torch.reshape(tensor,shape=[-1])

物物不物于物

已于 2023-03-17 21:38:01 修改

阅读量359

点赞数 1

文章标签： python 深度学习人工智能 Powered by 金山文档

于 2023-02-08 10:11:11 首次发布

本文链接：https://blog.csdn.net/qq_50799276/article/details/128930550

版权

将张量的shape改为一维张量

a = torch.arange(12).reshape(3,4)
a

# tensor([[ 0,  1,  2,  3],
#        [ 4,  5,  6,  7],
#        [ 8,  9, 10, 11]])

b = torch.reshape(a,shape=[-1])
b

# tensor([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])

b.shape

# torch.Size([12])

x = np.array([[[1,2,3],[4,5,6]],[[7,8,9],[10,11,12]]])
x

# array([[[ 1,  2,  3],
#        [ 4,  5,  6]],
#       [[ 7,  8,  9],
#        [10, 11, 12]]])

x = torch.from_numpy(x)
x

# tensor([[[ 1,  2,  3],
#         [ 4,  5,  6]],
#        [[ 7,  8,  9],
#         [10, 11, 12]]], dtype=torch.int32)

res = torch.reshape(x,shape=[-1])
res

# tensor([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12], dtype=torch.int32)