import torch import copy import gradio as gr import spaces from llama_cpp import Llama import os from huggingface_hub import hf_hub_download from threading import Thread HF_TOKEN = os.environ.get("HF_TOKEN", None) MODEL_ID = "google/gemma-2-27b-it" MODEL_NAME = MODEL_ID.split("/")[-1] MODEL_FILE = "gemma-2-27b-it-Q4_K_M.gguf" os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1" llm = Llama( model_path=hf_hub_download( repo_id=os.environ.get(MODEL_ID), filename=os.environ.get(MODEL_FILE), ), n_ctx=4096, n_gpu_layers=-1, chat_format="gemma", ) TITLE = "

Chatbox

" DESCRIPTION = f"""

MODEL: {MODEL_NAME}

Gemma is the large language model built by Google.
Feel free to test without log.

""" CSS = """ .duplicate-button { margin: auto !important; color: white !important; background: black !important; border-radius: 100vh !important; } h3 { text-align: center; } """ @spaces.GPU(duration=90) def stream_chat(message: str, history: list, temperature: float, max_new_tokens: int, top_p: float, top_k: int, penalty: float): print(f'message is - {message}') print(f'history is - {history}') conversation = [] for prompt, answer in history: conversation.extend([{"role": "user", "content": prompt}, {"role": "assistant", "content": answer}]) conversation.append({"role": "user", "content": message}) print(f"Conversation is -\n{conversation}") output = llm.create_chat_completion( messages=conversation, top_k=top_k, top_p=top_p, repeat_penalty=penalty, max_tokens=max_new_tokens, stream =True, temperature=temperature, ) for out in output: stream = copy.deepcopy(out) temp += stream["choices"][0]["text"] yield temp chatbot = gr.Chatbot(height=600) with gr.Blocks(css=CSS, theme="soft") as demo: gr.HTML(TITLE) gr.HTML(DESCRIPTION) gr.DuplicateButton(value="Duplicate Space for private use", elem_classes="duplicate-button") gr.ChatInterface( fn=stream_chat, chatbot=chatbot, fill_height=True, additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False), additional_inputs=[ gr.Slider( minimum=0, maximum=1, step=0.1, value=0.8, label="Temperature", render=False, ), gr.Slider( minimum=128, maximum=2048, step=1, value=1024, label="Max Tokens", render=False, ), gr.Slider( minimum=0.0, maximum=1.0, step=0.1, value=0.8, label="top_p", render=False, ), gr.Slider( minimum=1, maximum=20, step=1, value=20, label="top_k", render=False, ), gr.Slider( minimum=0.0, maximum=2.0, step=0.1, value=1.0, label="Repetition penalty", render=False, ), ], examples=[ ["Help me study vocabulary: write a sentence for me to fill in the blank, and I'll try to pick the correct option."], ["What are 5 creative things I could do with my kids' art? I don't want to throw them away, but it's also so much clutter."], ["Tell me a random fun fact about the Roman Empire."], ["Show me a code snippet of a website's sticky header in CSS and JavaScript."], ], cache_examples=False, ) if __name__ == "__main__": demo.launch()