Single

In [ ]:

from torch import nn
import torch

from torch import nn import torch

In [ ]:

class single(nn.Module):
    def __init__(self, dim, D):
        super(single, self).__init__()

        self.fc = nn.Linear(dim, D)
        self.output = nn.Linear(D, 3)

        self.bn = nn.BatchNorm1d(D)
        #self.drop = nn.Dropout(0.3)
        self.relu = nn.ReLU()

    def forward(self, emb):
        #out = self.drop(self.relu(self.bn(self.fc(emb))))
        out = self.relu(self.bn(self.fc(emb)))
        return self.output(out)

class single(nn.Module): def __init__(self, dim, D): super(single, self).__init__() self.fc = nn.Linear(dim, D) self.output = nn.Linear(D, 3) self.bn = nn.BatchNorm1d(D) #self.drop = nn.Dropout(0.3) self.relu = nn.ReLU() def forward(self, emb): #out = self.drop(self.relu(self.bn(self.fc(emb)))) out = self.relu(self.bn(self.fc(emb))) return self.output(out)