Struct VisionInput 

pub struct VisionInput {
    pub pw: usize,
    pub ph: usize,
    pub pixels: Vec<f32>,
}

One image (or video frame) ready for the tower: its patch grid and the patch rows, [pw·ph, 3·patch²], already in the tower’s [-1, 1].

Fields

pw: usize

ph: usize

pixels: Vec<f32>

Implementations

impl VisionInput

pub fn n_patches(&self) -> usize

pub fn n_soft(&self, k: usize) -> usize

Soft tokens this input becomes ((pw/k)·(ph/k)).

Trait Implementations

impl Clone for VisionInput

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for VisionInput

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

Formats the value using the given formatter.