From 80974027c79143b2f95695b924df21d50981bd39 Mon Sep 17 00:00:00 2001
From: Adam Wojdyla <adam.wojdyla@outlook.com>
Date: Sun, 24 Apr 2022 09:06:57 +0200
Subject: [PATCH] Added pytorch DL python script

---
 DeepLearning/lab05_deepLearning.py | 94 ++++++++++++++++++++++++++++++
 1 file changed, 94 insertions(+)
 create mode 100644 DeepLearning/lab05_deepLearning.py

diff --git a/DeepLearning/lab05_deepLearning.py b/DeepLearning/lab05_deepLearning.py
new file mode 100644
index 0000000..1f49020
--- /dev/null
+++ b/DeepLearning/lab05_deepLearning.py
@@ -0,0 +1,94 @@
+import numpy as np
+import torch
+from torch import nn
+from torch.autograd import Variable
+from sklearn.datasets import load_iris
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import accuracy_score
+import torch.nn.functional as F
+import pandas as pd
+from sklearn import preprocessing
+
+
+class Model(nn.Module):
+    def __init__(self, input_dim):
+        super(Model, self).__init__()
+        self.layer1 = nn.Linear(input_dim, 160)
+        # self.layer2 = nn.Linear(320, 160)
+        self.layer2 = nn.Linear(160, 80)
+        self.layer3 = nn.Linear(80, 23)
+
+    def forward(self, x):
+        x = F.relu(self.layer1(x))
+        x = F.relu(self.layer2(x))
+        x = F.softmax(self.layer3(x))  # To check with the loss function
+        return x
+
+
+def load_dataset():
+    """ Load data from .csv file. """
+    cars = pd.read_csv('./Car_Prices_Poland_Kaggle.csv', usecols=[1, 4, 5, 6, 10], sep=',')
+    # cars = cars.iloc()
+    return cars
+
+
+def prepare_dataset(dataset):
+    """ Label make column"""
+    le = preprocessing.LabelEncoder()
+    mark_column = np.array(dataset[:]['mark'])
+    le.fit(mark_column)
+
+    print(list(le.classes_))
+    labels = le.transform(mark_column)
+    features = dataset.drop(['mark'], axis=1).to_numpy()
+
+    mm_scaler = preprocessing.MinMaxScaler()
+    features = mm_scaler.fit_transform(features)
+
+    return labels, features
+
+
+# Prepare dataset
+dataset = load_dataset()
+labels, features = prepare_dataset(dataset)
+features_train, features_test, labels_train, labels_test = train_test_split(features, labels, random_state=42, shuffle=True)
+#
+# import matplotlib
+#
+# plt = matplotlib.pyplot.hist(features, 16)
+
+# Training
+model = Model(features_train.shape[1])
+optimizer = torch.optim.Adam(model.parameters(), lr=0.01)
+loss_fn = nn.CrossEntropyLoss()
+epochs = 1000
+
+x_train, y_train = Variable(torch.from_numpy(features_train)).float(), Variable(torch.from_numpy(labels_train)).long()
+for epoch in range(1, epochs + 1):
+    print("Epoch #", epoch)
+    y_pred = model(x_train)
+    loss = loss_fn(y_pred, y_train)
+    print(f"The loss calculated: {loss}")
+
+    # Zero gradients
+    optimizer.zero_grad()
+    loss.backward()  # Gradients
+    optimizer.step()  # Update
+
+print(1)
+
+x_test = Variable(torch.from_numpy(features_test)).float()
+pred = model(x_test)
+
+pred = pred.detach().numpy()
+print(pred)
+
+print("The accuracy is", accuracy_score(labels_test, np.argmax(pred, axis=1)))
+
+# Checking for first value
+print(np.argmax(model(x_test[0]).detach().numpy(), axis=0))
+print(labels_test[0])
+torch.save(model, "iris-pytorch.pkl")
+
+saved_model = torch.load("iris-pytorch.pkl")
+print(np.argmax(saved_model(x_test[0]).detach().numpy(), axis=0))