pytorch 常用函数 max ,eq说明
时间:2022-02-22 10:57:45|栏目:Python代码|点击: 次
max找出tensor 的行或者列最大的值:
找出每行的最大值:
import torch outputs=torch.FloatTensor([[1],[2],[3]]) print(torch.max(outputs.data,1))
输出:
(tensor([ 1., 2., 3.]), tensor([ 0, 0, 0]))
找出每列的最大值:
import torch outputs=torch.FloatTensor([[1],[2],[3]]) print(torch.max(outputs.data,0))
输出结果:
(tensor([ 3.]), tensor([ 2]))
Tensor比较eq相等:
import torch outputs=torch.FloatTensor([[1],[2],[3]]) targets=torch.FloatTensor([[0],[2],[3]]) print(targets.eq(outputs.data))
输出结果:
tensor([[ 0], [ 1], [ 1]], dtype=torch.uint8)
使用sum() 统计相等的个数:
import torch outputs=torch.FloatTensor([[1],[2],[3]]) targets=torch.FloatTensor([[0],[2],[3]]) print(targets.eq(outputs.data).cpu().sum())
输出结果:
tensor(2)
补充知识:PyTorch - torch.eq、torch.ne、torch.gt、torch.lt、torch.ge、torch.le
flyfish
torch.eq、torch.ne、torch.gt、torch.lt、torch.ge、torch.le
以上全是简写
参数是input, other, out=None
逐元素比较input和other
返回是torch.BoolTensor
import torch a=torch.tensor([[1, 2], [3, 4]]) b=torch.tensor([[1, 2], [4, 3]]) print(torch.eq(a,b))#equals # tensor([[ True, True], # [False, False]]) print(torch.ne(a,b))#not equal to # tensor([[False, False], # [ True, True]]) print(torch.gt(a,b))#greater than # tensor([[False, False], # [False, True]]) print(torch.lt(a,b))#less than # tensor([[False, False], # [ True, False]]) print(torch.ge(a,b))#greater than or equal to # tensor([[ True, True], # [False, True]]) print(torch.le(a,b))#less than or equal to # tensor([[ True, True], # [ True, False]])
上一篇:Pytorch测试神经网络时出现 RuntimeError:的解决方案
栏 目:Python代码
本文地址:http://www.codeinn.net/misctech/194232.html