Skip to main content

➡️ Forward Propagation

This is a fancy term for "Making a Guess."

🌊 The River Analogy​

Forward propagation is just the data flowing down the river, from the Input Layer, through all the Hidden Layers, until it splashes out the Output Layer.

During this phase, the network is not learning or updating anything. It is strictly calculating its best guess based on whatever weights it currently has.

🐍 Python Implementation​

Whenever you call model(data) in PyTorch, you are executing Forward Propagation!

import torch
import torch.nn as nn

# Re-using our network from the previous chapter
model = nn.Sequential(
    nn.Linear(10, 64),
    nn.ReLU(),
    nn.Linear(64, 2)
)

# 1. Collect Data
data = torch.randn(1, 10) 

# 2. Forward Propagation! (Data flows through the network)
guess = model(data) 

print("The model's forward pass result:", guess)