print("START: BEFORE IMPORTS") import os import time import gradio as gr import copy from llama_cpp import Llama from huggingface_hub import hf_hub_download print("START: AFTER IMPORTS") try: print("START: BEFORE MODEL DOWNLOAD") start_load_time = time.time() model_path = hf_hub_download( repo_id="NousResearch/Hermes-2-Pro-Llama-3-8B-GGUF", filename="Hermes-2-Pro-Llama-3-8B-Q4_K_M.gguf", ) print(f"START: AFTER MODEL DOWNLOAD -- {time.time() - start_load_time}s") llm = Llama( model_path=model_path, n_ctx=2048, n_gpu_layers=-1, # change n_gpu_layers if you have more or less VRAM verbose=True ) print(f"START: AFTER LLAMA-CPP SETUP -- {time.time() - start_load_time}s") except Exception as e: print(e) def generate_text( message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p, ): messages = [{"role": "system", "content": system_message}] for val in history: if val[0]: messages.append({"role": "user", "content": val[0]}) if val[1]: messages.append({"role": "assistant", "content": val[1]}) messages.append({"role": "user", "content": message}) response = "" for chunk in llm.create_chat_completion( stream=True, max_tokens=max_tokens, temperature=temperature, top_p=top_p, messages=messages, ): part = chunk["choices"][0]["delta"].get("content", None) if part: response += part yield response demo = gr.ChatInterface( generate_text, title="llama-cpp-python on GPU", description="Running LLM with https://github.com/abetlen/llama-cpp-python", examples=[ ["How to setup a human base on Mars? Give short answer."], ["Explain theory of relativity to me like I’m 8 years old."], ["What is 9,000 * 9,000?"], ["Write a pun-filled happy birthday message to my friend Alex."], ["Justify why a penguin might make a good king of the jungle."], ], cache_examples=False, retry_btn=None, undo_btn="Delete Previous", clear_btn="Clear", additional_inputs=[ gr.Textbox(value="You are a friendly Chatbot.", label="System message"), gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"), gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"), gr.Slider( minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)", ), ], ) if __name__ == "__main__": demo.launch()