Functionality to call high-level code from C++.

Programs/Source/breast_logistic.mpc

@@ -22,16 +22,17 @@ elif 'vertical' in program.args:
     b = sfix.input_tensor_via(1, X_train[:,X_train.shape[1] // 2:])
     X_train = a.concat_columns(b)
     y_train = sint.input_tensor_via(0, y_train)
-elif 'party0' in program.args:
-    a = sfix.input_tensor_via(0, X_train[:,:X_train.shape[1] // 2])
-    b = sfix.input_tensor_via(1, shape=X_train[:,X_train.shape[1] // 2:].shape)
+elif 'party0' in program.args or 'party1' in program.args:
+    party = int('party1' in program.args)
+    a = sfix.input_tensor_via(
+        0, X_train[:,:X_train.shape[1] // 2] if party == 0 else None,
+        shape=X_train[:,:X_train.shape[1] // 2].shape)
+    b = sfix.input_tensor_via(
+        1, X_train[:,X_train.shape[1] // 2:] if party == 1 else None,
+        shape=X_train[:,X_train.shape[1] // 2:].shape)
     X_train = a.concat_columns(b)
-    y_train = sint.input_tensor_via(0, y_train)
-elif 'party1' in program.args:
-    a = sfix.input_tensor_via(0, shape=X_train[:,:X_train.shape[1] // 2].shape)
-    b = sfix.input_tensor_via(1, X_train[:,X_train.shape[1] // 2:])
-    X_train = a.concat_columns(b)
-    y_train = sint.input_tensor_via(0, shape=y_train.shape)
+    y_train = sint.input_tensor_via(0, y_train if party == 0 else None,
+                                    shape=y_train.shape)
 else:
     X_train = sfix.input_tensor_via(0, X_train)
     y_train = sint.input_tensor_via(0, y_train)

Programs/Source/export-a2b.py

@@ -0,0 +1,7 @@
+@export
+def a2b(x, res):
+    print_ln('x=%s', x.reveal())
+    res[:] = sbitvec(x, length=16)
+    print_ln('res=%s', x.reveal())
+
+a2b(sint(size=10), sbitvec.get_type(16).Array(10))

Programs/Source/export-b2a.py

@@ -0,0 +1,7 @@
+@export
+def b2a(res, x):
+    print_ln('x=%s', x.reveal())
+    res[:] = sint(x[:])
+    print_ln('res=%s', x.reveal())
+
+b2a(sint.Array(size=10), sbitvec.get_type(16).Array(10))

Programs/Source/export-sort.py

@@ -0,0 +1,7 @@
+@export
+def sort(x):
+    print_ln('x=%s', x.reveal())
+    res = x.sort()
+    print_ln('res=%s', x.reveal())
+
+sort(sint.Array(1000))

Programs/Source/export-trunc.py

@@ -0,0 +1,8 @@
+@export
+def trunc_pr(x):
+    print_ln('x=%s', x.reveal())
+    res = x.round(32, 2)
+    print_ln('res=%s', res.reveal())
+    return res
+
+trunc_pr(sint(0, size=1000))

Programs/Source/mnist_full_A.mpc

@@ -63,9 +63,6 @@ else:
 if 'nearest' in program.args:
     sfix.round_nearest = True
 
-if program.options.ring:
-    assert sfix.f * 4 == int(program.options.ring)
-
 debug_ml = ('debug_ml' in program.args) * 2 ** (sfix.f / 2)
 
 if '1dense' in program.args:

Programs/Source/test_flow_optimization.mpc

@@ -21,3 +21,13 @@ test(a, 10000, 10000)
 test(b, 10000, 20000)
 test(a, 1000000, 1000000)
 test(b, 1000000, 2000000)
+
+a = 1
+if True:
+    if True:
+        a = 2
+    if True:
+        a = 3
+else:
+    a = 4
+    crash()

Programs/Source/torch_resnet.py

@@ -41,7 +41,9 @@ secret_input = sfix.input_tensor_via(
 
 layers = ml.layers_from_torch(model, secret_input.shape, 1, input_via=0)
 
-optimizer = ml.Optimizer(layers)
+optimizer = ml.Optimizer(layers, time_layers='time_layers' in program.args)
 
+start_timer(1)
 print_ln('Secure computation says %s',
          optimizer.eval(secret_input, top=True)[0].reveal())
+stop_timer(1)

Programs/Source/torch_vgg.py

@@ -0,0 +1,42 @@
+# this tests the pretrained VGG in secure computation
+
+program.options_from_args()
+
+from Compiler import ml
+
+try:
+    ml.set_n_threads(int(program.args[2]))
+except:
+    pass
+
+import torchvision
+import torch
+import numpy
+import requests
+import io
+import PIL
+
+from torchvision import transforms
+
+name =  'vgg' + program.args[1]
+model = getattr(torchvision.models, name)(weights='DEFAULT')
+
+r = requests.get('https://github.com/pytorch/hub/raw/master/images/dog.jpg')
+input_image = PIL.Image.open(io.BytesIO(r.content))
+input_tensor = transforms._presets.ImageClassification(crop_size=32)(input_image)
+input_batch = input_tensor.unsqueeze(0) # create a mini-batch as expected by the model
+
+with torch.no_grad():
+    output = int(model(input_batch).argmax())
+    print('Model says %d' % output)
+
+secret_input = sfix.input_tensor_via(
+    0, numpy.moveaxis(input_batch.numpy(), 1, -1))
+
+layers = ml.layers_from_torch(model, secret_input.shape, 1, input_via=0)
+
+optimizer = ml.Optimizer(layers)
+optimizer.time_layers = True
+
+print_ln('Secure computation says %s',
+         optimizer.eval(secret_input, top=True)[0].reveal())