Run GGUF Quantized Models Easily with KoboldCPP on Runpod

Lower VRAM usage and improve inference speed using GGUF quantized models in KoboldCPP with just a few environment variables.