PyTorch学习笔记（一）：创建Tensor

1、Tensor的维度和尺寸

不同类型的数据可以用不同维度(dimension)的张量来表示，标量为0维张量，向量为1维张量，矩阵为2维以上的张量；可以简单地总结为：有几层中括号，就是多少维的张量；可以使用shape属性或者size()方法查看张量在每一维的长度；

标量：

# 标量（0维张量）
scalar = torch.tensor(1)
print(scalar)
print(scalar.dim()) 
print(scalar.size())
print(scalar.shape)
# 输出
tensor(1)
0
torch.Size([])
torch.Size([])

向量：

#向量（1维张量）
vector = torch.tensor([1.0, 2.0, 3.0, 4.0]) 
print(vector)
print(vector.dim())
print(vector.size())
print(vector.shape)
# 输出
tensor([1., 2., 3., 4.])
1
torch.Size([4])
torch.Size([4])

矩阵

# 2维张量
matrix = torch.tensor([[1.0,2.0],[3.0,4.0]]) 
print(matrix)
print(matrix.dim())
print(matrix.size())
print(matrix.shape)
# 输出
tensor([[1., 2.],
        [3., 4.]])
2
torch.Size([2, 2])
torch.Size([2, 2])

# 3维张量
tensor3 = torch.tensor([[[1.0,2.0],[3.0,4.0]],[[5.0,6.0],[7.0,8.0]]])  
print(tensor3)
print(tensor3.dim())
print(tensor3.size())
print(tensor3.shape)
# 输出
tensor([[[1., 2.],
         [3., 4.]],

        [[5., 6.],
         [7., 8.]]])
3
torch.Size([2, 2, 2])
torch.Size([2, 2, 2])

# 4维张量
tensor4 = torch.tensor([[[[1.0,1.0],[2.0,2.0]],[[3.0,3.0],[4.0,4.0]]],
                        [[[5.0,5.0],[6.0,6.0]],[[7.0,7.0],[8.0,8.0]]]])  
print(tensor4)
print(tensor4.dim())
print(tensor4.size())
print(tensor4.shape)
# 输出
tensor([[[[1., 1.],
          [2., 2.]],

         [[3., 3.],
          [4., 4.]]],


        [[[5., 5.],
          [6., 6.]],

         [[7., 7.],
          [8., 8.]]]])
4
torch.Size([2, 2, 2, 2])
torch.Size([2, 2, 2, 2])

2、创建Tensor

Import from List:

a = [1, 2]
a = torch.tensor(a)
print(a)
# 输出
tensor([1, 2])

Import from numpy:

a = np.array([1, 2])
a = torch.from_numpy(a)
print(a)
# 输出
tensor([1, 2], dtype=torch.int32)

Torch.FloatTensor(d1, d2, d3)

tensor = torch.FloatTensor(1,2)
print(tensor)
print(tensor.dim())
print(tensor.shape)
# 输出
tensor([[-1.2401e-17,  4.5907e-41]])
2
torch.Size([1, 2])


tensor = torch.FloatTensor(2, 3, 3)
print(tensor)
print(tensor.dim())
print(tensor.shape)
# 输出
tensor([[[9.2755e-39, 1.0561e-38, 1.0929e-38],
         [9.6429e-39, 8.9082e-39, 9.4592e-39],
         [5.0510e-39, 4.2246e-39, 1.0286e-38]],

        [[1.0653e-38, 1.0194e-38, 8.4490e-39],
         [1.0469e-38, 9.3674e-39, 9.9184e-39],
         [8.7245e-39, 9.2755e-39, 8.9082e-39]]])
3
torch.Size([2, 3, 3])

rand

a = torch.rand(2,3)
print(a)
print(a.dim())
print(a.shape)

# 输出
tensor([[0.4862, 0.3772, 0.4883],
        [0.8780, 0.4097, 0.4205]])
2
torch.Size([2, 3])

rand_like

a = torch.rand(2,3)
b = torch.rand_like(a)
print(b)
print(b.dim())
print(b.shape)
# 输出
tensor([[0.6247, 0.7405, 0.7412],
        [0.8705, 0.7099, 0.0307]])
2
torch.Size([2, 3])

randint

a = torch.randint(1, 10, (3, 3))
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[6, 3, 1],
        [1, 4, 5],
        [9, 5, 9]])
2
torch.Size([3, 3])

randn

a = torch.randn(3,3)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[ 0.8001,  1.7763,  0.3723],
        [ 0.0721,  0.6027,  0.1838],
        [-0.9242, -1.8679, -0.1747]])
2
torch.Size([3, 3])

full

a = torch.full([2, 3], 10)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[10, 10, 10],
        [10, 10, 10]])
2
torch.Size([2, 3])

arange

a = torch.arange(0, 10)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
1
torch.Size([10])

a = torch.arange(0, 10, 2)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([0, 2, 4, 6, 8])
1
torch.Size([5])

ones

a = torch.ones(3, 3)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.]])
2
torch.Size([3, 3])

zeros

a = torch.zeros(3, 3)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[0., 0., 0.],
        [0., 0., 0.],
        [0., 0., 0.]])
2
torch.Size([3, 3])

eye

a = torch.eye(3)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([[1., 0., 0.],
        [0., 1., 0.],
        [0., 0., 1.]])
2
torch.Size([3, 3])

randperm

a = torch.randperm(10)
print(a)
print(a.dim())
print(a.shape)
# 输出
tensor([3, 4, 2, 8, 7, 5, 0, 9, 1, 6])
1
torch.Size([10])