Loading…
Glossary term
The learned numbers in a model that determine how inputs are transformed into outputs.
Weights are the numbers on the connections of a neural network. Each one scales how strongly one unit influences the next.
Training is the process of finding good values for them. When people say a model has billions of parameters, most of those are weights.
A single unit's weights
import numpy as np
x = np.array([1.0, 2.0, 3.0]) # inputs
w = np.array([0.5, -1.0, 0.25]) # learned weights
b = 0.1 # bias
print(x @ w + b) # output = w·x + bThe weights are the model. Sharing the weights file is what 'open weights' means.