For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).
Mojo struct
VariadicTensors
struct VariadicTensors[mut: Bool, input: IO, dtype: DType, rank: Int, //, size: Int, io_spec: IOSpec[mut, input], *, static_specs: StaticTensorSpecList[static_specs.internals_list, static_specs.shapes_list, static_specs.strides_list]]
A tuple-like container of tensors representing variadic arguments from the graph compiler.
Implemented traits
AnyType,
Copyable,
ImplicitlyCopyable,
ImplicitlyDeletable,
Movable,
RegisterPassable,
Sized,
TrivialRegisterPassable
Methods
__init__
def __init__(ptrs: StaticTuple[UnsafePointer[Scalar[dtype], MutAnyOrigin], size], shapes: StaticTuple[IndexList[rank], size]) -> Self
Initialize the variadic tensor from tuples of pointers and shapes.
This is a bulk initialization of the VariadicTensors value from an array of pointers and an array of runtime shapes. This allows the graph compiler to avoid generating code to construct DynamicTensor values directly.
__len__
def __len__(self) -> Int
Returns the number of variadic arguments in the pack.
Returns:
Int: The number of variadic arguments.
__getitem_param__
def __getitem_param__[index: Int](self) -> ManagedTensorSlice[io_spec, static_spec=static_specs.__getitem_param__[index]()]
Returns the tensor at the given position in the variadic argument argument pack.
Parameters:
- index (
Int): The index into the variadic tensor arguments.
Returns:
ManagedTensorSlice[io_spec, static_spec=static_specs.__getitem_param__[index]()]: The tensor at the specified index.
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!