Can't get models to run on my 7900xtx (Windows 11).

[Matt2111]

Installs I've used (cmd):
set FORCE_CMAKE=1 && set CMAKE_ARGS=-DLLAMA_CLBLAST=on (Multiple variations of this including HIPBLAS)

pip install llama-cpp-python --upgrade --force-reinstall --no-cache-dir

Code:

from llama_cpp import Llama

llama = Llama("Models/mistral-7b-instruct-v0.1.Q8_0.gguf", n_gpu_layers=-1)

output = llama(
      "The following is a list of the 3 largest cities in the world by population within city limits.\n\n1.",
      max_tokens=1000,
)

print(output)

Tried various different n_gpu_layers including the total layer count of the model.

Any help would be massively appreciated!!🙏🙏🙏