def layer_starting_weights(self):
random_weights = np.append(
np.random.normal(loc=0, scale=1, size=self.n_scaling_weights),
np.random.uniform(high=4 * np.pi,
size=self.n_rotation_weights +
self.n_entangling_weights))
return random_weights
def successive_params(par1, par2):
"""Return a list of parameter configurations, successively walking from
par1 to par2 coordinate-wise."""
par1_flat = np.fromiter(_flatten(par1), dtype=float)
par2_flat = np.fromiter(_flatten(par2), dtype=float)
walking_param = []
for i in range(len(par1_flat) + 1):
walking_param.append(unflatten(np.append(par2_flat[:i], par1_flat[i:]), par1))
return walking_param
def gen_data(thresh):
#thresh = 0.3
X = np.array([])
Y = np.array([])
ctr = 0 # num valid data pts
maxval = 0.0
minval = 0.0
np.random.seed(0)
while ctr < 40:
x = np.random.rand(N) * 2 * np.pi
for pair in itertools.combinations(range(N), r=2):
x = np.append(x, (np.pi - x[pair[0]]) * (np.pi - x[pair[1]]))
y = []
for i in range(N):
y.append(data_label(x, i=i))
y_prod = reduce((lambda x, y: x * y), y)
#print(y, y_prod)
if (y_prod > maxval):
maxval = y_prod
print("new max separation: ", maxval)
elif (y_prod < minval):
minval = y_prod
print("new min separation: ", minval)
if y_prod > thresh:
Y = np.append(Y, +1)
X = np.append(X, x)
ctr += 1
#print("+1")
elif y_prod < -1 * thresh:
Y = np.append(Y, -1)
X = np.append(X, x)
ctr += 1
#print("-1")
X = X.reshape(-1, 3)
print("Data: ", list(zip(X, Y)))
return X, Y
def pad_data(data: np.ndarray, axis: int, padding: int) -> np.ndarray:
"""
Function pads 0 data to the end of the given axis.
:param data: The data to pad 0 data to
:param axis: Axis to pad on
:param padding: How many elements to pad
"""
size = data.shape[0]
if size == 0:
return data.reshape((size, data.shape[axis] + padding))
return np.append(data, [[0] * padding for _ in range(size)], axis)
def layer_starting_weights(self):
random_weights = np.append(
np.random.normal(loc=0, scale=1, size=self.n_qubits),
np.random.uniform(low=0, high=4 * np.pi, size=self.n_qubits))
return random_weights
def _feature_padding(self, features):
res = features
for i in range(self.n_repeats_per_layer-1):
res = np.append(res, features)
return np.append(res, np.zeros(self.padding_width))