Distributed Distillation for On-Device Learning

Neural Information Processing Systems 

Transmitting model weights requires huge communication overhead and means only devices with identical model architectures can be included.

Similar Docs  Excel Report  more

TitleSimilaritySource
None found