Details
-
Improvement
-
Status: To Do
-
Major
-
Resolution: Unresolved
-
None
-
None
Description
The new subgraph API is a natural fit for TensorRT. To help make the codebase consistent we'd like to port the current TensorRT integration to use the new API. The current experimental integration into MXNet requires us to use contrib API calls. Once integration has moved to use the subgraph API users will be able to use TensorRT with a consistent API. Porting should also enable acceleration of gluon and module base models.
As a MXNet core developer I would like to see TensorRT support migrated to the subgraph API to ease the maintenance burden of the feature.
As an MXNet user I would like to be able to make use of TensorRT support without using custom, non-forward compatible APIs.