GRU4RecF

Reference:

Balázs Hidasi et al. “Parallel Recurrent Neural Network Architectures for Feature-rich Session-based Recommendations.” in RecSys 2016.

class recbole.model.sequential_recommender.gru4recf.GRU4RecF(config, dataset)[source]

Bases: recbole.model.abstract_recommender.SequentialRecommender

In the original paper, the authors proposed several architectures. We compared 3 different architectures:

  1. Concatenate item input and feature input and use single RNN,

  2. Concatenate outputs from two different RNNs,

  3. Weighted sum of outputs from two different RNNs.

We implemented the optimal parallel version(2), which uses different RNNs to encode items and features respectively and concatenates the two subparts’s outputs as the final output. The different RNN encoders are trained simultaneously.

calculate_loss(interaction)[source]

Calculate the training loss for a batch data.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Training loss, shape: []

Return type

torch.Tensor

forward(item_seq, item_seq_len)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

full_sort_predict(interaction)[source]

full sort prediction function. Given users, calculate the scores between users and all candidate items.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Predicted scores for given users and all candidate items, shape: [n_batch_users * n_candidate_items]

Return type

torch.Tensor

predict(interaction)[source]

Predict the scores between users and items.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Predicted scores for given users and items, shape: [batch_size]

Return type

torch.Tensor