Alpenglow Save

Project README


Build Status Docs Status

  • Alpenglow is a free and open source C++ framework with easy-to-use Python API.
  • Alpenglow is capable of training and evaluating industry standard recommendation algorithms including variants of popularity, nearest neighbor, and factorization models.
  • Besides batch training and evaluation, Alpenglow supports online training of recommendation models capable of adapting to concept drift in non-stationary environments.


  • Anaconda + Python 3.5+

Conda package:

Conda Version Conda Downloads Linux OSX Windows

Installation from conda repositories

  • conda install -c conda-forge alpenglow

In case you also intend to run sample code and tutorials, you should install matplotlib as well:
conda install matplotlib

If you encounter any conflict or error, try installing Alpenglow in a clean conda environment.

Installation from source on Linux

  • cd Alpenglow
  • conda activate [your_cond_env_name]
  • ./
  • conda install libgcc
  • conda install -c conda-forge eigen
  • pip install .

It is also possible on other plaforms to compile from source similarly, however we don't maintain exact instructions here. For reference on how the official binaries are built, please see the corresponding feedstock.

Getting Started

from alpenglow.experiments import FactorExperiment
from alpenglow.evaluation import DcgScore
import pandas as pd
import matplotlib
import matplotlib.pyplot as plt

data = pd.read_csv("")

factor_model_experiment = FactorExperiment(

rankings =, verbose=True)
rankings['dcg'] = DcgScore(rankings)
day = 86400
averages = rankings['dcg'].groupby((rankings['time']-rankings['time'].min())//day).mean()


  • For faster recompilation, use export CC="ccache cc"
  • E.g. to enable compilation on 4 threads, use echo 4 > .parallel
  • Reinstall modified version using pip install --upgrade --force-reinstall --no-deps .
  • To build and use in the current folder, use pip install --upgrade --force-reinstall --no-deps -e . and export PYTHONPATH="$(pwd)/python:$PYTHONPATH". You will also need to install the local sip module to this directory: ./ --destdir $(pwd)/python
Open Source Agenda is not affiliated with "Alpenglow" Project. README Source: rpalovics/Alpenglow
Open Issues
Last Commit
1 year ago

Open Source Agenda Badge

Open Source Agenda Rating