jvmncs

import tensorflow as tf
import tf_encrypted as tfe

# define the parties
tfe.set_config(tfe.RemoteConfig({
    "model-owner": "localhost:2222",
    "prediction-client": "1.2.3.4:2222",
    "server0": "1.1.1.1:4444",
    "server1": "1.1.1.2:4444"})

jvmncs

import tensorflow as tf

# generic functions for loading model weights and input data
def provide_weights(): """Load model weights as TensorFlow objects."""
def provide_input(): """Load input data as TensorFlow objects."""

# get model weights/input data (both unencrypted)
w0, b0, w1, b1, w2, b2 = provide_weights()
x = provide_input()

jvmncs

import tensorflow as tf
import tf_encrypted as tfe

# generic remote procedure calls
def provide_weights(): """Load model weights."""
def provide_input(): """Load input data."""
def receive_output(): """Receive and decrypt output."""

# get model weights/input data
# as private tensors from each party

jvmncs

import tensorflow as tf


# generic function stubs
def provide_weights(): """Load model weights."""
def provide_input(): """Load input data."""


# get model weights/input data
# (both unencrypted)

jvmncs

def matmul(prot: Pond,
           x: PondMaskedTensor,
           y: PondMaskedTensor):
    a, a0, a1, alpha_on_0, alpha_on_1 = x.unwrapped
    b, b0, b1, beta_on_0, beta_on_1 = y.unwrapped

    with tf.device(prot.crypto_producer.device_name):
        ab = a.matmul(b)
        ab0, ab1 = prot._share(ab)

jvmncs

import tensorflow as tf
import syft

hook = syft.TensorFlowHook(tf)

jvmncs

alice = syft.VirtualWorker(hook, “alice”)

x = tf.constant([1., 2., 3., 4.])
x_ptr = x.send(alice)

print(x_ptr)
# ==> (Wrapper)>[PointerTensor | me:random_id1 -> alice:random_id2]

jvmncs

y_ptr = x_ptr + x_ptr
y = tf.reshape(y_ptr, shape=[2, 2])
id = tf.constant([[1., 0.], [0., 1.]]).send(alice)
z = tf.matmul(y, id).get()

print(z)  
# ==> tf.Tensor([[2. 4.]
#                [6. 8.]], shape=(2, 2), dtype=float32)

jvmncs

x = tf.expand_dims(id[0], 0)

# Initialize the weight
w_init = tf.initializers.glorot_normal()
w = tf.Variable(w_init(shape=(2, 1), dtype=tf.float32)).send(alice)
z = tf.matmul(x, w)

# Manual differentiation & update
dzdx = tf.transpose(x)
w.assign_sub(dzdx)

jvmncs

mnist = tf.keras.datasets.mnist
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

# Converting the data from numpy to tf.Tensor in order to have PySyft functionalities.
x_train, y_train = tf.convert_to_tensor(x_train), tf.convert_to_tensor(y_train)
x_test, y_test = tf.convert_to_tensor(x_test), tf.convert_to_tensor(y_test)

# Send data to Alice (for demonstration purposes)
x_train_ptr = x_train.send(alice)