Tensorflow error on Android

Question

I am learning Tensorflow and following a tutorial I was able to make a custom model to run it in an Android App but I am having problems with it. I have the following code:

    public void testModel(Context ctx) {
        String model_file = "file:///android_asset/model_graph.pb";
        int[] result = new int[2];
        float[] input = new float[]{0.0F, 1.0F, 0.0F, 1.0F, 1.0F, 0.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 1.0F, 0.0F, 0.0F, 0.0F, 0.0F, 1.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 0.0F, 1.0F, 1.0F, 0.0F, 0.0F, 0.0F, 1.0F, 0.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 1.0F, 0.0F, 0.0F, 1.0F, 0.0F, 0.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 1.0F, 0.0F, 1.0F, 0.0F, 0.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F, 1.0F, 0.0F};
        TensorFlowInferenceInterface inferenceInterface;
        inferenceInterface = new TensorFlowInferenceInterface(ctx.getAssets(), model_file);
        inferenceInterface.feed("input", input, 68);
        inferenceInterface.run(new String[]{"output"});
        inferenceInterface.fetch("output", result);
        Log.v(TAG, Arrays.toString(result));
    }

I got an error when the app try to run the inferenceInterface.run(new String[]{"output"}) method:

java.lang.IllegalArgumentException: In[0] is not a matrix
[[Node: MatMul = MatMul[T=DT_FLOAT, transpose_a=false, transpose_b=false, _device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_input_0_0, W1)]]

I don't believe the model I created is the problem because I was able to use it in a Python code with positive result.

ash · Accepted Answer · 2017-11-14 06:26:29Z

1

From the error message (In[0] is not a matrix), it appears that your model requires the input to be a matrix (i.e., a two dimensional tensor), while you are feeding a one dimensional tensor (vector) with 68 elements.

In particular, the dims argument to TensorFlowInferenceInterface.feed seems incorrect in the line:

inferenceInterface.feed("input", input, 68);

Instead, it should be something like:

inferenceInterface.feed("input", input, 68, 1);

if your model expects a 68x1 matrix (or 34, 2 if it expects a 34x2 matrix, 17, 4 for a 17x4 matrix etc.)

Hope that helps.

answered Nov 14, 2017 at 6:26

ash

6,7714 gold badges20 silver badges30 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Fernando Berra Over a year ago

Yes, the problem was that the model was expecting a matrix not an array. I modified the model a op tf.reshape,that way I could set an input of an array and the model changes it to a matrix later

Collectives™ on Stack Overflow

Tensorflow error on Android

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related