Error with quantization and lora

When I use a model with lora it works as expected but when I pass `--type q8_0` it does not work at all.
The model works without lora and with quantization.
I tested different loras and models with the same result.
