Method rand

tinygrad/tensor.py:567–591

Creates a tensor with the given shape, filled with random values from a uniform distribution over the interval `[0, 1)`. You can pass in `dtype` and `device` keyword arguments to control the data type and device of the tensor. ```python exec="true" source="above" session="tensor" resu

(*shape, device:str|None=None, dtype:DTypeLike|None=None, contiguous:bool=True)

  @staticmethod
  def rand(*shape, device:str|None=None, dtype:DTypeLike|None=None, contiguous:bool=True) -> Tensor:
    """
    Creates a tensor with the given shape, filled with random values from a uniform distribution over the interval `[0, 1)`.

    You can pass in `dtype` and `device` keyword arguments to control the data type and device of the tensor.

    ```python exec="true" source="above" session="tensor" result="python"
    Tensor.manual_seed(42)
    t = Tensor.rand(2, 3)
    print(t.numpy())
    ```
    """
    dt = to_dtype(dtype or dtypes.default_float)
    if not dtypes.is_float(dt): raise ValueError(f"rand only supports float dtypes, got {dt}")
    if not all_int(shape:=argfix(*shape)) or not all(s >= 0 for s in shape): raise ValueError(f"invalid input {shape=}")
    if device is not None and not isinstance(device, str): raise ValueError(f"rand only supports single device, got {device=}")
    device = cast(str, canonicalize_device(device))

    # if shape has 0, return zero tensor
    if (numel := prod(shape)) == 0: return Tensor.zeros(shape, device=device, dtype=dt)
    num = ceildiv(numel * dt.itemsize, 4)
    key, counter = Tensor._next_counter(device, num)
    bits = Tensor.random_bits(key, counter, num)
    out = Tensor._bits_to_rand(bits, shape, dt)
    return out.contiguous() if contiguous else out

  # ***** creation helper functions *****