pyhanabi/net.py [279:299]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        self.priv_net = nn.Sequential(
            nn.Linear(self.priv_in_dim, self.hid_dim),
            nn.ReLU(),
            nn.Linear(self.hid_dim, self.hid_dim),
            nn.ReLU(),
            nn.Linear(self.hid_dim, self.hid_dim),
            nn.ReLU(),
        )

        ff_layers = [nn.Linear(self.publ_in_dim, self.hid_dim), nn.ReLU()]
        for i in range(1, self.num_ff_layer):
            ff_layers.append(nn.Linear(self.hid_dim, self.hid_dim))
            ff_layers.append(nn.ReLU())
        self.publ_net = nn.Sequential(*ff_layers)

        self.lstm = nn.LSTM(
            self.hid_dim,
            self.hid_dim,
            num_layers=self.num_lstm_layer,
        ).to(device)
        self.lstm.flatten_parameters()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



pyhanabi/supervised_model.py [76:96]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        self.priv_net = nn.Sequential(
            nn.Linear(self.priv_in_dim, self.hid_dim),
            nn.ReLU(),
            nn.Linear(self.hid_dim, self.hid_dim),
            nn.ReLU(),
            nn.Linear(self.hid_dim, self.hid_dim),
            nn.ReLU(),
        )

        ff_layers = [nn.Linear(self.publ_in_dim, self.hid_dim), nn.ReLU()]
        for i in range(1, self.num_ff_layer):
            ff_layers.append(nn.Linear(self.hid_dim, self.hid_dim))
            ff_layers.append(nn.ReLU())
        self.publ_net = nn.Sequential(*ff_layers)

        self.lstm = nn.LSTM(
            self.hid_dim,
            self.hid_dim,
            num_layers=self.num_lstm_layer,
        ).to(device)
        self.lstm.flatten_parameters()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -