Jupyter notebook for building recommender systems. The current system only supports Python 3.6.6 and Scala with Spark support. In addition, the following libraries are available:
- Numpy
- Scipy
- Scikit-learn
- Cython
- LightFM
- Surprise
- msgpack, msgpack-numpy
- nltk, gensim
- keras, tensorflow, pytorch
- tensorrec
- spotlight
- pandas
- vega
- ML evaluation metrics ( See https://github.com/benhamner/Metrics )
To start the jupyter notebook:
docker run -v host_work_dir:/work -p 8888:8888 fred4freedom/recommender-notebook
where host_work_dir is a local directory in the host machine to be mounted into the notebook
and point your browser to http://localhost:8888