Repository files navigation

Keras Recommenders is a library for building recommender systems on top ofKeras 3. Keras Recommenders works natively with TensorFlow, JAX, or PyTorch. Itprovides a collection of building blocks which help with the full workflow ofcreating a recommender system. As it's built on Keras 3, models can be trainedand serialized in any framework and re-used in another without costlymigrations.

This library is an extension of the core Keras API; all high-level modulesreceive that same level of polish as core Keras. If you are familiar with Keras,congratulations! You already understand most of Keras Recommenders.

• Home page
• Examples
• API documentation

Choose a backend:

importosos.environ["KERAS_BACKEND"]="jax"# Or "tensorflow" or "torch"!

Import KerasRS and other libraries:

importkerasimportkeras_rsimportnumpyasnp

Define a simple model using theFeatureCross layer:

vocabulary_size=32embedding_dim=6inputs=keras.Input(shape=(),name='indices',dtype="int32")x0=keras.layers.Embedding(input_dim=vocabulary_size,output_dim=embedding_dim)(inputs)x1=keras_rs.layers.FeatureCross()(x0,x0)x2=keras_rs.layers.FeatureCross()(x0,x1)output=keras.layers.Dense(units=10)(x2)model=keras.Model(inputs,output)

Compile the model:

model.compile(loss=keras.losses.MeanSquaredError(),optimizer=keras.optimizers.Adam(learning_rate=3e-4))

Callmodel.fit() on dummy data:

batch_size=2x=np.random.randint(0,vocabulary_size,size=(batch_size,))y=np.random.random(size=(batch_size,))model.fit(x,y=y)

If your task is to rank items in a list, you can make use of the ranking lossesand metrics which KerasRS provides. Below, we use the pairwise hinge loss andtrack the nDCG metric:

model.compile(loss=keras_rs.losses.PairwiseHingeLoss(),metrics=[keras_rs.metrics.NDCG()],optimizer=keras.optimizers.Adam(learning_rate=3e-4),)

Keras Recommenders is available on PyPI askeras-rs:

pip install keras-rs

To try out the latest version of Keras Recommenders, you can use our nightlypackage:

pip install keras-rs-nightly

ReadGetting started with Keras for moreinformation on installing Keras 3 and compatibility with different frameworks.

If you have Keras 3 installed in your environment (see installation above), youcan use Keras Recommenders with any of JAX, TensorFlow and PyTorch. To do so,set theKERAS_BACKEND environment variable. For example:

export KERAS_BACKEND=jax

Or in Colab, with:

importosos.environ["KERAS_BACKEND"]="jax"importkeras_rs

We followSemantic Versioning, and plan to providebackwards compatibility guarantees both for code and saved models built with ourcomponents. While we continue with pre-release0.y.z development, we may breakcompatibility at any time and APIs should not be considered stable.

If Keras Recommenders helps your research, we appreciate your citations.Here is the BibTeX entry:

@misc{kerasrecommenders2024,title={KerasRecommenders},author={Hertschuh, Fabien and  Chollet, Fran\c{c}ois and Sharma, Abheesht and others},year={2024},howpublished={\url{https://github.com/keras-team/keras-rs}},}

Thank you to all of our wonderful contributors!