- Sort Score
- Result 10 results
- Languages All
Results 1 - 2 of 2 for CreateEmbeddingPipeliningPass (0.23 sec)
-
tensorflow/compiler/mlir/tensorflow/transforms/sparsecore/sparsecore_passes.h
// This is a strictly sequential and formally correct fallback option for the // embedding pipelining pass intended for debugging during pipelining // development. std::unique_ptr<OperationPass<ModuleOp>> CreateEmbeddingPipeliningPass(); // Passes in the program key to embedding ops, by moving the embedding ops // after the _TPUCompileMlir op. std::unique_ptr<OperationPass<func::FuncOp>> CreateEmbeddingProgramKeyPass();
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Mar 28 23:42:09 UTC 2024 - 2.1K bytes - Viewed (0) -
tensorflow/compiler/mlir/tensorflow/transforms/sparsecore/sparsecore_passes.td
include "mlir/Pass/PassBase.td" def EmbeddingPipeliningPass : Pass<"tf-embedding-pipelining", "mlir::ModuleOp"> { let summary = "Rewrite graph for embedding pipelining"; let constructor = "TFDevice::CreateEmbeddingPipeliningPass()"; let description = [{ For architectures that support accelerated embedding lookups, this pass will rewrite the graph to use pipelining for better device utilization. }]; }
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Mar 28 23:42:09 UTC 2024 - 3.9K bytes - Viewed (0)