Deep Learning PyTorch Course, Anaconda Installation

Deep learning is a field of artificial intelligence that is especially used to learn patterns from large amounts of data and make predictions based on it. PyTorch is a popular library that helps implement deep learning easily. In this course, we will introduce how to install and set up PyTorch using the Anaconda environment.

1. What is Anaconda?

Anaconda is a Python distribution for data science, machine learning, and deep learning. This distribution includes a variety of libraries and tools, providing easy package management and environment management. By using Anaconda, you can easily create and manage Python environments suited for specific projects, which greatly helps prevent version conflicts between libraries.

1.1. Features of Anaconda

  • Package management: You can install and manage various packages through the conda package manager.
  • Environment management: You can create independent Python environments for each project to prevent library conflicts.
  • Diverse libraries: It includes many libraries related to data science such as NumPy, SciPy, Pandas, and Matplotlib.

2. Installing Anaconda

The process of installing Anaconda is simple. Let’s download and install Anaconda by following the steps below.

2.1. Downloading Anaconda

You can download the installation file from the official Anaconda website. Click the following link to go to the download page: Anaconda Distribution.

Choose the installation file suitable for your operating system (supports Windows, macOS, Linux).

2.2. Installation Process

Once the download is complete, run the installer. Although the process varies by operating system, it generally proceeds with the following steps.

  • Run the installer: Double-click the downloaded installation file to run it.
  • License agreement: Select the checkbox agreeing to the license agreement and click “Next”.
  • Select installation type: Choosing “Just Me” installs it only for personal use. Selecting “All Users” allows all users to use it.
  • Select installation path: You can leave the default installation path or change it to your desired path.
  • Other settings: You can choose whether to set environment variables (recommended).
  • Proceed with installation: Click the “Install” button to begin the installation.
  • Installation complete: Click the “Finish” button to end the installation.

2.3. Verifying Anaconda Installation

Once Anaconda is installed, open the Anaconda Prompt to verify that the installation was successful. You can search for “Anaconda Prompt” in the start menu to open it. 

conda --version

By entering the above command, the version of the installed conda will be displayed. If there is no output, the installation was not successful. In this case, please check the installation process again.

3. Creating a New Anaconda Environment

Now, let’s create a new environment to install the libraries needed for deep learning using Anaconda. Please proceed with the steps below.

3.1. Creating a New Environment

conda create --name mypytorch python=3.8

By entering the above command, a new environment named “mypytorch” will be created. Here, “python=3.8” sets the version of Python to be used in that environment.

3.2. Activating the Environment

conda activate mypytorch

Activate the newly created environment. The name of the prompt will change when the environment is activated.

3.3. Installing PyTorch

After activating the environment, install PyTorch using the command provided on the PyTorch official website. (It can be configured differently depending on the CUDA version you want to install.) 

conda install pytorch torchvision torchaudio cpuonly -c pytorch

The above command installs PyTorch, TorchVision, and Torchaudio for CPU only. To install for a GPU that supports CUDA, you can choose the corresponding CUDA version to install.

4. Verifying PyTorch Installation

To check whether PyTorch was installed correctly, run the Python interpreter and input the following code. 

python
import torch
print(torch.__version__)

If you enter the above code, the version of the installed PyTorch will be displayed. If no errors occur and the version is displayed, PyTorch has been successfully installed.

5. Simple PyTorch Code Example

Now that PyTorch is successfully installed, let’s write code to train a simple deep learning model. We will implement a simple linear regression model.

5.1. Generating Data

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

# Generate data
x = np.random.rand(100, 1) * 10  # Random values between 0 and 10
y = 2 * x + 1 + np.random.randn(100, 1)  # y = 2x + 1 + noise

# Visualize data
plt.scatter(x, y)
plt.xlabel('x')
plt.ylabel('y')
plt.title('Generated Data')
plt.show()

5.2. Defining the Model

import torch.nn as nn

# Define the linear regression model
class LinearRegressionModel(nn.Module):
    def __init__(self):
        super(LinearRegressionModel, self).__init__()
        self.linear = nn.Linear(1, 1)  # 1 input, 1 output

    def forward(self, x):
        return self.linear(x)

5.3. Defining the Loss Function and Optimizer

# Setting the loss function and optimizer
model = LinearRegressionModel()
criterion = nn.MSELoss()  # Mean Squared Error
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)  # Stochastic Gradient Descent

5.4. Training the Model

# Convert data to tensors
X = torch.from_numpy(x).float()  # Input
Y = torch.from_numpy(y).float()  # Output

# Train the model
for epoch in range(100):  # 100 epochs
    optimizer.zero_grad()  # Zero the gradients
    outputs = model(X)  # Model prediction
    loss = criterion(outputs, Y)  # Calculate loss
    loss.backward()  # Compute gradients
    optimizer.step()  # Update parameters

    if (epoch+1) % 10 == 0:
        print(f'Epoch [{epoch+1}/100], Loss: {loss.item():.4f}')

5.5. Visualizing the Training Result

# Visualize the training result
predicted = model(X).detach().numpy()  # Model predictions

plt.scatter(x, y, label='Original Data')
plt.plot(x, predicted, color='red', label='Fitted Line')
plt.xlabel('x')
plt.ylabel('y')
plt.title('Linear Regression Result')
plt.legend()
plt.show()

6. Conclusion

In this course, we covered how to install Anaconda and how to install PyTorch in that environment, as well as how to implement a simple linear regression model. Efficiently managing the Python environment through Anaconda and installing libraries related to deep learning is a very important first step in the field of data science and machine learning.

Furthermore, I recommend getting familiar with the basic structure and usage of PyTorch through practice and trying to implement various deep learning models. I hope your deep learning journey is an interesting and beneficial experience.