Missing Interpreter set_tensor method

[ghost]

I am trying to convert a python implementation to be used in Flutter.

This is a function definition I am unable to find interpreter.set_tensor it is not part the flutter-tflite Interpreter API set_tensor is defined here https://www.tensorflow.org/api_docs/python/tf/lite/Interpreter#set_tensor

def tflite_predict(tflite_interpreter, input_index, output_index, x):
    tflite_interpreter.set_tensor(input_index, x)
    tflite_interpreter.invoke()
    return tflite_interpreter.get_tensor(output_index)[None, ]

Here is what I have

  Tensor tfliteMissPredict(
    Interpreter interpreter,
    int inputIndex,
    int outputIndex,
    dynamic data,
  ) {
    // interpreter.setTensor(data) <--- Not part of the API
    interpreter.invoke();
    return interpreter.getOutputTensor(outputIndex);
  }

Is there another approach for this case?

[saurabhkumar8112]

I don't think you need to do that step. After _interpreter.allocateTensors(); interpreter.getOutputTensor should work fine.

[ghost]

I don't think you need to do that step. After _interpreter.allocateTensors(); interpreter.getOutputTensor should work fine.

There is data passed into tfliteMissPredict and the function is called in a stream many times very quickly with the input index being updated with each call.

I have updated the dart code to reflect the data being passed in.

[lyngbach]

Any luck on solving this or finding another approach? Currently facing the same issue.