Function dequantize_embedding 

pub fn dequantize_embedding(q: &QuantizedVector) -> Result<Vector, QuantError>

Dequantizes a quantized vector back to approximate F32 representation

Arguments

• q - The quantized vector to dequantize

Returns

• Ok(Vector) - Successfully dequantized vector (F32 type)
• Err(QuantError) - If dequantization fails

Example

use lnmp_quant::{quantize_embedding, dequantize_embedding, QuantScheme};
use lnmp_embedding::Vector;

let original = Vector::from_f32(vec![0.12, -0.45, 0.33]);
let quantized = quantize_embedding(&original, QuantScheme::QInt8).unwrap();
let restored = dequantize_embedding(&quantized).unwrap();