Considere o código python a seguir.

import torch
from torch import nn, Tensor

class CVMNet(nn.Module):
  def __init__(self,
              dim_in: int,
              dim_hidden: int,
              n_classes: int):
    super().__init__()
    self.i_layer = nn.Linear(dim_in, dim_hidden)
    self.h_layer = nn.Sequential(
      nn.Linear(dim_hidden, dim_hidden // 2),
      nn.Tanh()
  )
  self.o_layer = nn.Sequential(
    nn.Dropout(0.2),
    nn.Linear(dim_hidden // 2, n_classes),
    nn.Softmax(dim=-1)
  )

  def forward(self, x: Tensor):
    return self.o_layer(self.h_layer(self.i_layer(x)))

model = CVMNet(400, 100, 3)
pred = model(torch.eye(20).flatten()).detach()

Sobre a variável pred, é correto afirmar que: