Struct PPLCNetPreprocessConfig 

pub struct PPLCNetPreprocessConfig {
    pub input_shape: (u32, u32),
    pub resize_filter: FilterType,
    pub resize_short: Option<u32>,
    pub normalize_scale: f32,
    pub normalize_mean: Vec<f32>,
    pub normalize_std: Vec<f32>,
    pub tensor_layout: TensorLayout,
}

Configuration for PP-LCNet model preprocessing.

Fields

input_shape: (u32, u32)

Input shape (height, width)

resize_filter: FilterType

Resizing filter to use

resize_short: Option<u32>

When set, resize by short edge to this size (keep ratio) then center-crop to input_shape.

This matches ResizeImage(resize_short=256) + CropImage(size=224) used by most PP-LCNet classifiers (e.g. PP-LCNet_x1_0_doc_ori, PP-LCNet_x1_0_table_cls).

Some specialized PP-LCNet classifiers (e.g. PP-LCNet_x1_0_textline_ori) use a direct resize to a fixed (height,width) without short-edge resize/crop; set this to None to match ResizeImage(size=[w,h]).

normalize_scale: f32

Scaling factor applied before normalization (defaults to 1.0 / 255.0)

normalize_mean: Vec<f32>

Mean values for normalization

normalize_std: Vec<f32>

Standard deviation values for normalization

tensor_layout: TensorLayout

Tensor data layout (CHW or HWC)

Trait Implementations

impl Clone for PPLCNetPreprocessConfig

fn clone(&self) -> PPLCNetPreprocessConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PPLCNetPreprocessConfig

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

Formats the value using the given formatter.

impl Default for PPLCNetPreprocessConfig

fn default() -> PPLCNetPreprocessConfig

Returns the “default value” for a type.