[Python知识库]Python中的list、numpy、torch.tensor相互转换

参考：Python中的list、numpy、torch.tensor相互转换 - 百度文库

import torch

state = torch.tensor([[0.2687, 6.5176],
        [0.2683, 6.5396],
        [0.2690, 6.5230],
        [0.2668, 6.5259]])

state = state.numpy().tolist()
print('state,',state)


policy_out = []
for i in range(4):
    print('state[i]:',state[i])
    print('type(state[i]):', type(state[i]))
    policy_out.append(state[i])
print('policy_out',policy_out)

policy_out=torch.Tensor(policy_out)
policy_out=policy_out.cuda()
print('policy_out',policy_out)


c=[]
a= torch.tensor([[ 0.0389, -0.0858]])
b = torch.tensor([[ 0.0369, -0.0861]])
a=a.numpy().tolist()
b=b.numpy().tolist()
print('a[0]:',a[0])
print('type:',type(a[0]))
c.append(a[0])
print('c',c)

开发者涨薪指南

48位大咖的思考法则、工作方式、逻辑体系