Function dequantize_bitsandbytes 

pub fn dequantize_bitsandbytes(
    state: &QuantState,
    config: &BitsAndBytesConfig,
) -> Result<Tensor>

Dequantize tensor from bitsandbytes format