A Haskell library for tensors and linear algebra.
it's so pythonic!
Right now the library is split into two parts. The first has a type Vector a = [a] and type Matrix a = [[a]] and operations for these types. This is practically useless with the existence of the other half, which just has a datatype called tensor.
Tensors work like n-dimensional arrays. A tensor can either be a Value a or a list of type Tensor a. So a tensor can have as much "depth" as you'd like (a tensor of shape [3,2,8] has a depth of 3) and it always bottoms out with Values of type a.
The two most important functions that operate on tensors are broadcast and zipWithTen. These functions do what you'd expect them to. With them we can construct add and hadamard (multiplication) like add x y = zipWithTen (+) x y.