Struct icu_segmenter::provider::LstmDataFloat32

pub struct LstmDataFloat32<'data> {
    pub(crate) model: ModelType,
    pub(crate) dic: ZeroMap<'data, UnvalidatedStr, u16>,
    pub(crate) embedding: LstmMatrix2<'data>,
    pub(crate) fw_w: LstmMatrix3<'data>,
    pub(crate) fw_u: LstmMatrix3<'data>,
    pub(crate) fw_b: LstmMatrix2<'data>,
    pub(crate) bw_w: LstmMatrix3<'data>,
    pub(crate) bw_u: LstmMatrix3<'data>,
    pub(crate) bw_b: LstmMatrix2<'data>,
    pub(crate) time_w: LstmMatrix3<'data>,
    pub(crate) time_b: LstmMatrix1<'data>,
}

The struct that stores a LSTM model.

🚧 This code is considered unstable; it may change at any time, in breaking or non-breaking ways, including in SemVer minor releases. While the serde representation of data structs is guaranteed to be stable, their Rust representation might not be. Use with caution.

Fields

model: ModelType

Type of the model

dic: ZeroMap<'data, UnvalidatedStr, u16>

The grapheme cluster dictionary used to train the model

embedding: LstmMatrix2<'data>

The embedding layer. Shape (dic.len + 1, e)

fw_w: LstmMatrix3<'data>

The forward layer’s first matrix. Shape (h, 4, e)

fw_u: LstmMatrix3<'data>

The forward layer’s second matrix. Shape (h, 4, h)

fw_b: LstmMatrix2<'data>

The forward layer’s bias. Shape (h, 4)

bw_w: LstmMatrix3<'data>

The backward layer’s first matrix. Shape (h, 4, e)

bw_u: LstmMatrix3<'data>

The backward layer’s second matrix. Shape (h, 4, h)

bw_b: LstmMatrix2<'data>

The backward layer’s bias. Shape (h, 4)

time_w: LstmMatrix3<'data>

The output layer’s weights. Shape (2, 4, h)

time_b: LstmMatrix1<'data>

The output layer’s bias. Shape (4)

Trait Implementations

impl<'data> Clone for LstmDataFloat32<'data>

fn clone(&self) -> LstmDataFloat32<'data>

Returns a copy of the value.

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

Performs copy-assignment from source.

impl<'data> Debug for LstmDataFloat32<'data>

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

Formats the value using the given formatter.

impl<'data> PartialEq for LstmDataFloat32<'data>

fn eq(&self, other: &LstmDataFloat32<'data>) -> bool

This method tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

This method tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl<'a> Yokeable<'a> for LstmDataFloat32<'static>

type Output = LstmDataFloat32<'a>

This type MUST be Self with the 'static replaced with 'a, i.e. Self<'a>

fn transform(&'a self) -> &'a Self::Output

This method must cast self between &'a Self<'static> and &'a Self<'a>.

fn transform_owned(self) -> Self::Output

This method must cast self between Self<'static> and Self<'a>.

unsafe fn make(this: Self::Output) -> Self

This method can be used to cast away Self<'a>’s lifetime.

fn transform_mut<F>(&'a mut self, f: F)

where F: 'static + for<'b> FnOnce(&'b mut Self::Output),

This method must cast self between &'a mut Self<'static> and &'a mut Self<'a>, and pass it to f.

impl<'zf, 'zf_inner> ZeroFrom<'zf, LstmDataFloat32<'zf_inner>> for LstmDataFloat32<'zf>

fn zero_from(this: &'zf LstmDataFloat32<'zf_inner>) -> Self

Clone the other C into a struct that may retain references into C.

impl<'data> StructuralPartialEq for LstmDataFloat32<'data>