Tensors

Mix.install([
  {:nx, "~> 0.5"}
])

Playing With Tensors

Nx.tensor(44)

The tensor arg can be one of many things

• A number (aka scalar)
• A boolean (synatic sugar on a special 0/1 tensor)
• A list of numbers, booleans, or list of numbers and booleans
• A tensor (why? well the opts we haven’t talke about yet would then be used to tweak an existing tensor)

scalar = Nx.tensor(48)
true_scalar = Nx.tensor(true)
false_scalar = Nx.tensor(false)
vector = Nx.tensor([4, 8])
matrix = Nx.tensor([[1, 1, 2], [2, 2, 4]])

{scalar, true_scalar, false_scalar, vector, matrix}

Tensor Options

You can explicitly specify the data type

• :u8
• :u16
• :u32
• :u64
• :s8
• :s16
• :s32
• :s64
• :bf16
• :f16
• :f32
• :f64

implicit_bool = Nx.tensor(1, type: :u8)
named_columns = Nx.tensor([[0, 1], [1, 3], [2, 5]], names: [:rows, :cols])

{implicit_bool, named_columns}

Tensor Operations

You will be manipulating data a lot with Tensors, and it will take some getting used to. Try changing the second tensor between

• 10
• [10]
• [10, 20, 30]
• [10, 20, 30, 40]

[1,2,3]
|> Nx.tensor()
|> Nx.add(Nx.tensor([10, 20, 30]))