PyTorch torch.Tensor

torch.Tensor
https://pytorch.org/docs/stable/tensors.html

A torch.Tensor is a multi-dimensional matrix containing elements of a single data type.
torch.Tensor 是一种包含单一数据类型元素的多维矩阵。

1. Data types

Torch defines tensor types with the following data types:

torch.Tensor is an alias for the default tensor type (torch.FloatTensor).
torch.Tensor 是默认的 tensor 类型 (torch.FloatTensor) 的简称。

A tensor can be constructed from a Python list or sequence using the torch.tensor() constructor:
Tensor 可以用 torch.tensor() 转换 Python 的 list 或序列​​生成：

(pt-1.4_py-3.6) yongqiang@yongqiang:~$ python
Python 3.6.10 |Anaconda, Inc.| (default, May  8 2020, 02:54:21)
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> import numpy as np
>>>
>>> torch.tensor([[1., -1.], [1., -1.]])
tensor([[ 1., -1.],
        [ 1., -1.]])
>>>
>>> torch.tensor(np.array([[1, 2, 3], [4, 5, 6]]))
tensor([[1, 2, 3],
        [4, 5, 6]])
>>>
>>> exit()
(pt-1.4_py-3.6) yongqiang@yongqiang:~$

torch.tensor() always copies data. If you have a Tensor data and just want to change its requires_grad flag, use requires_grad_() or detach() to avoid a copy. If you have a numpy array and want to avoid a copy, use torch.as_tensor().
torch.tensor() 总是拷贝 data。如果你有一个 Tensor data 并且仅仅想改变它的 requires_grad 属性, 可用 requires_grad_() or detach() 来避免拷贝。如果你有一个 numpy 数组并且想避免拷贝, 请使用 torch.as_tensor().

A tensor of specific data type can be constructed by passing a torch.dtype and/or a torch.device to a constructor or tensor creation op:
指定数据类型的 Tensor 可以通过传递参数 torch.dtype 和/或 torch.device 到构造函数生成：

>>> torch.zeros([2, 4], dtype=torch.int32)
tensor([[ 0,  0,  0,  0],
        [ 0,  0,  0,  0]], dtype=torch.int32)
>>> cuda0 = torch.device('cuda:0')
>>> torch.ones([2, 4], dtype=torch.float64, device=cuda0)
tensor([[ 1.0000,  1.0000,  1.0000,  1.0000],
        [ 1.0000,  1.0000,  1.0000,  1.0000]], dtype=torch.float64, device='cuda:0')

The contents of a tensor can be accessed and modified using Python’s indexing and slicing notation:
Tensor 的内容可以通过 Python 索引或者切片访问以及修改：

>>> x = torch.tensor([[1, 2, 3], [4, 5, 6]])
>>> print(x[1][2])
tensor(6)
>>> x[0][1] = 8
>>> print(x)
tensor([[ 1,  8,  3],
        [ 4,  5,  6]])

Use torch.Tensor.item() to get a Python number from a tensor containing a single value:
使用 torch.Tensor.item() 从只有一个值的 Tensor 中获取 Python Number：

>>> x = torch.tensor([[1]])
>>> x
tensor([[ 1]])
>>> x.item()
1
>>> x = torch.tensor(2.5)
>>> x
tensor(2.5000)
>>> x.item()
2.5

A tensor can be created with requires_grad=True so that torch.autograd records operations on them for automatic differentiation.
Tensor 可以通过参数 requires_grad=True 创建, 这样 torch.autograd 会记录相关的运算实现自动求导。

(pt-1.4_py-3.6) yongqiang@yongqiang:~$ python
Python 3.6.10 |Anaconda, Inc.| (default, May  8 2020, 02:54:21)
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> data = torch.tensor([[1., -1.], [1., 1.]], requires_grad=True)
>>> data
tensor([[ 1., -1.],
        [ 1.,  1.]], requires_grad=True)
>>>
>>> data_pow = data.pow(2)
>>> data_pow
tensor([[1., 1.],
        [1., 1.]], grad_fn=<PowBackward0>)
>>>
>>> data_sum = data_pow.sum()
>>> data_sum
tensor(4., grad_fn=<SumBackward0>)
>>>
>>> data_sum.backward()
>>>
>>> data.grad
tensor([[ 2., -2.],
        [ 2.,  2.]])
>>> exit()
(pt-1.4_py-3.6) yongqiang@yongqiang:~$

Each tensor has an associated torch.Storage, which holds its data. The tensor class also provides multi-dimensional, strided view of a storage and defines numeric operations on it.
每一个 tensor 都有一个相应的 torch.Storage 保存其数据。tensor 类提供了一个多维的、strided 视图，并定义了数值操作。

Methods which mutate a tensor are marked with an underscore suffix. For example, torch.FloatTensor.abs_() computes the absolute value in-place and returns the modified tensor, while torch.FloatTensor.abs() computes the result in a new tensor.
修改 tensor 的方法可以用一个下划线后缀来标示。比如 torch.FloatTensor.abs_() 会在原地计算绝对值并返回修改的张量，而 torch.FloatTensor.abs() 将会在新张量中计算结果.

To change an existing tensor’s torch.device and/or torch.dtype, consider using to() method on the tensor.
为了改变已有的 tensor 的 torch.device 和/或 torch.dtype，考虑使用 to() 方法。

Current implementation of torch.Tensor introduces memory overhead, thus it might lead to unexpectedly high memory usage in the applications with many tiny tensors. If this is your case, consider using one large structure.
Torch.Tensor 的当前实现会引入内存开销，因此可能会在具有许多微小张量的应用程序中导致意外的高内存使用率。如果是这种情况，请考虑使用一种大型结构。

References

[1] Yongqiang Cheng, https://yongqiang.blog.csdn.net/