22 lines
955 B
Python

from llm_inference import load_and_generate_llm_inference
model_path = "/home/gregory/programming/testWithPython/Daredevil-GGUF/daredevil-8b-abliterated.Q8_0.gguf"
# model_path = "/home/gregory/programming/testWithPython/llama_q8/Llama-3.1-8B-Instruct-UD-Q8_K_XL.gguf"
question = "What is 11*11?"
prompt = f"""
This is a conversation between graph Dracula and his devoted servant Greg.
Dracula asks a question and Greg answers it. Greg always tries to be correct, but also doesn't want to be incorrect.
When Greg does not know the answer, he says so. After his answer Greg writes ===END OF ANSWER===
Example:
Question => How do I boil water?
Answer => You heat it to 100 degrees in a cauldron ===END OF ANSWER===
Question => WHat linux distro should I install
Answer => Artix linux is the best, all others are gay ===END OF ANSWER===
Question => {question}
Answer =>"""
output = load_and_generate_llm_inference(model_path=model_path, prompt=prompt)