Struct FusionTensor 

pub struct FusionTensor<R>
where
    R: FusionRuntime,
{
    pub id: TensorId,
    pub shape: Shape,
    pub client: GlobalFusionClient<R>,
    pub dtype: DType,
    pub stream: StreamId,
    /* private fields */
}

Tensor primitive for the fusion backend for all kind.

Fields

id: TensorId

Tensor id.

shape: Shape

The shape of the tensor.

client: GlobalFusionClient<R>

The fusion client.

dtype: DType

The datatype of the tensor.

stream: StreamId

The current stream id this tensor is on.

Implementations

impl<R> FusionTensor<R>

where R: FusionRuntime,

pub fn to_ir_out(&self) -> TensorIr

Intermediate representation to be used when using an uninitialized tensor as output.

pub fn into_ir(self) -> TensorIr

Intermediate representation to be used when using an initialized tensor used as input.

Trait Implementations

impl<R> Clone for FusionTensor<R>

where R: FusionRuntime,

fn clone(&self) -> FusionTensor<R>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<R> Debug for FusionTensor<R>

where R: FusionRuntime,

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl<R> Drop for FusionTensor<R>

where R: FusionRuntime,

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.

impl<R> TensorMetadata for FusionTensor<R>

where R: FusionRuntime,

type Device = <R as FusionRuntime>::FusionDevice

The device type associated with the tensor.

fn dtype(&self) -> DType

Get the dtype of the tensor.

fn shape(&self) -> Shape

Get the shape of the tensor.

fn rank(&self) -> usize

Get the number of dimensions of the tensor.

fn device(&self) -> <FusionTensor<R> as TensorMetadata>::Device

Get the device associated with the tensor.

fn can_mut(&self) -> bool

Whether the tensor’s buffer can be mutated in place — i.e. this handle uniquely owns it, so an in-place op (slice_assign, an inplace kernel) writes the existing allocation instead of copying it first.

fn scheme(&self) -> QuantScheme

Get the quantization scheme for a quantized float tensor.