ngraph-bridge
ngraph-bridge copied to clipboard
TensorFlow-nGraph bridge
I tried using script of image label classification from examples folder where I tried to compare two scripts one with ngraph bridge configuration and other without it, I found that...
* Unconditionally enable "ngraph-optimizer" grappler pass * Remove is_grappler_enabled() API call * Run tests using the grappler pass
This PR includes the new backend engine. It separates the OpenVINO execution from the Executable into an isolated class, IEBackendEngine. This will create an abstraction for integrating other device based...
Use replace_node_update_name to preserve friendly names of replaced nodes in the transpose sinking pass.
Avoid usage of `get_shape()` in `Size` and `Split` op so that they can work for dynamic inputs
Transpose sinking should copy/preserve the friendly name of nodes that it replaces or materializes whenever possible.