45. Your First Neural Network in PyTorch

45.1 Build a Tiny Sequential Net

By Ravindra Bagale Last updated: 24 October 2026

Why. Put Chapters 41–44 together: a tiny net in PyTorch code you can run.

What. torch.nn.Sequential stacks layers. For a table with 3 features → hidden 8 → output 1 (binary score before sigmoid).

How.

Build the module

Import torch and torch.nn as nn.
Define Sequential: Linear → ReLU → Linear.
Print the module.

import torch
import torch.nn as nn

model = nn.Sequential(
 nn.Linear(3, 8),
 nn.ReLU(),
 nn.Linear(8, 1),
)
print(model)
x = torch.randn(4, 3) # batch of 4
print(model(x).shape) # (4, 1)

What you should see. Layer list and output shape (4, 1).