GAN training loop in PyTorch

Introduction

A GAN training loop teaches two models to work together: one creates fake data, the other learns to spot fakes. This helps computers make new, realistic data like images or sounds.

When you want to create new images that look real, like faces or art.

When you want to improve data for training other AI models by generating more examples.

When you want to learn how two models can compete and improve each other.

When you want to explore creative AI applications like style transfer or image editing.

Syntax

PyTorch

for epoch in range(num_epochs):
    for real_data in dataloader:
        batch_size = real_data.size(0)
        noise = torch.randn(batch_size, noise_dim)
        # Train Discriminator
        optimizer_d.zero_grad()
        real_output = discriminator(real_data)
        fake_data = generator(noise)
        fake_output = discriminator(fake_data.detach())
        loss_d_real = loss_function(real_output, torch.ones_like(real_output))
        loss_d_fake = loss_function(fake_output, torch.zeros_like(fake_output))
        loss_d = loss_d_real + loss_d_fake
        loss_d.backward()
        optimizer_d.step()

        # Train Generator
        optimizer_g.zero_grad()
        fake_output = discriminator(fake_data)
        loss_g = loss_function(fake_output, torch.ones_like(fake_output))
        loss_g.backward()
        optimizer_g.step()

The loop trains the discriminator first, then the generator.

Detach is used to avoid updating the generator when training the discriminator.

Examples

Basic loop over 5 epochs with noise input for generator.

PyTorch

for epoch in range(5):
    for real_images in dataloader:
        batch_size = real_images.size(0)
        noise = torch.randn(batch_size, noise_dim)
        fake_images = generator(noise)
        # Train discriminator and generator here

Shows how to calculate discriminator loss from real and fake data.

PyTorch

optimizer_d.zero_grad()
real_loss = criterion(discriminator(real_data), real_labels)
fake_loss = criterion(discriminator(fake_data.detach()), fake_labels)
loss_d = real_loss + fake_loss
loss_d.backward()
optimizer_d.step()

Sample Model

This code trains a simple GAN with a tiny generator and discriminator on fake 2D points. It prints losses for both models each epoch.

PyTorch

import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader, TensorDataset

# Simple Generator
class Generator(nn.Module):
    def __init__(self):
        super().__init__()
        self.net = nn.Sequential(
            nn.Linear(10, 16),
            nn.ReLU(),
            nn.Linear(16, 2),
            nn.Tanh()
        )
    def forward(self, x):
        return self.net(x)

# Simple Discriminator
class Discriminator(nn.Module):
    def __init__(self):
        super().__init__()
        self.net = nn.Sequential(
            nn.Linear(2, 16),
            nn.ReLU(),
            nn.Linear(16, 1),
            nn.Sigmoid()
        )
    def forward(self, x):
        return self.net(x)

# Create fake real data: points near (1,1)
real_data = torch.randn(100, 2) * 0.1 + 1
real_labels = torch.ones(100, 1)
fake_labels = torch.zeros(100, 1)

# DataLoader
dataset = TensorDataset(real_data, real_labels)
dataloader = DataLoader(dataset, batch_size=10, shuffle=True)

# Initialize models
generator = Generator()
discriminator = Discriminator()

# Loss and optimizers
criterion = nn.BCELoss()
optimizer_d = optim.Adam(discriminator.parameters(), lr=0.01)
optimizer_g = optim.Adam(generator.parameters(), lr=0.01)

num_epochs = 3
noise_dim = 10

for epoch in range(num_epochs):
    for real_batch, _ in dataloader:
        batch_size = real_batch.size(0)

        # Train Discriminator
        optimizer_d.zero_grad()
        real_output = discriminator(real_batch)
        real_loss = criterion(real_output, torch.ones_like(real_output))

        noise = torch.randn(batch_size, noise_dim)
        fake_data = generator(noise)
        fake_output = discriminator(fake_data.detach())
        fake_loss = criterion(fake_output, torch.zeros_like(fake_output))

        loss_d = real_loss + fake_loss
        loss_d.backward()
        optimizer_d.step()

        # Train Generator
        optimizer_g.zero_grad()
        fake_output = discriminator(fake_data)
        loss_g = criterion(fake_output, torch.ones_like(fake_output))
        loss_g.backward()
        optimizer_g.step()

    print(f"Epoch {epoch+1}/{num_epochs} - Loss D: {loss_d.item():.4f}, Loss G: {loss_g.item():.4f}")

Important Notes

Use .detach() on fake data when training the discriminator to avoid updating the generator.

Generator tries to fool the discriminator by making fake data look real.

Discriminator tries to correctly tell real from fake data.

Summary

A GAN training loop alternates training two models: discriminator and generator.

Discriminator learns to spot fake data, generator learns to create better fakes.

Loss values show how well each model is doing during training.

Practice

1. In a GAN training loop, what is the main role of the discriminator?

easy

A. To calculate the learning rate

B. To generate new fake data samples

C. To distinguish between real and fake data

D. To optimize the generator's weights directly

Solution

1. Step 1: Understand the discriminator's purpose

   The discriminator's job is to tell apart real data from fake data created by the generator.

2. Step 2: Differentiate roles in GAN

   The generator creates fake data, while the discriminator judges it. It does not generate data or adjust learning rates.

3. Final Answer:

   To distinguish between real and fake data -> Option C

4. Quick Check:

   Discriminator = spot fake data [OK]

Hint: Discriminator judges data authenticity, generator creates data [OK]

Common Mistakes:

• Confusing discriminator with generator
• Thinking discriminator generates data
• Assuming discriminator sets learning rate

2. Which of the following is the correct way to zero the gradients of both generator and discriminator optimizers in PyTorch during a GAN training loop?

easy

A. optimizer_G.clear(); optimizer_D.clear()

B. optimizer_G.zero_grad(); optimizer_D.zero_grad()

C. optimizer_G.reset_grad(); optimizer_D.reset_grad()

D. optimizer_G.zero(); optimizer_D.zero()

Solution

1. Step 1: Recall PyTorch optimizer method

   In PyTorch, the method to clear gradients before backpropagation is zero_grad().

2. Step 2: Check each option's validity

   Only optimizer_G.zero_grad(); optimizer_D.zero_grad() uses zero_grad(), which is the correct method. Other options use non-existent methods.

3. Final Answer:

   optimizer_G.zero_grad(); optimizer_D.zero_grad() -> Option B

4. Quick Check:

   Clear gradients = zero_grad() [OK]

Hint: Use zero_grad() to clear gradients before backward pass [OK]

Common Mistakes:

• Using non-existent optimizer methods
• Forgetting to zero gradients before backward
• Confusing zero_grad() with optimizer step()

3. Given the following PyTorch GAN training snippet, what will be the printed output for print(loss_D.item(), loss_G.item()) after one iteration?

real_data = torch.ones(4)
fake_data = torch.zeros(4)
criterion = torch.nn.BCELoss()

# Discriminator output
output_real = torch.tensor([0.9, 0.8, 0.95, 0.85])
output_fake = torch.tensor([0.1, 0.2, 0.05, 0.15])

loss_real = criterion(output_real, real_data)
loss_fake = criterion(output_fake, fake_data)
loss_D = (loss_real + loss_fake) / 2

# Generator tries to fool discriminator
output_fake_for_G = torch.tensor([0.4, 0.3, 0.5, 0.6])
loss_G = criterion(output_fake_for_G, real_data)

print(loss_D.item(), loss_G.item())

medium

A. Approximately 0.693 and 0.105

B. Approximately 0.0 and 1.0

C. Approximately 0.5 and 0.5

D. Approximately 0.136 and 0.831

Solution

1. Step 1: Calculate discriminator loss

   Loss_real = BCELoss([0.9,0.8,0.95,0.85], [1,1,1,1]) ≈ 0.136; Loss_fake = BCELoss([0.1,0.2,0.05,0.15], [0,0,0,0]) ≈ 0.136; Average loss_D ≈ 0.136.

2. Step 2: Calculate generator loss

   Generator tries to fool discriminator, so loss_G = BCELoss([0.4,0.3,0.5,0.6], [1,1,1,1]) ≈ 0.831.

3. Final Answer:

   Approximately 0.136 and 0.831 -> Option D

4. Quick Check:

   Discriminator loss < Generator loss [OK]

Hint: Discriminator loss low near 0, generator loss higher when fooled less [OK]

Common Mistakes:

• Mixing up real and fake labels in loss
• Confusing which loss belongs to generator
• Ignoring averaging of discriminator losses

4. Identify the error in this GAN training loop snippet and select the fix:

for real_data in dataloader:
    optimizer_D.zero_grad()
    output_real = discriminator(real_data)
    loss_real = criterion(output_real, torch.ones_like(output_real))

    fake_data = generator(torch.randn(batch_size, noise_dim))
    output_fake = discriminator(fake_data.detach())
    loss_fake = criterion(output_fake, torch.zeros_like(output_fake))

    loss_D = loss_real + loss_fake
    loss_D.backward()
    optimizer_D.step()

    optimizer_G.zero_grad()
    output_fake_for_G = discriminator(fake_data)
    loss_G = criterion(output_fake_for_G, torch.ones_like(output_fake_for_G))
    loss_G.backward()
    optimizer_D.step()

medium

A. Replace last optimizer_D.step() with optimizer_G.step()

B. Add optimizer_G.zero_grad() before loss_G.backward()

C. Detach fake_data before passing to generator

D. Use torch.zeros_like for output_fake_for_G labels

Solution

1. Step 1: Analyze optimizer steps

   After computing generator loss and backward, the generator's optimizer should update weights, not discriminator's.

2. Step 2: Identify incorrect optimizer call

   The code calls optimizer_D.step() twice; the second should be optimizer_G.step() to update generator.

3. Final Answer:

   Replace last optimizer_D.step() with optimizer_G.step() -> Option A

4. Quick Check:

   Generator update uses optimizer_G.step() [OK]

Hint: Generator updates need optimizer_G.step(), not optimizer_D.step() [OK]

Common Mistakes:

• Calling discriminator optimizer twice
• Forgetting to update generator weights
• Misplacing zero_grad calls

5. You want to improve GAN training stability by modifying the training loop. Which change below best helps prevent the discriminator from overpowering the generator?

hard

A. Train the discriminator less frequently than the generator

B. Increase the learning rate of the discriminator only

C. Remove the generator's loss calculation

D. Use the same optimizer instance for both models

Solution

1. Step 1: Understand discriminator overpowering

   If the discriminator learns too fast, the generator receives poor feedback and training stalls.

2. Step 2: Evaluate options to balance training

   Training discriminator less often than generator slows its learning, giving generator a chance to improve.

3. Final Answer:

   Train the discriminator less frequently than the generator -> Option A

4. Quick Check:

   Balance training frequency to stabilize GAN [OK]

Hint: Train discriminator less often to balance GAN training [OK]

Common Mistakes:

• Increasing discriminator learning rate worsens imbalance
• Removing generator loss stops generator learning
• Sharing optimizer causes parameter update conflicts