import gradio as gr import os from pathlib import Path import argparse model_file = "openhermes-2-mistral-7b.Q4_0.gguf" if not os.path.isfile(model_file): os.system("wget -c https://huggingface.co./TheBloke/OpenHermes-2-Mistral-7B-GGUF/resolve/main/openhermes-2-mistral-7b.Q4_0.gguf") DEFAULT_MODEL_PATH = model_file from llama_cpp import Llama llm = Llama(model_path=model_file, model_type="mistral") def predict(input, chatbot, max_length, top_p, temperature, history): chatbot.append((input, "")) response = "" history.append(input) for output in llm(input, stream=True, temperature=temperature, top_p=top_p, max_tokens=max_length, ): piece = output['choices'][0]['text'] response += piece chatbot[-1] = (chatbot[-1][0], response) yield chatbot, history history.append(response) yield chatbot, history def reset_user_input(): return gr.update(value="") def reset_state(): return [], [] with gr.Blocks() as demo: gr.HTML("""