Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
PipelineConfig, InputTensors -> OutputTensors
Route the output tensors of one model directly into the input buffers of another within shared memory.
Problem it solves
Returning intermediate tensors to clients between pipeline steps wastes network bandwidth and introduces latency.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.