Struct ClipVisionConfig 

pub struct ClipVisionConfig {
    pub vit_config: ViTConfig,
}

Configuration for the CLIP vision encoder.

Wraps a ViTConfig so that CLIP can share the same architectural hyper-parameter vocabulary as a standalone ViT model.

Fields

vit_config: ViTConfig

Underlying ViT hyper-parameters (image size, patch size, depth, …).

Implementations

impl ClipVisionConfig

pub fn new(vit_config: ViTConfig) -> Self

Create a ClipVisionConfig from an existing ViTConfig.

pub fn tiny() -> Self

Convenience constructor for a tiny CLIP encoder suitable for tests.

Delegates to ViTConfig::tiny.

Trait Implementations

impl Clone for ClipVisionConfig

fn clone(&self) -> ClipVisionConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ClipVisionConfig

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

Formats the value using the given formatter.