Function dequantize 

pub fn dequantize(
    tensor: &Tensor,
    scale: f32,
    zero_point: i32,
) -> Result<Tensor>

Dequantize a quantized tensor using scale and zero_point