The weather may be getting colder in the northern hemisphere, but because our team is a dedicated bunch of awesomeness, we keep announcing hot stuff.
Just a few days ago, we announced that Sanofi has joined the AI Structural Biology Consortium, contributed the industry perspective on federated learning in life sciences to Cell Patterns with our friends at Roche and NVIDIA… and today we are pleased to announce the release of Apheris 3.4.
With today's release, we continue our journey to provide the simplest, most secure solution for building and joining life sciences and drug discovery data networks.
TL;DR version of this release:
Reduced cognitive load for data scientists: Idle Compute Specifications (Compute Specs) are now deactivated automatically for peace of mind and to minimize resource usage.
Closer to a self-serve Model Registry: You can now update your own models in the Model Registry.
Smoother data science experience: A host of improvements to make your day-to-day experience easier and more efficient.
Good to know more: You can now access more details about compute specifications within the Governance Portal.
Let’s dive into the nitty gritty
This release is packed with new features and enhancements, specifically designed to improve the data science experience.
New features include
Environment Culling has been introduced to allow the system to automatically deactivate idle Compute Specs (without any Jobs running for a certain amount of time). This eliminates the need for the Data Scientist to remember to deactivate Compute Specs that are no longer in use. This allows the system to free up resources when they are no longer needed. See documentation for more details.
You can now add new versions to an existing model in the Model Registry. Prior to 3.4, your Model Registry was entirely managed by the Apheris team. By using the Apheris CLI models add-version command, you can now be more self-sufficient with your model building workflow. Find more details in the documentation.
More information is shown for Compute Specs. You can now access the Activation Status of a Compute Spec as well as the number of jobs running for that Compute Spec. All neatly accessible through the Governance Portal.
In addition to these new features, we also smoothed out a few things
Dataset descriptions support full markdown: Data Scientists can never see the raw data, so a clear and easy-to-understand description of the dataset is very helpful. Now you can provide a detailed, nicely formatted description of any dataset with full markdown support.
Faster and easier daily work with the Apheris CLI: We reworked parts of the Apheris CLI authentication and now you only have to login once (a long awaited improvement, we know) and while we were at it, we also speeded up things like getting details on Compute Specs via apheris compute list.
If you are looking for a new adventure and want to help the world collaborate on data safely, join us!
We are looking for a Senior Deployment Engineer and a Senior ML Field Engineer: Drug Discovery and Life Sciences
See you, latest with Apheris 3.5
