avoid Pylint 'torch' has no 'from_numpy' member error

torch has no memer function from_numpy because it is actually _C.from_numpy. This induces an error when running lint. See pytorch/pytorch#701
qiskit-community · Aug 9, 2019 · ba17e45 · ba17e45
1 parent 4f5ec89
commit ba17e45
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 10 deletions.
diff --git a/qiskit/aqua/components/neural_networks/numpy_discriminator.py b/qiskit/aqua/components/neural_networks/numpy_discriminator.py
@@ -39,10 +39,9 @@ def __init__(self, n_features=1, n_out=1):
             n_out: int, output dimension
         """
         self.architecture = [
-            {"input_dim": n_features, "output_dim": 4, "activation": "leaky_relu"},
-            {"input_dim": 4, "output_dim": 156, "activation": "leaky_relu"},
-            {"input_dim": 156, "output_dim": 52, "activation": "leaky_relu"},
-            {"input_dim": 52, "output_dim": n_out, "activation": "sigmoid"},
+            {"input_dim": n_features, "output_dim":50, "activation": "leaky_relu"},
+            {"input_dim": 50, "output_dim": 20, "activation": "leaky_relu"},
+            {"input_dim": 20, "output_dim": n_out, "activation": "sigmoid"},
         ]
 
         self.parameters = []
@@ -229,7 +228,7 @@ def __init__(self, n_features=1, n_out=1):
         self._n_features = n_features
         self._n_out = n_out
         self._discriminator = DiscriminatorNet(self._n_features, self._n_out)
-        self._optimizer = ADAM(maxiter=1, tol=1e-6, lr=1e-5, beta_1=0.7, beta_2=0.99, noise_factor=1e-4,
+        self._optimizer = ADAM(maxiter=1, tol=1e-6, lr=1e-3, beta_1=0.7, beta_2=0.99, noise_factor=1e-4,
                                eps=1e-6, amsgrad=True)
 
         self._ret = {}

diff --git a/qiskit/aqua/components/neural_networks/pytorch_discriminator.py b/qiskit/aqua/components/neural_networks/pytorch_discriminator.py
@@ -278,7 +278,8 @@ def train(self, data, weights, penalty=True, quantum_instance=None, shots=None):
         Returns: dict, with Discriminator loss (torch.Tensor) and updated parameters (array).
 
         """
-
+        # pylint: disable=E1101
+        # pylint: disable=E1102
         # Reset gradients
         self._optimizer.zero_grad()
         real_batch = data[0]
@@ -311,7 +312,8 @@ def train(self, data, weights, penalty=True, quantum_instance=None, shots=None):
 
         if penalty:
             self.gradient_penalty(real_batch).backward()
-
+        # pylint: enable=E1101
+        # pylint: enable=E1102
         # Update weights with gradients
         self._optimizer.step()
 

diff --git a/qiskit/aqua/components/neural_networks/quantum_generator.py b/qiskit/aqua/components/neural_networks/quantum_generator.py
@@ -126,8 +126,8 @@ def __init__(self, bounds, num_qubits, generator_circuit=None, init_params=None,
             else:
                 raise AquaError('Set univariate variational distribution to represent univariate data')
         # Set optimizer for updating the generator network
-        self._optimizer = ADAM(maxiter=1, tol=1e-6, lr=1e-5, beta_1=0.9, beta_2=0.99, noise_factor=1e-8,
-                               eps=1e-10, amsgrad=True, snapshot_dir=snapshot_dir)
+        self._optimizer = ADAM(maxiter=1, tol=1e-6, lr=1e-3, beta_1=0.7, beta_2=0.99, noise_factor=1e-6,
+                               eps=1e-6, amsgrad=True, snapshot_dir=snapshot_dir)
 
         if np.ndim(self._bounds) == 1:
             bounds = np.reshape(self._bounds, (1, len(self._bounds)))

diff --git a/test/aqua/test_qgan.py b/test/aqua/test_qgan.py
@@ -107,7 +107,6 @@ def setUp(self):
         g_circuit = UnivariateVariationalDistribution(sum(num_qubits), var_form, init_params,
                                                       low=self._bounds[0],
                                                       high=self._bounds[1])
-        # initial_distribution=init_distribution,
         # Set quantum generator
         self.qgan.set_generator(generator_circuit=g_circuit)