The Flower team is excited to announce the release of Flower 1.44 stable, packed with new updates!
Flower is a friendly framework for collaborative AI and data science. It makes novel approaches such as federated learning, federated evaluation, federated analytics, and fleet learning accessible to a wide audience of researchers and engineers.
Thanks to our contributors
We would like to give our special thanks to all the contributors who made the new version of Flower possible (in git shortlog order):
Charles Beauville, Daniel J. Beutel, Daniel Nata Nugraha, Heng Pan, Javier, Taner Topal
What's new?
-
Improve agent runtime infrastructure (#8444, #8453, #8477, #8481, #8484, #8485, #8494, #8498, #8500, #8507, #8511, #8512, #8517, #8521, #8523, #8525, #8527, #8528, #8532, #8533, #8534, #8536, #8545, #8546, #8547, #8548, #8550, #8551, #8552, #8563)
AgentApps can use compacted context from earlier runs and access tools from multiple connected accounts. Experimental file storage also lets AgentApps read and write files scoped to the current agent.
-
Start work faster on prewarmed workers (#8482, #8487, #8490)
Prewarmed AgentApp and connector workers now start handling tasks sooner because the handoff loads fewer Python modules. In local Kubernetes benchmarks, median task handoff time fell from 480 ms to 101 ms.
-
General improvements (#8469, #8480, #8483, #8489, #8497, #8503, #8504, #8505, #8506, #8508, #8509, #8513, #8514, #8516, #8520, #8522, #8530, #8539, #8554, #8555, #8556, #8558)
As always, many parts of the Flower framework and quality infrastructure were improved and updated.
