Opti 2

In [0]:

import numpy as np
import matplotlib.pyplot as plt
import torch

In [0]:

import torch.nn as nn

In [0]:

class LogisticRegression(nn.Module):
    def __init__(self):
        super().__init__()
        self.l1 = nn.Linear(8,6)
        self.l2 = nn.Linear(6,1)
    def forward(self,x):
        x = torch.relu(self.l1(x))
        x = torch.sigmoid(self.l2(x))
        return x

In [0]:

import pandas as pd

In [0]:

df = pd.read_csv('diabetes.csv.gz', compression='gzip', header=None, dtype=np.float32)

In [131]:

df

Out[131]:

	0	1	2	3	4	5	6	7	8
0	-0.294118	0.487437	0.180328	-0.292929	0.000000	0.001490	-0.531170	-0.033333	0.0
1	-0.882353	-0.145729	0.081967	-0.414141	0.000000	-0.207153	-0.766866	-0.666667	1.0
2	-0.058823	0.839196	0.049180	0.000000	0.000000	-0.305514	-0.492741	-0.633333	0.0
3	-0.882353	-0.105528	0.081967	-0.535354	-0.777778	-0.162444	-0.923997	0.000000	1.0
4	0.000000	0.376884	-0.344262	-0.292929	-0.602837	0.284650	0.887276	-0.600000	0.0
...	...	...	...	...	...	...	...	...	...
754	0.176471	0.015075	0.245902	-0.030303	-0.574468	-0.019374	-0.920581	0.400000	1.0
755	-0.764706	0.226131	0.147541	-0.454545	0.000000	0.096870	-0.776260	-0.800000	1.0
756	-0.411765	0.216080	0.180328	-0.535354	-0.735225	-0.219076	-0.857387	-0.700000	1.0
757	-0.882353	0.266332	-0.016393	0.000000	0.000000	-0.102832	-0.768574	-0.133333	0.0
758	-0.882353	-0.065327	0.147541	-0.373737	0.000000	-0.093890	-0.797609	-0.933333	1.0

759 rows × 9 columns

In [0]:

from sklearn.model_selection import train_test_split

In [0]:

X_train, X_test, y_train, y_test = train_test_split(df.iloc[:,:8], df.iloc[:,8], random_state=8)

In [148]:

X_train.shape, X_test.shape

Out[148]:

((569, 8), (190, 8))

In [0]:

from sklearn.preprocessing import StandardScaler

In [0]:

sc = StandardScaler()
X_train = sc.fit_transform(X_train)
X_test = sc.transform(X_test)

In [0]:

X_train = torch.from_numpy(X_train).float()
X_test = torch.from_numpy(X_test).float()
y_train = torch.from_numpy(y_train.values)
y_test = torch.from_numpy(y_test.values)

In [0]:

def accuracy(truth,preds):
    return (truth==preds).float().mean()

SGD¶

In [0]:

model = LogisticRegression()
criterion = nn.BCELoss()
optimizer = torch.optim.SGD(model.parameters(), lr = 0.001)

In [154]:

for epoch in range(501):
    pred = model.forward(X_train)
    loss = criterion(pred, y_train)
    loss.backward()
    optimizer.step()
    optimizer.zero_grad()
    with torch.no_grad():
        test_preds = model.forward(X_test)
        test_preds = test_preds>0.5
        acc = accuracy(y_test, test_preds.view(y_test.shape))
        if epoch%50==0:print(epoch, acc)

/usr/local/lib/python3.6/dist-packages/torch/nn/modules/loss.py:498: UserWarning: Using a target size (torch.Size([569])) that is different to the input size (torch.Size([569, 1])) is deprecated. Please ensure they have the same size.
  return F.binary_cross_entropy(input, target, weight=self.weight, reduction=self.reduction)

0 tensor(0.6421)
50 tensor(0.6368)
100 tensor(0.6368)
150 tensor(0.6421)
200 tensor(0.6474)
250 tensor(0.6474)
300 tensor(0.6526)
350 tensor(0.6579)
400 tensor(0.6737)
450 tensor(0.6842)
500 tensor(0.6842)

In [0]:

model = LogisticRegression()
criterion = nn.BCELoss()
optimizer = torch.optim.RMSprop(model.parameters(), lr = 0.001)

In [156]:

for epoch in range(501):
    pred = model.forward(X_train)
    loss = criterion(pred, y_train)
    loss.backward()
    optimizer.step()
    optimizer.zero_grad()
    with torch.no_grad():
        test_preds = model.forward(X_test)
        test_preds = test_preds>0.5
        acc = accuracy(y_test, test_preds.view(y_test.shape))
        if epoch%50==0:print(epoch, acc)

/usr/local/lib/python3.6/dist-packages/torch/nn/modules/loss.py:498: UserWarning: Using a target size (torch.Size([569])) that is different to the input size (torch.Size([569, 1])) is deprecated. Please ensure they have the same size.
  return F.binary_cross_entropy(input, target, weight=self.weight, reduction=self.reduction)

0 tensor(0.5368)
50 tensor(0.7526)
100 tensor(0.8000)
150 tensor(0.7895)
200 tensor(0.7842)
250 tensor(0.7737)
300 tensor(0.7684)
350 tensor(0.7684)
400 tensor(0.7737)
450 tensor(0.7737)
500 tensor(0.7895)

In [0]:

model = LogisticRegression()
criterion = nn.BCELoss()
optimizer = torch.optim.Adam(model.parameters(), lr = 0.001)

In [158]:

for epoch in range(501):
    pred = model.forward(X_train)
    loss = criterion(pred, y_train)
    loss.backward()
    optimizer.step()
    optimizer.zero_grad()
    with torch.no_grad():
        test_preds = model.forward(X_test)
        test_preds = test_preds>0.5
        acc = accuracy(y_test, test_preds.view(y_test.shape))
        if epoch%50==0:print(epoch, acc)

/usr/local/lib/python3.6/dist-packages/torch/nn/modules/loss.py:498: UserWarning: Using a target size (torch.Size([569])) that is different to the input size (torch.Size([569, 1])) is deprecated. Please ensure they have the same size.
  return F.binary_cross_entropy(input, target, weight=self.weight, reduction=self.reduction)

0 tensor(0.6158)
50 tensor(0.6842)
100 tensor(0.6842)
150 tensor(0.6842)
200 tensor(0.7211)
250 tensor(0.7632)
300 tensor(0.7789)
350 tensor(0.7842)
400 tensor(0.7842)
450 tensor(0.7789)
500 tensor(0.7842)

In [0]:

In [0]:

In [0]:

In [0]:

arr = np.arange(16).reshape((2,2,4))

In [119]:

arr

Out[119]:

array([[[ 0,  1,  2,  3],
        [ 4,  5,  6,  7]],

       [[ 8,  9, 10, 11],
        [12, 13, 14, 15]]])

In [120]:

arr.shape

Out[120]:

(2, 2, 4)

In [121]:

arr[0]

Out[121]:

array([[0, 1, 2, 3],
       [4, 5, 6, 7]])

In [122]:

arr[1]

Out[122]:

array([[ 8,  9, 10, 11],
       [12, 13, 14, 15]])

In [124]:

arr.transpose(1,0,2).shape

Out[124]:

(2, 2, 4)

In [125]:

arr.transpose(1,0,2)

Out[125]:

array([[[ 0,  1,  2,  3],
        [ 8,  9, 10, 11]],

       [[ 4,  5,  6,  7],
        [12, 13, 14, 15]]])

In [0]: