Logistic Regression¶

In [1]:

# Uncomment the next line and run this cell to install sorix
#!pip install 'sorix @ git+https://github.com/Mitchell-Mirano/sorix.git@main'

# Uncomment the next line and run this cell to install sorix #!pip install 'sorix @ git+https://github.com/Mitchell-Mirano/sorix.git@main'

In [2]:

import os
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns

import os import numpy as np import matplotlib.pyplot as plt import pandas as pd import seaborn as sns

In [3]:

import sorix
from sorix.model_selection import train_test_split
from sorix.metrics import classification_report, confusion_matrix
from sorix.nn import Linear, BCEWithLogitsLoss
from sorix.optim import RMSprop, Adam

import sorix from sorix.model_selection import train_test_split from sorix.metrics import classification_report, confusion_matrix from sorix.nn import Linear, BCEWithLogitsLoss from sorix.optim import RMSprop, Adam

In [4]:

device = 'cuda' if sorix.cuda.is_available() else 'cpu'
device

device = 'cuda' if sorix.cuda.is_available() else 'cpu' device

✅ GPU basic operation passed
✅ GPU available: NVIDIA GeForce RTX 4070 Laptop GPU
CUDA runtime version: 13000
CuPy version: 13.6.0

Out[4]:

'cuda'

In [5]:

data=pd.read_csv("../data/Iris.csv")
data.head()

data=pd.read_csv("../data/Iris.csv") data.head()

Out[5]:

	Id	SepalLengthCm	SepalWidthCm	PetalLengthCm	PetalWidthCm	Species
0	1	5.1	3.5	1.4	0.2	Iris-setosa
1	2	4.9	3.0	1.4	0.2	Iris-setosa
2	3	4.7	3.2	1.3	0.2	Iris-setosa
3	4	4.6	3.1	1.5	0.2	Iris-setosa
4	5	5.0	3.6	1.4	0.2	Iris-setosa

In [6]:

data['Species'].unique()
data = data[data['Species'].isin(['Iris-setosa', 'Iris-versicolor'])]
data

data['Species'].unique() data = data[data['Species'].isin(['Iris-setosa', 'Iris-versicolor'])] data

Out[6]:

	Id	SepalLengthCm	SepalWidthCm	PetalLengthCm	PetalWidthCm	Species
0	1	5.1	3.5	1.4	0.2	Iris-setosa
1	2	4.9	3.0	1.4	0.2	Iris-setosa
2	3	4.7	3.2	1.3	0.2	Iris-setosa
3	4	4.6	3.1	1.5	0.2	Iris-setosa
4	5	5.0	3.6	1.4	0.2	Iris-setosa
...	...	...	...	...	...	...
95	96	5.7	3.0	4.2	1.2	Iris-versicolor
96	97	5.7	2.9	4.2	1.3	Iris-versicolor
97	98	6.2	2.9	4.3	1.3	Iris-versicolor
98	99	5.1	2.5	3.0	1.1	Iris-versicolor
99	100	5.7	2.8	4.1	1.3	Iris-versicolor

100 rows × 6 columns

In [7]:

labels2id = {label: i for i, label in enumerate(data['Species'].unique())}
id2labels = {i: label for i, label in enumerate(data['Species'].unique())}

labels2id = {label: i for i, label in enumerate(data['Species'].unique())} id2labels = {i: label for i, label in enumerate(data['Species'].unique())}

In [8]:

data['labels']=data['Species'].map(labels2id)
data = data[data['labels'].isin([0,1])]
data.head()

data['labels']=data['Species'].map(labels2id) data = data[data['labels'].isin([0,1])] data.head()

Out[8]:

	Id	SepalLengthCm	SepalWidthCm	PetalLengthCm	PetalWidthCm	Species	labels
0	1	5.1	3.5	1.4	0.2	Iris-setosa	0
1	2	4.9	3.0	1.4	0.2	Iris-setosa	0
2	3	4.7	3.2	1.3	0.2	Iris-setosa	0
3	4	4.6	3.1	1.5	0.2	Iris-setosa	0
4	5	5.0	3.6	1.4	0.2	Iris-setosa	0

In [9]:

plt.figure(figsize=(10,6))
for specie in data['Species'].unique():
    specie_data = data[data['Species'] == specie]
    plt.scatter(specie_data['PetalLengthCm'],specie_data['SepalWidthCm'],s=80 ,label=specie )
    plt.legend()
    plt.xlabel("Petal Length")
    plt.ylabel("Sepal Width")

plt.figure(figsize=(10,6)) for specie in data['Species'].unique(): specie_data = data[data['Species'] == specie] plt.scatter(specie_data['PetalLengthCm'],specie_data['SepalWidthCm'],s=80 ,label=specie ) plt.legend() plt.xlabel("Petal Length") plt.ylabel("Sepal Width")

In [10]:

independent_features=['PetalLengthCm','SepalWidthCm','PetalWidthCm','SepalLengthCm']
dependent_feature=['labels']

independent_features=['PetalLengthCm','SepalWidthCm','PetalWidthCm','SepalLengthCm'] dependent_feature=['labels']

In [11]:

data_train,data_test=train_test_split(data,test_size=0.2)

X_train=sorix.tensor(data_train[independent_features],device=device)
Y_train=sorix.tensor(data_train[dependent_feature],device=device)

X_test=sorix.tensor(data_test[independent_features],device=device)
Y_test=sorix.tensor(data_test[dependent_feature],device=device)

X_train.shape,Y_train.shape,X_test.shape,Y_test.shape

data_train,data_test=train_test_split(data,test_size=0.2) X_train=sorix.tensor(data_train[independent_features],device=device) Y_train=sorix.tensor(data_train[dependent_feature],device=device) X_test=sorix.tensor(data_test[independent_features],device=device) Y_test=sorix.tensor(data_test[dependent_feature],device=device) X_train.shape,Y_train.shape,X_test.shape,Y_test.shape

Out[11]:

(sorix.Size([80, 4]),
 sorix.Size([80, 1]),
 sorix.Size([20, 4]),
 sorix.Size([20, 1]))

In [12]:

model = Linear(4, 1).to(device)

loss_fn = BCEWithLogitsLoss()
optimizer = Adam(model.parameters(), lr=0.01)

model = Linear(4, 1).to(device) loss_fn = BCEWithLogitsLoss() optimizer = Adam(model.parameters(), lr=0.01)

In [13]:

for epoch in range(1000+1):
    y_pred = model(X_train)
    loss = loss_fn(y_pred, Y_train)
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()
    if epoch % 100 == 0:
        print(f"Epoch: {epoch} | Loss: {loss.item():.4f}")

for epoch in range(1000+1): y_pred = model(X_train) loss = loss_fn(y_pred, Y_train) optimizer.zero_grad() loss.backward() optimizer.step() if epoch % 100 == 0: print(f"Epoch: {epoch} | Loss: {loss.item():.4f}")

Epoch: 0 | Loss: 1.6742
Epoch: 100 | Loss: 0.2143
Epoch: 200 | Loss: 0.1048
Epoch: 300 | Loss: 0.0634

Epoch: 400 | Loss: 0.0433
Epoch: 500 | Loss: 0.0319
Epoch: 600 | Loss: 0.0247
Epoch: 700 | Loss: 0.0198

Epoch: 800 | Loss: 0.0163
Epoch: 900 | Loss: 0.0137
Epoch: 1000 | Loss: 0.0117

In [14]:

with sorix.no_grad():
    logits = model(X_test)
probs = sorix.sigmoid(logits)
preds = (probs > 0.5).astype('uint8')
acc_test = (preds == Y_test).mean()
plt.scatter(X_test[:,0],X_test[:,1],c=preds)
plt.title(f"Logistic Regression on Test Data(Accuracy:{100*acc_test.item():.2f}%)")
plt.show()

with sorix.no_grad(): logits = model(X_test) probs = sorix.sigmoid(logits) preds = (probs > 0.5).astype('uint8') acc_test = (preds == Y_test).mean() plt.scatter(X_test[:,0],X_test[:,1],c=preds) plt.title(f"Logistic Regression on Test Data(Accuracy:{100*acc_test.item():.2f}%)") plt.show()

Save Model¶

In [15]:

sorix.save(model.state_dict(),"logistic_model.sor")

sorix.save(model.state_dict(),"logistic_model.sor")

In [16]:

model2 = Linear(4, 1)
model2.load_state_dict(sorix.load("logistic_model.sor"))
model2.to(device)

model2 = Linear(4, 1) model2.load_state_dict(sorix.load("logistic_model.sor")) model2.to(device)

Out[16]:

Linear(in_features=4, out_features=1, bias=True)

In [17]:

with sorix.no_grad():
    logits = model2(X_test)
probs = sorix.sigmoid(logits)
preds = (probs.data > 0.5).astype('uint8')
acc_test = (preds == Y_test.data).mean()
preds = preds.get() if device == 'cuda' else preds

plt.scatter(X_test[:,0],X_test[:,1],c=preds)
plt.title(f"Logistic Regression on Test Data(Accuracy:{100*acc_test:.2f}%)")
plt.show()

with sorix.no_grad(): logits = model2(X_test) probs = sorix.sigmoid(logits) preds = (probs.data > 0.5).astype('uint8') acc_test = (preds == Y_test.data).mean() preds = preds.get() if device == 'cuda' else preds plt.scatter(X_test[:,0],X_test[:,1],c=preds) plt.title(f"Logistic Regression on Test Data(Accuracy:{100*acc_test:.2f}%)") plt.show()