Design and implement a back-end plugin that implements the vAccel inference operation #20, that supports TensorFlow models.
This plugin will be built on top of the TensorFlow C++ API. It will consume a model that has been previously registered with the session #19, and it will receive in some input data and a buffer in which to place output results.
Internally the plugin will need to transform the input data in TensorFlow tensor data types and pass them to the appropriate TensorFlow function calls.
Design and implement a back-end plugin that implements the vAccel inference operation #20, that supports TensorFlow models.
This plugin will be built on top of the TensorFlow C++ API. It will consume a model that has been previously registered with the session #19, and it will receive in some input data and a buffer in which to place output results.
Internally the plugin will need to transform the input data in TensorFlow tensor data types and pass them to the appropriate TensorFlow function calls.