PyTorch | torch.manual_seed是什么意思？torch随机数manual_seed有什么用？如何理解torch.manual_seed

Posted 2022-09-22 人工智能博士

torch.manual_seed(1)是为了设置CPU的的随机数固定，使得紧跟着的rand（）函数生成的值是固定的随机！

# 王博Kings
import torch
torch.manual_seed(1)
print(torch.rand(5))

tensor([0.7576, 0.2793, 0.4031, 0.7347, 0.0293])

# 王博Kings
torch.manual_seed(2)
print(torch.rand(5))

tensor([0.6147, 0.3810, 0.6371, 0.4745, 0.7136])

# 王博Kings
torch.manual_seed(1)
print(torch.rand(5))

tensor([0.7576, 0.2793, 0.4031, 0.7347, 0.0293])

# 王博Kings
torch.manual_seed(1)
print(torch.rand(5))
print(torch.rand(5))

tensor([0.7576, 0.2793, 0.4031, 0.7347, 0.0293])
tensor([0.7999, 0.3971, 0.7544, 0.5695, 0.4388])

# 王博Kings
torch.manual_seed(1)
print(torch.rand(5))
torch.manual_seed(1)
print(torch.rand(5))

tensor([0.7576, 0.2793, 0.4031, 0.7347, 0.0293])
tensor([0.6147, 0.3810, 0.6371, 0.4745, 0.7136])

# 王博Kings
torch.manual_seed(1)
print(torch.rand(5))
torch.manual_seed(2)
print(torch.rand(5))

 

tensor([0.7576, 0.2793, 0.4031, 0.7347, 0.0293])
tensor([0.6147, 0.3810, 0.6371, 0.4745, 0.7136])