model.py

from typing import Iterator

from ctransformers import AutoModelForCausalLM

# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("OpenBuddy/openbuddy-gguf", model_file="openbuddy-mistral-7b-v13.1-Q3_K.gguf", model_type="mistral", gpu_layers=0)

def run(message: str,
        chat_history: list[tuple[str, str]],
        system_prompt: str,
        max_new_tokens: int = 1024,
        temperature: float = 0.3,
        top_p: float = 0.85,
        top_k: int = 5) -> Iterator[str]:
    history = []
    print(chat_history)
    result=""
    for i in chat_history:
        history.append({"role": "user", "content": i[0]})
        history.append({"role": "assistant", "content": i[1]})
    print(history)
    history.append({"role": "user", "content": message})
    for response in llm.create_chat_completion(history,stop=["</s>"],stream=True,max_tokens=-1,temperature=temperature,top_k=top_k,top_p=top_p,repeat_penalty=1.1):
        if "content" in response["choices"][0]["delta"]:
            result = result + response["choices"][0]["delta"]["content"]
            yield result