Fondant is an open source framework for data preparation and fine-tuning of foundation models, developed by ML6 together with the open source community. Our goal is to make it easy and efficient to fine-tune large foundation models based on specific knowledge domain data.
Connexion is the Python web framework that enables you to build API-first. We have been maintaining it since 2021, working towards a major 3.0 release. With around 5 million downloads per month, it is a staple of the Python API ecosystem.
Apache Beam is a framework for unified batch and streaming pipelines, which is mostly known as the framework underlying Google Cloud Dataflow, but can be run on almost any distributed processing backend such as Apache Flink or Spark. We have ported Apache Beam to Python 3, added ML inference functionality, and created a lot of content for the community.
We leverage Hugging Face a lot in our solutions, but we also contribute back. We have published models, datasets, and demos for the community. Check our space for more.
Github is the home for all our open source code projects. Check our Github page for an overview of all our repositories.