We have a new release Recommenders 1.2.0!
So many changes since our last release. We have full tests on Python 3.8 to 3.11 (around 1800 tests), upgraded performance in many algorithms, reviewed notebooks, and many more improvements.
We are pleased to announce that this repository (formerly known as Microsoft Recommenders, https://github.com/microsoft/recommenders), has joined the Linux Foundation of AI and Data (LF AI & Data)! The new organization, recommenders-team, reflects this change.
We hope this move makes it easy for anyone to contribute! Our objective continues to be building an ecosystem and a community to sustain open source innovations and collaborations in recommendation systems.
Now to access the repo, instead of going to https://github.com/microsoft/recommenders, you need to go to https://github.com/recommenders-team/recommenders. The old URL will still resolve to the new one, but we recommend that you update your bookmarks.
We moved to a new organization! Now to access the repo, instead of going to https://github.com/microsoft/recommenders, you need to go to https://github.com/recommenders-team/recommenders. The old URL will still resolve to the new one, but we recommend you to update your bookmarks.
We reached 15,000 stars!!
We have a new release Recommenders 1.1.1!
We have introduced a new way of testing our repository using AzureML. With AzureML we are able to distribute our tests to different machines and run them in parallel. This allows us to test our repository on a wider range of machines and provides us with a much faster test cycle. Our total computation time went from around 9h to 35min, and we were able to reduce the costs by half. See more details here.
We also made other improvements like faster evaluation metrics and improving SAR algorithm.
We have a new release Recommenders 1.1.0! We have introduced the SASRec and SSEPT algorithms that are based on transformers. In addition, we now have enabled Python 3.8 and 3.9. We have also made improvements on the SARPlus algorithm, including support for Azure Synapse and Spark 3.2. There are also bug fixes and improvements on NCF, RBM, LightGBM, LightFM, Scikit-Surprise, the stratified splitter, dockerfile and upgrade to Scikit-Learn 1.0.2.
We have a new release Recommenders 1.0.0! The codebase has now migrated to TensorFlow versions 2.6 / 2.7 and to Spark version 3. In addition, there are a few changes in the dependencies and extras installed by pip (see this guide). We have also made improvements in the code and the CI / CD pipelines.
We have a new release Recommenders 0.7.0!
In this, we have changed the names of the folders which contain the source code, so that they are more informative. This implies that you will need to change any import statements that reference the recommenders package. Specifically, the folder reco_utils has been renamed to recommenders and its subfolders have been renamed according to issue 1390.
The recommenders package now supports three types of environments: venv, virtualenv and conda with Python versions 3.6 and 3.7.
We have also added new evaluation metrics: novelty, serendipity, diversity and coverage (see the evalution notebooks).
Code coverage reports are now generated for every PR, using Codecov.
We have a new release Recommenders 0.6.0!
Recommenders is now on PyPI and can be installed using pip! In addition there are lots of bug fixes and utilities improvements.
Here you can find the PyPi page: https://pypi.org/project/recommenders/
Here you can find the package documentation: https://microsoft-recommenders.readthedocs.io/en/latest/
We have surpassed 10k stars!
Microsoft Recommenders repository has reached 10k stars and has become the most starred open-source recommender system project on GitHub.
Many thanks and congratulations to all the contributors to this repository! More advanced algorithms and best practices are yet to come!
We have a new release Recommenders 0.5.0!
It comes with lots of bug fixes, optimizations and 3 new algorithms, GeoIMC, Standard VAE and Multinomial VAE. We also added tools to facilitate the use of Microsoft News dataset (MIND). In addition, we published our KDD2020 tutorial where we built a recommender of COVID papers using Microsoft Academic Graph.
We also changed the default branch from master to main. Now when you download the repo, you will get main branch.
Leaderboard Reopen!
Microsoft News Recommendation Competition Winners Announced
Congratulations to all participants and winners of the Microsoft News Recommendation Competition! In the last two months, over 200 participants from more than 90 institutions in 19 countries and regions joined the competition and collectively advanced the state of the art of news recommendation.
The competition is based on the recently released MIND dataset, an open, large-scale English news dataset with impression logs. Details of the dataset are available at this ACL paper.
With the competition successfully closed, the leaderboard is now reopn. Want to see if you can grab the top spot? Get familiar with the news recommendation scenario. Then dive into some baselines such as DKN, LSTUR, NAML, NPA and NRMS and start hacking!
Microsoft News Recommendation Competition Winners Announced, Leaderboard to Reopen!
Congratulations to all participants and winners of the Microsoft News Recommendation Competition! In the last two months, over 200 participants from more than 90 institutions in 19 countries and regions joined the competition and collectively advanced the state of the art of news recommendation.
The competition is based on the recently released MIND dataset, an open, large-scale English news dataset with impression logs. Details of the dataset are available at this ACL paper.
With the competition successfully closed, the leaderboard will reopen soon. Want to see if you can grab the top spot? Get familiar with the news recommendation scenario. Then dive into some baselines such as DKN, LSTUR, NAML, NPA and NRMS and get ready!
Microsoft is hosting a News Recommendation competition based on the MIND dataset, a large-scale English news dataset with impression logs. Check out the ACL paper, get familiar with the news recommendation scenario, and dive into the quick start example using the DKN algorithm. Then try some other algorithms (NAML, NPA, NRMS, LSTUR) and tools in recommenders and submit your entry!
New release: Recommenders 0.4.0
13 new algos and multiple fixes and new features
New release: Recommenders 0.3.1
We reached 5000 stars!!
New release: Recommenders 0.3.0
New release: Recommenders 0.2.0
We reached 1000 stars!!
First release: Recommenders 0.1.1
First pre-release: Recommenders 0.1.0