Skip to content

Refactor Noisy Parity dataset generation into a function #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
BooBSD wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Refactor Noisy Parity dataset generation into a function #4

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 38 additions & 32 deletions examples/NoisyParityData.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,35 +1,41 @@
import numpy as np


noise = 0.2
number_of_features = 12
number_of_variables = 4
number_of_examples = 20000

X_train = np.random.randint(2, size=(number_of_examples, number_of_features), dtype=np.uint32)
Y_train = np.zeros(number_of_examples, dtype=np.uint32)

for i in range(number_of_examples):
for j in range(number_of_features):
X_train[i, j] = np.random.randint(2)

set_bit_count = 0
for j in range(number_of_variables):
set_bit_count += X_train[i, j * number_of_features // number_of_variables:j * number_of_features // number_of_variables + 2].sum()
Y_train[i] = set_bit_count % 2

Y_train = np.where(np.random.rand(number_of_examples) <= noise, 1-Y_train, Y_train) # Adds noise
np.savetxt("examples/NoisyParityTrainingData.txt", np.append(X_train, Y_train.reshape((number_of_examples, 1)), axis=1), fmt='%d')

X_test = np.random.randint(2, size=(number_of_examples, number_of_features), dtype=np.uint32)
Y_test = np.zeros(number_of_examples, dtype=np.uint32)
for i in range(number_of_examples):
for j in range(number_of_features):
X_test[i, j] = np.random.randint(2)

set_bit_count = 0
for j in range(number_of_variables):
set_bit_count += X_test[i, j * number_of_features // number_of_variables:j * number_of_features // number_of_variables + 2].sum()
Y_test[i] = set_bit_count % 2

np.savetxt("examples/NoisyParityTestingData.txt", np.append(X_test, Y_test.reshape((number_of_examples, 1)), axis=1), fmt='%d')
NOISE = 0.2

NUMBER_OF_FEATURES = 12
NUMBER_OF_VARIABLES = 4
NUMBER_OF_EXAMPLES = 20_000

# NUMBER_OF_FEATURES = 24
# NUMBER_OF_VARIABLES = 8
# NUMBER_OF_EXAMPLES = 100_000


def generate_dataset(
number_of_features: int,
number_of_variables: int,
number_of_examples: int,
noise: float = 0.0,
) -> tuple[np.ndarray, np.ndarray]:
X = np.random.randint(2, size=(number_of_examples, number_of_features), dtype=np.uint32)
Y = np.zeros(number_of_examples, dtype=np.uint32)

for i in range(number_of_examples):
set_bit_count = 0
for j in range(number_of_variables):
start = j * number_of_features // number_of_variables
set_bit_count += X[i, start : start + 2].sum()
Y[i] = set_bit_count % 2

if noise > 0:
Y = np.where(np.random.rand(number_of_examples) <= noise, 1 - Y, Y) # Adds noise

return X, Y


X_train, Y_train = generate_dataset(NUMBER_OF_FEATURES, NUMBER_OF_VARIABLES, NUMBER_OF_EXAMPLES, noise=NOISE)
X_test, Y_test = generate_dataset(NUMBER_OF_FEATURES, NUMBER_OF_VARIABLES, NUMBER_OF_EXAMPLES)

np.savetxt("examples/NoisyParityTrainingData.txt", np.hstack([X_train, Y_train.reshape(-1, 1)]), fmt='%d')
np.savetxt("examples/NoisyParityTestingData.txt", np.hstack([X_test, Y_test.reshape(-1, 1)]), fmt='%d')