Run onnxruntime with uint8 input

Question

Is there any possible way to run onnxruntime session with uint8 numpy array input?

[ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Unexpected input data type. Actual: (tensor(uint8)) , expected: (tensor(float))

Above comment pops up everytime. Thx in advance!

Answer 1

Looks like the error came from the model which expects float type but you provided uint8. It is not a problem of onnxruntime. In onnxruntime, a session is binded to a model.

To know what data type a model expects, netron can be used to check it.

Here documents the operators and data types supported by onnxruntime.