Iris Classifier based on Flask using Pytorch neural networks

---

Four sliders to enter data.

---

Enter data and press "Calculate".

---

---

How to install?

Clone

Clone this repo to your local machine using git clone https://github.com/CosmoSt4r/Iris-Classification

Required packages

To start the server you need the following packages:

• Flask
• Pytorch
• Pandas
• Scikit-learn

To install packages:

pip install flask torch pandas scikit-learn 

Starting the server

python main.py

or just open main.py

Opening app in browser

Open your browser and go to the 127.0.0.1:5000 address. You will see the login page.

Neural network

Neural network is built using Pytorch and named IrisNet.

---

• Structure: 4 x 16 x 16 x 3
• Activation functions: 2 Sigmoid functions and Softmax on output
• Loss function: Cross entropy loss
• Optimizer: Adam
• Learning rate: 0.001

You can change parameters as you want for better accuracy. Now average accuracy is 95-97%.

---