massquantity

Hi, I wouldn't recommend doing multi-model inference with the models in this library as they are primarily designed for single tasks. I'm planning a major upgrade of the whole library,...

`libsering.serialization.tf_saved` is not compatible with `tf.saved_model.load`. If you want to save and load a model, use [`Save/Load`](https://librecommender.readthedocs.io/en/latest/user_guide/evaluation_save_load.html) API.

Hi, The TwoTower class inherits from the [DynEmbedBase](https://github.com/massquantity/LibRecommender/blob/master/libreco/bases/dyn_embed_base.py#L17) class, which inherits from the [EmbedBase](https://github.com/massquantity/LibRecommender/blob/master/libreco/bases/embed_base.py) class. The `build_model` method is called in the [fit](https://github.com/massquantity/LibRecommender/blob/master/libreco/bases/embed_base.py#L133) method of EmbedBase. `self.serving_topk()` is used in...

`model.recommend_user` mainly inherits from the [base classes](https://librecommender.readthedocs.io/en/latest/api/algorithms/bases.html), and different models have different base classes. For example, the [DeepFM doc](https://librecommender.readthedocs.io/en/latest/api/algorithms/deepfm.html) indicates the base class is `TfBase`, so the link is https://github.com/massquantity/LibRecommender/blob/master/libreco/bases/tf_base.py#L186

Thanks for the question! Currently, the library doesn’t support precomputed embeddings directly—the workaround of splitting them into individual dense features works but isn’t ideal. We may consider adding this feature...

Sorry there is no way to stream the data. The library is built with the assumption that the entire dataset can be loaded into memory. While we recognize the need...

This warning is because your test data contains users or items not existing in the training data. They are "unknown" to the trained model. If your data is large, they...

Hi, I don't think the growing graph is causing the problem, since the graph remains on CPU and only the sampled batches and subgraphs are transferred to GPU. Given that...

Each user and item requires an embedding vector in the model. With incremental training, the growing number of users and items will increase memory usage.

This library does not support the removal of non-active users and items, as it is designed for scenarios where the entire model fits into memory. For larger, real-world applications with...