Polyaxon is an open source, self-hosted, platform for model management, and machine learning (ML) and deep learning(DL) workload orchestration on Kubernetes.
Versioning
In our previous blog post we briefly mentioned that Polyaxon is moving from it’s current release methodology to semantic versioning, but we did not develop enough about the reasoning behind this update. We have now a new section on our documentation website with more information, this new documentation section will be a living document, and will get updated whenever we change something about our versioning mechanism or release cycle.
Roadmap
Although we have been discussing features and platform improvements either publicly through github issues or privately through email with our enterprise customers. We were always asked by several of our users to publish a roadmap outlining major product features and product direction.
Today we are opening a very short-term version of our roadmap about our open source offering, to give our community users an idea about where the product is heading. With this new document we would like to involve our community edition users as well as our enterprise customers in shaping the future of Polyaxon. Our enterprise users will keep being involved in prioritizing features of our products and our enterprise edition, and will have access to a much long-term roadmap.
Guides and resources
In addition to this new product updates, we published several guides, integrations, and resources about deploying and using Polyaxon, some notable guides are about manifests and documentation to deploy High Available Postgres database on AWS, Azure, and GCP to use with Polyaxon, a new deployment guide fro Minikube, a new upcoming guide to deploy Polyaxon on microK8S, an integration with the Pytorch Ignite library, an upcoming integration with the fastai library, a better tracking documentation, and a new guide on how to use Python for parsing and loading polyaxonfiles.
