Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
Model<ONNX> + EPConfig -> PartitionedSession
Partition an ONNX computational graph to route supported subgraphs to a specialized NPU execution provider while falling back to the CPU for unsupported operations.
Problem it solves
NPUs often lack implementation for all mathematical operators in a model graph, causing complete execution failures without fallbacks.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.