MAGICGrants/Monero-Dataset-Pipeline
1
0
Fork 0
mirror of https://github.com/MAGICGrants/Monero-Dataset-Pipeline.git synced 2026-01-09 13:37:57 -05:00
Code Issues Packages Projects Releases Wiki Activity

reduced runs from 10 to 5

Browse Source
This commit is contained in:
ACK-J
2022-08-10 17:15:34 -04:00
parent 7b1c511408
commit 1ed0157962
4 changed files with 17 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
DataScience/GradientBoostedClassifier.py
View File

@@ -112,12 +112,14 @@ def gradient_boosted(X_train, X_test, y_train, y_test, RANDOM_STATE, X_Validatio
NUM_PROCESSES = cpu_count()
if NUM_PROCESSES > 10:
NUM_PROCESSES = 10
Num_Iterations = 5
if NUM_PROCESSES > Num_Iterations:
NUM_PROCESSES = Num_Iterations
with Manager() as manager:
with manager.Pool(processes=NUM_PROCESSES) as pool:
for returned_data in tqdm(pool.imap_unordered(func=run_model_wrapper, iterable=zip(repeat(X_train,10), repeat(X_test,10), repeat(y_train,10), repeat(y_test,10), list(range(10)), repeat(X_Validation,10), repeat(y_Validation,10))), desc="(Multiprocessing) Training GBC", total=10, colour='blue'):
for returned_data in tqdm(pool.imap_unordered(func=run_model_wrapper, iterable=zip(repeat(X_train,Num_Iterations), repeat(X_test,Num_Iterations), repeat(y_train,Num_Iterations), repeat(y_test,Num_Iterations), list(range(Num_Iterations)), repeat(X_Validation,Num_Iterations), repeat(y_Validation,Num_Iterations))), desc="(Multiprocessing) Training GBC", total=Num_Iterations, colour='blue'):
weighted_f1, weighted_f1_mainnet = returned_data
out_of_sample_f1.append(weighted_f1)
mainnet_f1.append(weighted_f1_mainnet)

6
DataScience/NeuralNetwork.py
View File

@@ -89,7 +89,7 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
scaler = StandardScaler().fit(X_Validation)
X_Validation = scaler.transform(X_Validation)
for i in range(10):
for i in range(5):
from keras_visualizer import visualizer
model = Sequential()
model.add(Dense(11, input_shape=(X_train.shape[1],), activation='relu'))
@@ -153,7 +153,7 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
mainnet_f1.append(weighted_f1_mainnet)
if stagenet:
cm = confusion_matrix(y_test, y_pred)
cm = confusion_matrix(y_test_copy, y_pred)
# Heat map
plt.figure(figsize=(10, 7))
sn.heatmap(cm, annot=True)
@@ -161,7 +161,7 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
plt.ylabel('Truth')
plt.savefig("./models/GBC/stagenet/CM_epochs_" + str(EPOCHS) + "_batch_size_" + str(BATCH_SIZE) + "_i_" + str(i) + "_accuracy_" + str(weighted_f1_mainnet) + ".png")
else:
cm = confusion_matrix(y_test, y_pred)
cm = confusion_matrix(y_test_copy, y_pred)
# Heat map
plt.figure(figsize=(10, 7))
sn.heatmap(cm, annot=True)

13
DataScience/RandomForest.py
View File

@@ -19,7 +19,7 @@ LR = .25
N_ESTIMATORS = 700
def run_model(X_train, X_test, y_train, y_test, RANDOM_STATE, X_Validation, y_Validation):
def run_model_rf(X_train, X_test, y_train, y_test, RANDOM_STATE, X_Validation, y_Validation):
global stagenet
model = AdaBoostClassifier(n_estimators=N_ESTIMATORS, random_state=RANDOM_STATE, learning_rate=LR)
# Train the model
@@ -77,8 +77,8 @@ def run_model(X_train, X_test, y_train, y_test, RANDOM_STATE, X_Validation, y_Va
return weighted_f1, weighted_f1_mainnet
def run_model_wrapper(data):
return run_model(*data)
def run_model_wrapper_rf(data):
return run_model_rf(*data)
def random_forest(X_train, X_test, y_train, y_test, N_ESTIMATORS, MAX_DEPTH, RANDOM_STATE, X_Validation, y_Validation, stagenet_val=True):
@@ -88,13 +88,14 @@ def random_forest(X_train, X_test, y_train, y_test, N_ESTIMATORS, MAX_DEPTH, RAN
mainnet_f1 = []
NUM_PROCESSES = cpu_count()
Num_Iterations = 5
if NUM_PROCESSES > 10:
NUM_PROCESSES = 10
if NUM_PROCESSES > Num_Iterations:
NUM_PROCESSES = Num_Iterations
with Manager() as manager:
with manager.Pool(processes=NUM_PROCESSES) as pool:
for returned_data in tqdm(pool.imap_unordered(func=run_model_wrapper, iterable=zip(repeat(X_train, 10), repeat(X_test, 10), repeat(y_train, 10), repeat(y_test, 10), list(range(10)), repeat(X_Validation, 10), repeat(y_Validation, 10))), desc="(Multiprocessing) Training RF", total=10, colour='blue'):
for returned_data in tqdm(pool.imap_unordered(func=run_model_wrapper_rf, iterable=zip(repeat(X_train, Num_Iterations), repeat(X_test, Num_Iterations), repeat(y_train, Num_Iterations), repeat(y_test, Num_Iterations), list(range(Num_Iterations)), repeat(X_Validation, Num_Iterations), repeat(y_Validation, Num_Iterations))), desc="(Multiprocessing) Training RF", total=Num_Iterations, colour='blue'):
weighted_f1, weighted_f1_mainnet = returned_data
out_of_sample_f1.append(weighted_f1)
mainnet_f1.append(weighted_f1_mainnet)

1
requirements.txt
View File

@@ -10,3 +10,4 @@ xgboost
graphviz
colorama
psycopg2-binaray
keras