app.py

import spaces
import json
import subprocess
from llama_cpp import Llama
from llama_cpp_agent import LlamaCppAgent, MessagesFormatterType
from llama_cpp_agent.providers import LlamaCppPythonProvider
from llama_cpp_agent.chat_history import BasicChatHistory
from llama_cpp_agent.chat_history.messages import Roles
import gradio as gr
from huggingface_hub import hf_hub_download

# モデルのダウンロード
hf_hub_download(
   repo_id="bartowski/gemma-2-9b-it-GGUF",
   filename="gemma-2-9b-it-Q5_K_M.gguf",
   local_dir="./models"
)

hf_hub_download(
   repo_id="bartowski/gemma-2-27b-it-GGUF",
   filename="gemma-2-27b-it-Q5_K_M.gguf",
   local_dir="./models"
)

# 推論関数
@spaces.GPU(duration=120)
def respond(
   message,
   history: list[tuple[str, str]],
   model,
   system_message,
   max_tokens,
   temperature,
   top_p,
   top_k,
   repeat_penalty,
):
   chat_template = MessagesFormatterType.GEMMA_2

   llm = Llama(
       model_path=f"models/{model}",
       flash_attn=True,
       n_gpu_layers=81,
       n_batch=1024,
       n_ctx=8192,
   )
   provider = LlamaCppPythonProvider(llm)

   agent = LlamaCppAgent(
       provider,
       system_prompt=f"{system_message}",
       predefined_messages_formatter_type=chat_template,
       debug_output=True
   )

   settings = provider.get_provider_default_settings()
   settings.temperature = temperature
   settings.top_k = top_k
   settings.top_p = top_p
   settings.max_tokens = max_tokens
   settings.repeat_penalty = repeat_penalty
   settings.stream = True

   messages = BasicChatHistory()

   for msn in history:
       user = {
           'role': Roles.user,
           'content': msn[0]
       }
       assistant = {
           'role': Roles.assistant,
           'content': msn[1]
       }
       messages.add_message(user)
       messages.add_message(assistant)

   stream = agent.get_chat_response(
       message,
       llm_sampling_settings=settings,
       chat_history=messages,
       returns_streaming_generator=True,
       print_output=False
   )

   outputs = ""
   for output in stream:
       outputs += output
       yield outputs

# Gradioのインターフェースを作成
def create_interface(model_name, description):
   return gr.ChatInterface(
       respond,
       additional_inputs=[
           gr.Textbox(value=model_name, label="Model", interactive=False),
           gr.Textbox(value="You are a helpful assistant.", label="System message"),
           gr.Slider(minimum=1, maximum=4096, value=2048, step=1, label="Max 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",
           ),
           gr.Slider(
               minimum=0,
               maximum=100,
               value=40,
               step=1,
               label="Top-k",
           ),
           gr.Slider(
               minimum=0.0,
               maximum=2.0,
               value=1.1,
               step=0.1,
               label="Repetition penalty",
           ),
       ],
       retry_btn="Retry",
       undo_btn="Undo",
       clear_btn="Clear",
       submit_btn="Send",
       title=f"Chat with Gemma 2 using llama.cpp - {model_name}",
       description=description,
       chatbot=gr.Chatbot(
           scale=1,
           likeable=False,
           show_copy_button=True
       )
   )

# 各モデルのインターフェース
description_9b = """<p align="center">Gemma-2 9B it Model</p>"""
description_27b = """<p align="center">Gemma-2 27B it Model</p>"""

interface_9b = create_interface('gemma-2-9b-it-Q5_K_M.gguf', description_9b)
interface_27b = create_interface('gemma-2-27b-it-Q5_K_M.gguf', description_27b)

# Gradio Blocksで2つのインターフェースを並べて表示
with gr.Blocks() as demo:
   #gr.Markdown("# Compare Gemma-2 9B and 27B Models")
   with gr.Row():
       with gr.Column():
           interface_9b.render()
       with gr.Column():
           interface_27b.render()

if __name__ == "__main__":
   demo.launch()