Classification always giving same results

[Gustavo-Jodar]

Hi, I am trying to use the tensorflow implementation with intel sgx, but I keep getting the same outputs, regardless of the input values.

My model is in float32:

not quantized

Im saving it in a .tflite format
And also the inputs are saved like this: .astype('float32').tofile('input_tree.bin')

You can see the code from google colab that I use to generate the model and the data:
https://colab.research.google.com/drive/1yRhkVwxm100AaxGmSNco4AW26Uq_MOEd#scrollTo=mZMuiKZZwOju

The enclave starts and finishes with 0 problems, the inference is done with no errors, but always with the same results (different from Google Colab's code). I suppose that the input or the model are not well formatted. Is there a way you could show how to generate a proper model and input ?

Thank-you very much in advance, this is a very good project.