什麼要學習PyTorch?

有的人總是選擇,選擇的人最多的框架,來作為自己的初學框架,比如Tensorflow,但是大多論文的實現都是基於PyTorch的,如果我們要深入論文的細節,就必須選擇學習入門PyTorch

安裝PyTorch

一行命令即可 官網

pip install torch===1.6.0 torchvision===0.7.0 - https://download.pytorch.org/whl/torch_stable.html

時間較久,耐心等待

測試自己是否安裝成功

執行命令測試

import torch
x = torch.rand(5,3)
print(x)

輸出

tensor([[0.5096,0.1209,0.7721],
[0.9486,0.8676,0.2157],
[0.0586,0.3467,0.5015],
[0.9470,0.5654,0.9317],
[0.2127,0.2386,0.0629]])

開始學習PyTorch

不初始化的建立張量

import torch
x = torch.empty([5,5])
print(x)

輸出

tensor([[0.,0.,0.],
[0.,0.]])

隨機建立一個0-1的張量

import torch
x = torch.rand(5,5)
print(x)

輸出

tensor([[0.3369,0.5339,0.8419,0.6857,0.6241],

[0.4991,0.1691,0.8356,0.4574,0.0395],
[0.9714,0.2975,0.9322,0.5213,0.8509],
[0.3037,0.8690,0.3481,0.2538,0.9513],
[0.0156,0.9516,0.3674,0.1831,0.6466]])

建立全為0的張量

import torch
x = torch.zeros(5,5,dtype=torch.float32)
print(x)

建立的時候可以通過dtype指定資料型別

輸出

tensor([[0.,0.]])

使用資料來直接建立張量

import torch
x = torch.zeros([5,5],dtype=torch.float32)
print(x)

輸出

tensor([5.,5.])

使用原有tensor建立新的tensor

import torch
x = torch.tensor([5,dtype=torch.float32)
x = x.new_zeros(5,3)
y = torch.rand_like(x)
print(x)
print(y)

輸出

tensor([[0.,0.]])
tensor([[0.5552,0.3333,0.0426],
[0.3861,0.3945,0.6658],
[0.6978,0.3508,0.4813],
[0.8193,0.2274,0.8384],
[0.9360,0.9226,0.1453]])

觀察tensor的維度資訊

x = torch.rand(3,3)
x.size()

輸出

torch.Size([3,3])

一些簡單的運算

x = torch.tensor([1])
y = torch.tensor([3])
'''
方式1
'''
z = x + y
'''
方式2
''' 
z = torch.add(x,y)
'''
方式3
'''
result = torch.empty(1)
# 不初始化資料
torch.add(x,y,out=result)
# 將結果返回到result中
'''
方式4
'''
x.add_(y)

輸出

tensor([4])

索引操作

x = torch.rand(5,5)
x[:,:]
x[1,:]
x[:,1]
x[1,1]

分別輸出

tensor([[0.4012,0.2604,0.1720,0.0996,0.7806],
[0.8734,0.9087,0.4828,0.3543,0.2375],
[0.0924,0.9040,0.4408,0.9758,0.2250],
[0.7179,0.7244,0.6165,0.1142,0.7363],
[0.8504,0.0391,0.0753,0.4530,0.7372]])
tensor([0.8734,0.2375])
tensor([0.2604,0.0391])
tensor(0.9087)

維度變換

x = torch.rand(4,4)
x.view(16)
x.view(8,2)
x.view(-1,8)

分別輸出

tensor([0.9277,0.9547,0.9487,0.9841,0.4114,0.1693,0.8691,0.3954,0.4679,
0.7914,0.7456,0.0522,0.0043,0.2097,0.5932,0.9797])
tensor([[0.9277,0.9547],
[0.9487,0.9841],
[0.4114,0.1693],
[0.8691,0.3954],
[0.4679,0.7914],
[0.7456,0.0522],
[0.0043,0.2097],
[0.5932,0.9797]])
tensor([[0.9277,0.7914,0.9797]])

注意:必須維度變換資料的數量必須保持一致

到此這篇關於PyTorch安裝與基本使用詳解的文章就介紹到這了,更多相關PyTorch安裝與使用內容請搜尋我們以前的文章或繼續瀏覽下面的相關文章希望大家以後多多支援我們！