What an Input Layer Actually Does
A free lesson from Swipe AI · Deep Learning & Neural Networks, level 2 · updated October 2026
By the Swipe team, founded by Colin · how we make lessons
The input layer is where raw data enters a network. For an image, that might be the brightness value of each pixel; for text, it might be a number representing each word or token. The input layer does not transform anything or make any decisions — it simply presents the raw data in a shape the rest of the network can work with.
Each input unit usually corresponds to one piece of raw data: one pixel's brightness, one token's identifier, one measured number in a table of data. Because of that, the number of input units is not really a free design choice — it is fixed by the shape of whatever data you are feeding in.
Image: 4 pixels arranged as a 2x2 square Pixel values: 12, 200, 45, 90 Input layer: 4 input units Unit 1 = 12 Unit 2 = 200 Unit 3 = 45 Unit 4 = 90
Different kinds of data shape the input layer differently. An image contributes one input unit per pixel (sometimes per color channel). A sentence contributes one input unit per token slot. A row in a spreadsheet contributes one input unit per column. Whatever the source, the input layer's only job is getting that raw data into the network in a consistent, numeric form.
What is the input layer's main job in a neural network?
The number of units in a network's input layer is fixed by the ____ of the data being fed in, not chosen freely.
The input layer performs the same kind of transformation work that hidden layers do.
Free ai tools
Keep your streak going in Swipe AI
Swipe through every level of Deep Learning & Neural Networks on your iPhone — with XP, stars, streaks, a personal advisor and the full track beyond these free levels.
Get the Swipe apps →