In the directions for section “3 - The Sequential API,” we have the following:

For the first part of this assignment, you’ll create a model using TF Keras’ Sequential API, which allows you to build layer by layer, and is ideal for building models where each layer has exactly one input tensor and one output tensor.

My question: What is a tensor? This terminology was not discussed in the course videos for week 1.

Since this is the fourth course in a sequence of five, it assumes you already learned about TensorFlow (and therefore “tensors”) when it was covered in Course 2. That’s why it isn’t discussed in detail here.

From Wikipedia: a tensor is an algebraic object that describes a multilinear relationship between sets of algebraic objects related to a vector space.

Tensors are most easily understood as multi-dimensional matrices.

With regards to the error messages you’ll probably see a lot of, a tensor is a sort of data type that TensorFlow uses.

OK. Thanks. I did complete Course 2, and I do remember the introduction to TensorFlow, but I don’t remember a defintion being provided for the term “tensor.” I should probably go back and review Course 2.