Tensorflow Lite in Android Studio for Image Classfication

Have a doubt in java part with using the interpreter. I got the same Result whatever feed any images as an input. I follow the below steps:
1.Load my Tflite and Label FIle from Asset Folder with using Interpreter
2.Convert my Bitmap to Float32 (float)
3. After feed the byte buffer and output label to the interpreter
4. Got Output without Exception But the Outputs are the same whatever feed any inputs.

Please, suggest me which step I did a mistake. And how to solve it…

I’m a bit confused, are you actually using Deeplearning4j? The way you are describing this, it sounds like you are actually using tensorflow lite and not our tensorflow import.

Obviously mistake nr. 1 is posting here. We don’t support TensorFlow users, we support DL4J users :slight_smile:

Oh, But I get some ideas about Tensorflow and numpy from http://nd4j.org/.

Deeplearning4J, or rather its sub-project SameDiff does support import of tensorflow models. Unfortunately our time is limited, so unless you are actually using one of our projects, we can’t help you here.

Ok, Thanks for your response