6c17a1019665e1bbaee6826b3204bb7d14996a9f,Tars/models/vae.py,VAE,__init__,#VAE#Any#Any#Any#Any#Any#Any#,13

Before Change

        super().__init__()

        self.encoder = encoder
        self.decoder = decoder
        self.other_distributions = nn.ModuleList(tolist(other_distributions))

        // set losses
        reconstruction =\
            StochasticReconstructionLoss(self.encoder, self.decoder)
        loss_cls = (reconstruction + regularizer).mean()
        self.loss_cls = loss_cls
        self.test_loss_cls = loss_cls
        self.loss_text = str(loss_cls)

        // set params and optim
        q_params = list(self.encoder.parameters())
        p_params = list(self.decoder.parameters())
        other_params = list(self.other_distributions.parameters())
        params = q_params + p_params + other_params
        self.optimizer = optimizer(params, **optimizer_params)

    def train(self, train_x, **kwargs):
        self.decoder.train()

After Change

                 optimizer=optim.Adam,
                 optimizer_params={}):

        distributions = nn.ModuleList([encoder, decoder] + tolist(other_distributions))
        super().__init__(distributions)

        // set losses
        reconstruction =\

Instances