From fea7c768b20a7cbe002846774f9945acbcb59ad7 Mon Sep 17 00:00:00 2001
From: wikbom <gedin8@gmail.com>
Date: Wed, 10 May 2023 13:10:44 +0200
Subject: [PATCH] lab5 nn learning and testing scripts

---
 eval.py     | 43 ++++++++++++++++++++++++++++++++
 learning.py | 70 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 113 insertions(+)
 create mode 100755 eval.py
 create mode 100755 learning.py

diff --git a/eval.py b/eval.py
new file mode 100755
index 0000000..def2dae
--- /dev/null
+++ b/eval.py
@@ -0,0 +1,43 @@
+#! /usr/bin/python3
+import numpy as np
+import torch
+from torch import nn
+import pandas as pd
+
+from sklearn.metrics import accuracy_score
+from sklearn.preprocessing import LabelEncoder
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self, input_dim):
+        super(Model, self).__init__()
+        self.layer1 = nn.Linear(input_dim, 50)
+        self.layer2 = nn.Linear(50, 20)
+        self.layer3 = nn.Linear(20, 2)
+        
+    def forward(self, x):
+        x = F.relu(self.layer1(x))
+        x = F.relu(self.layer2(x))
+        x = F.softmax(self.layer3(x))
+        return x
+
+
+test_df = pd.read_csv('testing_data.csv')
+
+X = test_df[['Pclass', 'Sex', 'Age','SibSp', 'Fare']]
+Y = test_df[['Survived']]
+Y = np.ravel(Y) 
+encoder = LabelEncoder()
+encoder.fit(Y)
+Y = encoder.transform(Y)
+
+model = Model(X.shape[1])
+model.load_state_dict(torch.load('model.pt'))
+
+
+x_test = torch.tensor(X.values, dtype=torch.float32)
+pred = model(x_test)
+pred = pred.detach().numpy()
+print ("The accuracy is", accuracy_score(Y, np.argmax(pred, axis=1)))
+
+np.savetxt('prediction.tsv', pred, delimiter='\t')
\ No newline at end of file
diff --git a/learning.py b/learning.py
new file mode 100755
index 0000000..e0c2250
--- /dev/null
+++ b/learning.py
@@ -0,0 +1,70 @@
+#!/usr/bin/python3
+import numpy as np
+import torch
+from torch import nn
+import pandas as pd
+import subprocess
+
+from sklearn.model_selection import train_test_split
+import torch.nn.functional as F
+from sklearn.preprocessing import LabelEncoder
+
+
+class Model(nn.Module):
+    def __init__(self, input_dim):
+        super(Model, self).__init__()
+        self.layer1 = nn.Linear(input_dim, 50)
+        self.layer2 = nn.Linear(50, 20)
+        self.layer3 = nn.Linear(20, 2)
+        
+    def forward(self, x):
+        x = F.relu(self.layer1(x))
+        x = F.relu(self.layer2(x))
+        x = F.softmax(self.layer3(x))
+        return x
+
+def print_(loss):
+    print ("The loss calculated: ", loss)
+
+if __name__ == "__main__":
+    df = pd.read_csv("train.csv")
+    df = df.dropna() #drop NA values
+
+    columns_to_normalize=['Age','Fare'] #NORMALIZATION
+    for colname in columns_to_normalize:
+        df[colname]=(df[colname]-df[colname].min())/(df[colname].max()-df[colname].min())
+            
+    X = df[['Pclass', 'Sex', 'Age','SibSp', 'Fare']] #only reasonable numerical data
+    Y = df[['Survived']]
+    X.loc[:,('Sex')].replace(['female', 'male'], [0,1], inplace=True) #categorical data transformed to 
+
+    X_train, X_test, Y_train, Y_test = train_test_split(X,Y, random_state=45, test_size=0.2, shuffle=True) #split the date into train and test sets
+
+    testing_data = pd.concat([X_test, Y_test], axis=1)
+    testing_data.to_csv('testing_data.csv', sep=',')
+
+    Y_train = np.ravel(Y_train) 
+    encoder = LabelEncoder()
+    encoder.fit(Y_train)
+    Y_train = encoder.transform(Y_train)
+
+    Xt = torch.tensor(X_train.values, dtype = torch.float32)
+    Yt = torch.tensor(Y_train, dtype=torch.long)
+
+    model = Model(Xt.shape[1])
+    optimizer = torch.optim.Adam(model.parameters(), lr=0.01)
+    loss_fn = nn.CrossEntropyLoss()
+    epochs = 1000
+
+    #TRAINING LOOP
+    for epoch in range(1, epochs+1):
+        print("Epoch #", epoch)
+        y_pred = model(Xt)
+        loss = loss_fn(y_pred, Yt)
+        print_(loss.item())
+        
+        optimizer.zero_grad()
+        loss.backward()
+        optimizer.step()
+
+    torch.save(model.state_dict(), 'model.pt')
\ No newline at end of file