Function load_from_bytes 

pub fn load_from_bytes<M: DeserializeOwned>(
    data: &[u8],
    expected_type: ModelType,
) -> Result<M>

Load a model from a byte slice (spec 1.1 - Single Binary Deployment)

Enables the include_bytes!() pattern for embedding models directly in executables. This is the key function for zero-dependency ML deployment.

Arguments

• data - Raw .apr file bytes (e.g., from include_bytes!())
• expected_type - Expected model type (for type safety)

Example

use aprender::format::{load_from_bytes, ModelType};

// Embed model at compile time
const MODEL: &[u8] = include_bytes!("sentiment.apr");

fn main() -> Result<()> {
    let model: LogisticRegression = load_from_bytes(MODEL, ModelType::LogisticRegression)?;
    let prediction = model.predict(&input)?;
    Ok(())
}

Errors

Returns error on format error, type mismatch, or checksum failure