Collected sources and patterns will appear here. Add from search or the patterns library.
Stream<HostData> -> Stream<DeviceTensor>
Overlap host-side data loading with device-side augmentation and model execution using double-buffered queues.
Problem it solves
Synchronous data transfer and CPU-GPU serialization cause GPU starvation.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.