导入速度。

self.num_workers等于0的情况,也就是不采用多进程进行数据读取。先通过indices = next(self.sample_iter)获取长度为batch size的列表:indices,这个列表的每个值表示一个batch中每个数据的index,每执行一次next操作都会读取一批长度为batch size的indices列表。             然后通过self.collate_fn函数将batch size个tuple(每个tuple长度为2,其中第一个值是数据,Tensor类型,第二个值是标签,int类型)封装成一个list,这个list长度为2,两个值都是Tensor,一个是batch size个数据组成的FloatTensor,另一个是batch size个标签组成的LongTensor。所以简单讲self.collate_fn函数就是将batch size个分散的Tensor封装成一个Tensor。

self.num_workers语句是针对多进程或单进程的情况进行初始化,如果不是设置为多进程读取数据,那么就不需要这些初始化操作,后面会介绍单进程数据读取。

通过multiprocessing.SimpleQueue()类创建了一个简单的队列对象。multiprocessing.Process类就是构造进程的类,这里根据设定的进程数来启动,然后赋值给self.workers。接下来的一个for循环就通过调用start方法依次启动self.workers中的进程。

如果设置为多进程读取数据,那么就会采用队列的方式来读,如果不是采用多进程来读取数据,那就采用普通方式来读

2. DataLoader类源代码:

先看看__init__中的几个重要的输入,也就是参数,参数上面已经解释过了。

在__init__中,RandomSampler类表示随机采样且不重复,所以起到的就是shuffle的作用

BatchSampler类则是把batch size个RandomSampler类对象封装成一个,这样就实现了随机选取一个batch的目的。这两个采样类都;是定义在sampler.py脚本中,地址:https://github.com/pytorch/pytorch/blob/master/torch/utils/data/sampler.py。以上这些都是初始化的时候进行的。当代码运行到要从torch.utils.data.DataLoader类生成的对象中取数据的时候,比如: 
train_data=torch.utils.data.DataLoader(…) 
for i, (input, target) in enumerate(train_data): 
… 
就会调用DataLoader类的__iter__方法,__iter__方法就一行代码:return DataLoaderIter(self),输入正是DataLoader类的属性。因此当调用__iter__方法的时候就牵扯到另外一个类:DataLoaderIter,接下来介绍。

class DataLoader(object):
"""
    Data loader. Combines a dataset and a sampler, and provides
    single- or multi-process iterators over the dataset.

    Arguments:
        dataset (Dataset): dataset from which to load the data.
        batch_size (int, optional): how many samples per batch to load
            (default: 1).
        shuffle (bool, optional): set to ``True`` to have the data reshuffled
            at every epoch (default: False).
        sampler (Sampler, optional): defines the strategy to draw samples from
            the dataset. If specified, ``shuffle`` must be False.
        batch_sampler (Sampler, optional): like sampler, but returns a batch of
            indices at a time. Mutually exclusive with batch_size, shuffle,
            sampler, and drop_last.
        num_workers (int, optional): how many subprocesses to use for data
            loading. 0 means that the data will be loaded in the main process.
            (default: 0)
        collate_fn (callable, optional): merges a list of samples to form a mini-batch.
        pin_memory (bool, optional): If ``True``, the data loader will copy tensors
            into CUDA pinned memory before returning them.
        drop_last (bool, optional): set to ``True`` to drop the last incomplete batch,
            if the dataset size is not divisible by the batch size. If ``False`` and
            the size of dataset is not divisible by the batch size, then the last batch
            will be smaller. (default: False)
        timeout (numeric, optional): if positive, the timeout value for collecting a batch
            from workers. Should always be non-negative. (default: 0)
        worker_init_fn (callable, optional): If not None, this will be called on each
            worker subprocess with the worker id as input, after seeding and before data
            loading. (default: None)
"""

    def __init__(self, dataset, batch_size=1, shuffle=False, sampler=None, batch_sampler=None,
                 num_workers=0, collate_fn=default_collate, pin_memory=False, drop_last=False,
                 timeout=0, worker_init_fn=None):
        self.dataset = dataset
        self.batch_size = batch_size
        self.num_workers = num_workers
        self.collate_fn = collate_fn
        self.pin_memory = pin_memory
        self.drop_last = drop_last
        self.timeout = timeout
        self.worker_init_fn = worker_init_fn

        if timeout < 0:
            raise ValueError('timeout option should be non-negative')

        if batch_sampler is not None:
            if batch_size > 1 or shuffle or sampler is not None or drop_last:
                raise ValueError('batch_sampler is mutually exclusive with '
                                 'batch_size, shuffle, sampler, and drop_last')

        if sampler is not None and shuffle:
            raise ValueError('sampler is mutually exclusive with shuffle')

        if self.num_workers < 0:
            raise ValueError('num_workers cannot be negative; '
                             'use num_workers=0 to disable multiprocessing.')

        if batch_sampler is None:
            if sampler is None:
                if shuffle:
                    sampler = RandomSampler(dataset)
                else:
                    sampler = SequentialSampler(dataset)
            batch_sampler = BatchSampler(sampler, batch_size, drop_last)

        self.sampler = sampler
        self.batch_sampler = batch_sampler

    def __iter__(self):
        return DataLoaderIter(self)

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

原博客:https://blog.csdn.net/u014380165/article/details/79058479