pytorch学習ノート(1)-Tensorの作成から

文書ディレクトリ

Tensor

の作成

1.基本作成

2.特殊値

を作成

3.ランダム値と特定のシーケンスTensor

の作成

4.Numpy回転Tensor

5.Tensor回転Numpy

Tensorの作成
pytorch学習の最も基礎的な学習はテンソルの作成から始まる.
1.基本作成

import torch
#     
a = torch.Tensor([[1, 2],[3, 4]])
print(a)

#      ，    
b = torch.Tensor(2, 2)
print(b)

d = torch.tensor(((1, 2), (3, 4)))
print(d.type())
print(d.type_as(a))

tensor([[1., 2.], [3., 4.]]) torch.FloatTensor tensor([[1., 2.], [3., 4.]]) torch.LongTensor tensor([[1., 2.], [3., 4.]])
2.特別な値の作成

#   Tensor
d = torch.empty(2,3)
print(d.type())
print(d.type_as(a))

#  0 Tensor
d = torch.zeros(2,3)
print(d.type())
print(d.type_as(a))

#    0 Tensor
d = torch.zeros_like(d)
print(d.type())
print(d.type_as(a))

#    1
d = torch.eye(2, 2)
print(d.type())
print(d.type_as(a))

# 1
d = torch.ones(2, 2)
print(d.type())
print(d.type_as(a))

d = torch.ones_like(d)
print(d.type())
print(d.type_as(a))

torch.FloatTensor tensor([[0., 0., 0.], [0., 0., 0.]])
torch.FloatTensor tensor([[0., 0., 0.], [0., 0., 0.]])
torch.FloatTensor tensor([[0., 0., 0.], [0., 0., 0.]])
torch.FloatTensor tensor([[1., 0.], [0., 1.]])
torch.FloatTensor tensor([[1., 1.], [1., 1.]])
torch.FloatTensor tensor([[1., 1.], [1., 1.]])
3.ランダム値と特定シーケンスTensorの作成

#0-1   
d = torch.rand(2, 3)
print(d.type())
print(d.type_as(a))

d = torch.arange(2, 10, 2)
print(d.type())
print(d.type_as(a))

d = torch.linspace(10, 2, 3)
print(d.type())
print(d.type_as(a))

dd = torch.normal(mean=0, std=1, size=(2, 3), out=b)
print(b)
print(dd)

d = torch.normal(mean=torch.rand(5), std=torch.rand(5))
print(d.type())
print(d.type_as(a))


d = torch.Tensor(2, 2).uniform_(-1, 1)
print(d.type())
print(d.type_as(a))


d = torch.randperm(10)
print(d.type())
print(d.type_as(a))

torch.FloatTensor tensor([[0.6432, 0.4434, 0.3289], [0.6581, 0.7615, 0.6703]])
torch.LongTensor tensor([2., 4., 6., 8.])
torch.FloatTensor tensor([10., 6., 2.])
tensor([[-0.4051, -0.5710, -1.3798], [ 0.3047, -0.3695, -0.2271]])
tensor([[-0.4051, -0.5710, -1.3798], [ 0.3047, -0.3695, -0.2271]])
torch.FloatTensor tensor([0.4624, 1.2237, 1.1937, 1.3881, 0.5219])
torch.FloatTensor tensor([[ 0.9237, 0.2990], [-0.5562, -0.2350]])
torch.LongTensor tensor([2., 7., 4., 3., 5., 6., 9., 1., 0., 8.])
4.Numpy回転Tensor

a = torch.ones(5)
b = a.numpy()    
a.add_(1)
print(a)
print(b)

tensor([2., 2., 2., 2., 2.]) [2. 2. 2. 2. 2.]
5.Tensor回転Numpy

import numpy as np
a = np.ones(5)
b = torch.from_numpy(a)
np.add(a,1,out = a)
print(a)
print(b)

[2. 2. 2. 2. 2.] tensor([2., 2., 2., 2., 2.], dtype=torch.float64)