Closed Blacknight318 closed 1 year ago
Acceleration depends on onnxruntime execution providers. onnxruntime supports many different EPs but Coral TPUs are not among them yet. Also see https://github.com/microsoft/onnxruntime/issues/10248.
You could try to convert the ONNX model in deface
into TFLite format and replace the onnxruntime calls with the corresponding pycoral code in your own fork but that is out of scope of this project.
I apologize if this has already been looked at, but would there be a way to add the ability to use a Coral TPU(USB) in place of Nvidia Cuda to speed up processing?