Maison >développement back-end >Tutoriel Python >Déaplatir dans PyTorch
Achetez-moi un café☕
*Mémos :
Unflatten() peut ajouter zéro ou plusieurs dimensions au tenseur 1D ou plus D de zéro ou plusieurs éléments, obtenant le tenseur 1D ou plus D de zéro ou plusieurs éléments comme indiqué ci-dessous :
*Mémos :
import torch from torch import nn unflatten = nn.Unflatten() unflatten # Unflatten(dim=0, unflattened_size=(6,)) unflatten.dim # 0 unflatten.unflattened_size # (6,) my_tensor = torch.tensor([7, 1, -8, 3, -6, 0]) unflatten = nn.Unflatten(dim=0, unflattened_size=(6,)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1,)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(6,)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1,)) unflatten(input=my_tensor) # tensor([7, 1, -8, 3, -6, 0]) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 6)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1, 6)) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 6)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1, 6)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, -1)) unflatten(input=my_tensor) # tensor([[7, 1, -8, 3, -6, 0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2, 3)) unflatten = nn.Unflatten(dim=0, unflattened_size=(2, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(2, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(2, -1)) unflatten(input=my_tensor) # tensor([[7, 1, -8], [3, -6, 0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(3, 2)) unflatten = nn.Unflatten(dim=0, unflattened_size=(3, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(3, 2)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(3, -1)) unflatten(input=my_tensor) # tensor([[7, 1], [-8, 3], [-6, 0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(6, 1)) unflatten = nn.Unflatten(dim=0, unflattened_size=(6, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(6, 1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(6, -1)) unflatten(input=my_tensor) # tensor([[7], [1], [-8], [3], [-6], [0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 2, 3)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1, 2, 3)) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, -1, 3)) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 2, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 2, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1, 2, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, -1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 2, -1)) unflatten(input=my_tensor) # tensor([[[7, 1, -8], [3, -6, 0]]]) etc my_tensor = torch.tensor([[7, 1, -8], [3, -6, 0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2,)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1,)) unflatten = nn.Unflatten(dim=1, unflattened_size=(3,)) unflatten = nn.Unflatten(dim=1, unflattened_size=(-1,)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(3,)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1,)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(2,)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(-1,)) unflatten(input=my_tensor) # tensor([[7, 1, -8], [3, -6, 0]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 2)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1, 2)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(1, 2)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(-1, 2)) unflatten(input=my_tensor) # tensor([[[7, 1, -8], [3, -6, 0]]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2, 1)) unflatten = nn.Unflatten(dim=0, unflattened_size=(2, -1)) unflatten = nn.Unflatten(dim=1, unflattened_size=(1, 3)) unflatten = nn.Unflatten(dim=1, unflattened_size=(-1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1, 3)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(2, 1)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(2, -1)) unflatten(input=my_tensor) # tensor([[[7, 1, -8]], [[3, -6, 0]]]) unflatten = nn.Unflatten(dim=1, unflattened_size=(3, 1)) unflatten = nn.Unflatten(dim=1, unflattened_size=(3, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(3, 1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(3, -1)) unflatten(input=my_tensor) # tensor([[[7], [1], [-8]], [[3], [-6], [0]]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 1, 2)) unflatten = nn.Unflatten(dim=0, unflattened_size=(-1, 1, 2)) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, -1, 2)) unflatten = nn.Unflatten(dim=0, unflattened_size=(1, 1, -1)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(1, 1, 2)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(-1, 1, 2)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(1, -1, 2)) unflatten = nn.Unflatten(dim=-2, unflattened_size=(1, 1, -1)) unflatten(input=my_tensor) # tensor([[[[7, 1, -8], [3, -6, 0]]]]) unflatten = nn.Unflatten(dim=1, unflattened_size=(1, 1, 3)) unflatten = nn.Unflatten(dim=1, unflattened_size=(-1, 1, 3)) unflatten = nn.Unflatten(dim=1, unflattened_size=(1, -1, 3)) unflatten = nn.Unflatten(dim=1, unflattened_size=(1, 1, -1)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(-1, 1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, -1, 3)) unflatten = nn.Unflatten(dim=-1, unflattened_size=(1, 1, -1)) unflatten(input=my_tensor) # tensor([[[[7, 1, -8]]], [[[3, -6, 0]]]]) my_tensor = torch.tensor([[7., 1., -8.], [3., -6., 0.]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2,)) unflatten(input=my_tensor) # tensor([[7., 1., -8.], [3., -6., 0.]]) my_tensor = torch.tensor([[7.+0.j, 1.+0.j, -8.+0.j], [3.+0.j, -6.+0.j, 0.+0.j]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2,)) unflatten(input=my_tensor) # tensor([[7.+0.j, 1.+0.j, -8.+0.j], # [3.+0.j, -6.+0.j, 0.+0.j]]) my_tensor = torch.tensor([[True, False, True], [False, True, False]]) unflatten = nn.Unflatten(dim=0, unflattened_size=(2,)) unflatten(input=my_tensor) # tensor([[True, False, True], [False, True, False]])
Ce qui précède est le contenu détaillé de. pour plus d'informations, suivez d'autres articles connexes sur le site Web de PHP en chinois!