import torch
from torch.autograd import Variable
from torch import nn

torch.manual_seed(0)

x_restored = Variable(torch.load('x.pth'), requires_grad=False)
x_res_repeated = x_restored.repeat(10, 1)

class Network(nn.Module):
    def __init__(self, dim_in, dim_out):
        super(Network, self).__init__()
        self.m = nn.Sequential(
            nn.Linear(dim_in, 100),
            nn.BatchNorm1d(100),
            nn.ReLU(),
            nn.Linear(100, 100),
            nn.BatchNorm1d(100),
            nn.ReLU(),
            nn.Linear(100, dim_out*2)
        )

    def forward(self, X):
        return self.m(X)

net = Network(x_restored.size()[1], 2)
print(net(x_restored))
print(net(x_res_repeated))