Basic Tensor Operations

Demir's first pass at the amplifier is already on the main console when the shift starts, and so is its output: Error in matMul: inner shapes (64) and (1) of Tensors with shapes 1,64 and 1,64 and transposeA=false and transposeB=false must match. One line of code, one error, sixty-four channels of raw signal going nowhere.

Chief Nazari

Raw data is useless to me. Clean it, combine the feeds, amplify what matters. And it runs at sensor speed or it doesn't run. We take two million readings a minute. Whatever that red text is, it is not two million readings a minute.

Cut to: Briefing

This lesson starts where you are standing: at the bug. You'll read the error, meet the two multiplication ops it is confusing, and repair the pipeline.

Read the Error First

Here is the code that produced it. The intent: amplify each of the 64 channels by its own gain factor. Position by position, gain times reading.

The error is precise once you can read it: matMul multiplies rows into columns, so it needs the left tensor's column count (64) to equal the right tensor's row count (1). But matching positions was the intent all along, and that is a different op with no inner-dimension rule at all. TensorFlow.js has both multiplications; the bug is reaching for the wrong one.

Element-wise Operations: The Repair

When you write arr.map((x, i) => x + arr2[i]), you're doing element-wise addition. tf.mul, tf.add, and tf.sub are the same idea across any number of dimensions, on the GPU. gains.mul(window) is the one-word fix for the broken amplifier:

When You DO Want matMul

So why does matMul exist at all? Because "every row dot-producted with every column" is the operation neural networks are made of, and no chain of element-wise ops can express it. It comes with the one inflexible shape rule the error message was enforcing:

[m, n] × [n, p] → [m, p]: the inner dimensions must match.

The left tensor's column count must equal the right tensor's row count. When they don't, you flip one with transpose(), which swaps a tensor's rows and columns, turning [3, 2] into [2, 3]:

Why care? Every layer of every neural network you'll build is essentially matMul(input, weights) + bias. The shape rule above is the single most common source of errors in ML code. The error that opened this lesson will find you again the day you build your first network layer; next time you'll read it in one pass.

Challenge

Add two tensors together to combine sensor readings.