torch.utils.Dataset PyTorch数据集配置

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

    All datasets that represent a map from keys to data samples should subclass
    it. All subclasses should overwrite :meth:`__getitem__`, supporting fetching a
    data sample for a given key. Subclasses could also optionally overwrite
    :meth:`__len__`, which is expected to return the size of the dataset by many
    :class:`~torch.utils.data.Sampler` implementations and the default options
    of :class:`~torch.utils.data.DataLoader`.

    .. note::
      :class:`~torch.utils.data.DataLoader` by default constructs a index
      sampler that yields integral indices.  To make it work with a map-style
      dataset with non-integral indices/keys, a custom sampler must be provided.
    """

    def __getitem__(self, index):
        raise NotImplementedError

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

    # No `def __len__(self)` default?
    # See NOTE [ Lack of Default `__len__` in Python Abstract Base Classes ]
    # in pytorch/torch/utils/data/sampler.py

根据api描述，Dataset的子类必须overwrite两个方法：

• __getitem__: 支持获取给定key/index的数据样本。
• __len__: 支持获取数据size

再加上init方法，就是最基本的三个方法。

• __init__:

以sklearn中breast_cancer数据为例，做一个dataset的Demo [github source]：

import torch
from torch.utils.data import Dataset, DataLoader
from sklearn.datasets import load_breast_cancer
from sklearn.preprocessing import StandardScaler
import numpy as np
class MyDataSet(Dataset):
    def __init__(self) -> None:
        super().__init__()
        self._data = load_breast_cancer()
        self._X = self._data.data.astype(np.float32)
        self._y = self._data.target.astype(np.float32)
        self._sc = StandardScaler()
        self._X = self._sc.fit_transform(self._X)
        self._X = torch.from_numpy(self._X).type(torch.float32)
        self._y = torch.from_numpy(self._y).type(torch.float32)
        self._len = len(self._X)

    def __getitem__(self, index: int):
        return self._X[index], self._y[index]

    def __len__(self) -> int:
        return self._len

data = MyDataSet()
dataLoader = DataLoader(dataset=data, batch_size=10, shuffle=True, drop_last=False, num_workers=1)
n = 0
for data_val, label_val in dataLoader:
    print(f'iter: {n}')
    print(f'data_val: {len(data_val)}, label_val: {len(label_val)}')
    n += 1

# iter: 0
# data_val: 10, label_val: 10
# iter: 1
# data_val: 10, label_val: 10
# iter: 2
# data_val: 10, label_val: 10
# iter: 3
# data_val: 10, label_val: 10
# iter: 4
# data_val: 10, label_val: 10
...