Skip to content

Commit

Permalink
Merge pull request #2 from anh9895/main
Browse files Browse the repository at this point in the history 
Update all models based on Pytorch. Remove the dependent of skorch library
  • Loading branch information
@thieu1995
thieu1995 authored Nov 2, 2024
2 parents 7f4188f + 60e1fdb commit 862bfbb
Show file tree
Hide file tree
Showing 30 changed files with 2,157 additions and 2,866 deletions.
297 changes: 2 additions & 295 deletions examples/core/exam_base_mlp.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,11 +4,7 @@
# Github: https://github.com/thieu1995 %
# --------------------------------------------------%

from sklearn.datasets import load_iris, load_breast_cancer, load_diabetes, make_regression
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import accuracy_score
from metaperceptron import CustomMLP, MlpClassifier, MlpRegressor
from metaperceptron.core.base_mlp import CustomMLP


def check_CustomMLP_class():
Expand All @@ -23,293 +19,4 @@ def check_CustomMLP_class():
print(model)


def check_MlpClassifier_multi_class():
# Load and prepare the dataset
data = load_iris()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

net = MlpClassifier(hidden_layers=(100, ), act_names="ReLU", dropout_rates=None, act_output=None,
epochs=10, batch_size=16, optim="Adam", optim_paras=None,
early_stopping=True, n_patience=10, epsilon=0.001, valid_rate=0.1,
seed=42, verbose=True)
net.fit(X_train, y_train)
res = net.score(X_test, y_test)
print(res)
print(net.model)
print(net.model.get_weights())


def check_MlpClassifier_multi_class_gridsearch():
# Load and prepare the dataset
data = load_iris()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# Define parameters for grid search
param_grid = {
'hidden_layers': [(30,), (50,), [20, 10], (50, 20)],
'act_names': ["ReLU", "Tanh", "Sigmoid"],
'dropout_rates': [0.2, 0.3, None],
'epochs': [10, 20],
'batch_size': [16, 24],
'optim': ['Adam', 'SGD'],
"early_stopping": [True],
"valid_rate": [0.1, 0.2],
}

# Hyperparameter tuning with GridSearchCV
searcher = GridSearchCV(estimator=MlpClassifier(), param_grid=param_grid, cv=3, scoring='accuracy', verbose=2)
searcher.fit(X_train, y_train)

# Get best parameters and accuracy
print("Best Parameters:", searcher.best_params_)
print("Best Cross-Validation Accuracy:", searcher.best_score_)

# Evaluate on test set
best_net = searcher.best_estimator_
y_pred = best_net.predict(X_test)
test_accuracy = accuracy_score(y_test, y_pred)
print("Test Accuracy:", test_accuracy)

print(best_net.model)


def check_MlpClassifier_binary_class():
# Load and prepare the dataset
data = load_breast_cancer()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

net = MlpClassifier(hidden_layers=(100, 50), act_names="ReLU", dropout_rates=0.2, act_output=None,
epochs=50, batch_size=8, optim="Adam", optim_paras=None,
early_stopping=True, n_patience=5, epsilon=0.0, valid_rate=0.1,
seed=42, verbose=True)
net.fit(X_train, y_train)
res = net.score(X_test, y_test)
print(res)
print(net.model)


def check_MlpClassifier_binary_class_gridsearch():
# Load and prepare the dataset
data = load_breast_cancer()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# Define parameters for grid search
param_grid = {
'hidden_layers': [(60,), (100,)],
'act_names': ["ReLU", "Tanh", "Sigmoid"],
'dropout_rates': [0.2, None],
'epochs': [10, 20],
'batch_size': [16, 24],
'optim': ['Adam', 'SGD'],
"early_stopping": [True],
"valid_rate": [0.1, 0.2],
}

# Hyperparameter tuning with GridSearchCV
searcher = GridSearchCV(estimator=MlpClassifier(), param_grid=param_grid, cv=3, scoring='accuracy', verbose=2)
searcher.fit(X_train, y_train)

# Get best parameters and accuracy
print("Best Parameters:", searcher.best_params_)
print("Best Cross-Validation Accuracy:", searcher.best_score_)

# Evaluate on test set
best_net = searcher.best_estimator_
y_pred = best_net.predict(X_test)
test_accuracy = accuracy_score(y_test, y_pred)
print("Test Accuracy:", test_accuracy)

print(best_net.model)


def check_MlpRegressor_single_output():
# Load and prepare the dataset
data = load_diabetes()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
X_scaler = StandardScaler()
X_train = X_scaler.fit_transform(X_train)
X_test = X_scaler.transform(X_test)

y_scaler = StandardScaler()
y_train = y_scaler.fit_transform(y_train.reshape(-1, 1))
y_test = y_scaler.transform(y_test.reshape(-1, 1))

net = MlpRegressor(hidden_layers=(30, 15), act_names="ELU", dropout_rates=0.2, act_output=None,
epochs=50, batch_size=8, optim="Adam", optim_paras=None,
early_stopping=True, n_patience=5, epsilon=0.0, valid_rate=0.1,
seed=42, verbose=True)
net.fit(X_train, y_train)
res = net.score(X_test, y_test)
print(res)
print(net.model)


def check_MlpRegressor_single_output_gridsearch():
# Load and prepare the dataset
data = load_diabetes()
X = data.data
y = data.target

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
X_scaler = StandardScaler()
X_train = X_scaler.fit_transform(X_train)
X_test = X_scaler.transform(X_test)

y_scaler = StandardScaler()
y_train = y_scaler.fit_transform(y_train.reshape(-1, 1))
y_test = y_scaler.transform(y_test.reshape(-1, 1))

# Define parameters for grid search
param_grid = {
'hidden_layers': [(60,), (100,)],
'act_names': ["ELU", "Tanh", "Sigmoid"],
'dropout_rates': [0.2, None],
'epochs': [10, 20],
'batch_size': [16, 24],
'optim': ['Adam', 'SGD'],
"early_stopping": [True],
"valid_rate": [0.1, 0.2],
}

# Hyperparameter tuning with GridSearchCV
searcher = GridSearchCV(estimator=MlpRegressor(), param_grid=param_grid, cv=3,
scoring='neg_mean_squared_error', verbose=2)
searcher.fit(X_train, y_train)

# Get best parameters and accuracy
print("Best Parameters:", searcher.best_params_)
print("Best Cross-Validation Accuracy:", searcher.best_score_)

# Evaluate on test set
best_net = searcher.best_estimator_
# y_pred = best_net.predict(X_test)
print(best_net.score(X_test, y_test))
print(best_net.model)


def check_MlpRegressor_multi_output():
# Load and prepare the dataset
# 1. Generate synthetic multi-output regression dataset
# Setting n_targets=2 for multi-output (2 target variables)
X, y = make_regression(n_samples=1000, n_features=10, n_targets=2, noise=0.1, random_state=42)

# Split into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
X_scaler = StandardScaler()
X_train = X_scaler.fit_transform(X_train)
X_test = X_scaler.transform(X_test)

y_scaler = StandardScaler()
y_train = y_scaler.fit_transform(y_train)
y_test = y_scaler.transform(y_test)

net = MlpRegressor(hidden_layers=(30, 15), act_names="ELU", dropout_rates=0.2, act_output=None,
epochs=50, batch_size=8, optim="Adam", optim_paras=None,
early_stopping=True, n_patience=5, epsilon=0.0, valid_rate=0.1,
seed=42, verbose=True)
net.fit(X_train, y_train)
res = net.score(X_test, y_test)
print(res)
print(net.model)


def check_MlpRegressor_multi_output_gridsearch():
# Load and prepare the dataset
X, y = make_regression(n_samples=1000, n_features=10, n_targets=2, noise=0.1, random_state=42)

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardize the features
X_scaler = StandardScaler()
X_train = X_scaler.fit_transform(X_train)
X_test = X_scaler.transform(X_test)

y_scaler = StandardScaler()
y_train = y_scaler.fit_transform(y_train)
y_test = y_scaler.transform(y_test)

# Define parameters for grid search
param_grid = {
'hidden_layers': [(60,), (100,)],
'act_names': ["ELU", "Tanh", "Sigmoid"],
'dropout_rates': [0.2, None],
'epochs': [10, 20],
'batch_size': [16, 24],
'optim': ['Adam', 'SGD'],
"early_stopping": [True],
"valid_rate": [0.1, 0.2],
}

# Hyperparameter tuning with GridSearchCV
searcher = GridSearchCV(estimator=MlpRegressor(), param_grid=param_grid, cv=3,
scoring='neg_mean_squared_error', verbose=2)
searcher.fit(X_train, y_train)

# Get best parameters and accuracy
print("Best Parameters:", searcher.best_params_)
print("Best Cross-Validation Accuracy:", searcher.best_score_)

# Evaluate on test set
best_net = searcher.best_estimator_
# y_pred = best_net.predict(X_test)
print(best_net.score(X_test, y_test))
print(best_net.model)


# check_CustomMLP_class()
check_MlpClassifier_multi_class()
# check_MlpClassifier_multi_class_gridsearch()
# check_MlpClassifier_binary_class()
# check_MlpClassifier_binary_class_gridsearch()
# check_MlpRegressor_single_output()
# check_MlpRegressor_single_output_gridsearch()
# check_MlpRegressor_multi_output()
# check_MlpRegressor_multi_output_gridsearch()
check_CustomMLP_class()
Loading

0 comments on commit 862bfbb

Please sign in to comment.