Forward call for multiple user and item embeddings

torch_forward.py

    def forward(self, x, u_cats, i_cats):
        """
        Forward pass
        :param x: Float Tensor
        :param u_cats: User index tensor
        :param i_cats: Item index tensor
        :return: Predictions for this batch
        """
        curr_batch_size = len(u_cats)
        # Take User and Item embeddings for each value
        u_embs = self.user_embedding(u_cats.long())
        i_embs = self.item_embedding(i_cats.long())
        # Orient along the correct axis
        u_embs = u_embs.view(curr_batch_size, -1)
        i_embs = i_embs.view(curr_batch_size, -1)
        # Concat float values and embeddings together
        x = torch.cat([u_embs, i_embs, x], 1)
        return self.layers(x)